UNLOCKING THE POWER OF HEALTHY LONGEVITY Compendium of Research for the Healthy Longevity Initiative AUGUST 2024 Editors: Prabhat Jha George Alleyne Paul Isenman Gisela M. Garcia Seemeen Saadat Jeremy Veillard Sameera Altuwaijri ii U N LO CK I N G T H E P OW ER O F H E ALT H Y LO N G E V I T Y © 2024 The World Bank 1818 H Street NW, Washington DC 20433 Telephone: 202-473-1000; Internet: www.worldbank.org Some rights reserved. This work is a product of The World Bank. The findings, interpretations, and conclusions expressed in this work do not necessarily reflect the views of the Executive Directors of The World Bank or the governments they represent. The World Bank does not guarantee the accuracy, completeness, or currency of the data included in this work and does not assume responsibility for any errors, omissions, or discrepancies in the information, or liability with respect to the use of or failure to use the information, methods, processes, or conclusions set forth. The boundaries, colors, denominations, links/footnotes and other information shown in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. The citation of works authored by others does not mean the World Bank endorses the views expressed by those authors or the content of their works. Nothing herein shall constitute or be construed or considered to be a limitation upon or waiver of the privileges and immunities of The World Bank, all of which are specifically reserved. Chapter 4: The economic value associated with avoidable mortality: a systematic assessment by cause of death across world regions © Harvard University. Rights and Permissions The material in this work is subject to copyright. Because The World Bank encourages dissemination of its knowledge, this work may be reproduced, in whole or in part, for noncommercial purposes as long as full attribution to this work is given. Attribution—Please cite the work as follows: “World Bank. 2024. Unlocking the Power of Healthy Longevity: Compendium of Research for the Healthy Longevity Initiative. Washington, DC: World Bank. Creative Commons Attribution CC BY 3.0 IGO. © World Bank.” Any queries on rights and licenses, including subsidiary rights, should be addressed to World Bank Publications, The World Bank, 1818 H Street NW, Washington, DC 20433, USA; fax: 202-522-2625; email: pubrights@worldbank.org. Cover design and typesetting: Karim Ezzat Khedr, Creative Director; email: karim.ezzat@gmail.com. COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E iii Compendium at a Glance Theoretical framework Chapter 1: Enhancing human capital and boosting productivity by tackling non-communicable diseases: results of a research initiative Chapter 2: Toward s a framework for impact pathways between non-communicable diseases, human capital and healthy longevity, and wellbeing outcomes Economic cost of avoidable mortality Chapter 3: The economic value of avoidable mortality Chapter 4: The economic value associated with avoidable mortality: a systematic assessment by cause of death across world regions Chapter 5: Rates of progress in mortality decline, 2000-2019 Behavior change and healthy longevity Chapter 6: Behavioral aspects of healthy longevity Chapter 7: Taxation of harmful products, including tobacco, alcohol and sugar-sweetened beverages (SSBs), and related topics Gender and Aging Chapter 8: Gender gaps in health and well-being of older adults: A review of the burden of non-communicable diseases and barriers to healthcare for women and men Chapter 9: Gendered responsibilities, elderly care, and labor supply: evidence from four middle-income countries Long-term care Chapter 10: Health and long-term care needs in a context of rapid population aging Chapter 11: Demand for and supply of long-term care for older persons in low- and middle-income countries Social protection and jobs Chapter 12: Exploring the labour market outcomes of the risk factors for non-communicable diseases: a systematic review Chapter 13: Productive longevity: what can work in low- and middle-income countries? Chapter 14: Adequacy pensions and access to healthcare: maintaining human capital during old age Prioritizing action Chapter 15: Priority setting for NCD control and health systems investments Chapter 16: Control of non-communicable diseases for enhanced human capital: the case for whole-of- society action Monitoring indicators for healthy longevity Chapter 17: Healthy longevity initiative: a performance dashboard for decision-making in low- and middle- income countries Chapter 18: Assessing human capital, non-communicable diseases, and healthy longevity in low- and middle- income countries: healthy longevity dashboard and the case for india iv COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E Contents Foreword xv Preface xvi Acknowledgements xvii Overview xx Abbreviations xxxix I Theoretical Framework 1 Chapter 1 Enhancing human capital and boosting productivity by tackling non- communicable diseases: results of a research initiative 2 Introduction and Overview 3 Research inputs and policy discussions 5 Recommendations for key stakeholders 27 Conclusions and way forward 29 References 30 Annex 1.1: Toronto workshop agenda 31 Annex 1.2: Toronto workshop, list of participants 32 Chapter 2 Towards a framework for impact pathways between NCDs, human capital and healthy longevity, economic and wellbeing outcomes 33 Introduction 34 Factors in developing a framework, some definitions and cross-cutting themes 35 Human capital, growth and country wealth 40 Inequality and gender dimensions of NCD and human capital pathways 53 Wellbeing effects of addressing NCDs 58 Policies and interventions for NCDs and healthy longevity 60 Conclusion 66 References 67 Annex 2.1 79 II Economic Cost of Avoidable Mortality 83 Chapter 3 The Economic Value of Avoidable Mortality 84 Introduction 85 Conceptual Framework 85 Methods 87 Results 90 Discussion 99 Conclusion 102 Dedication 102 References 102 CONTENTS v Chapter 4 The economic value associated with avoidable mortality: a systematic assessment by cause of death across world regions 104 Introduction 105 Methods 106 Results 109 Discussion 119 References 121 Chapter 5 Rates of progress in mortality decline, 2000-2019 123 Introduction 124 Rates of progress over 2000-09 and 2010-19 125 Countries’ rates of progress during 2000-09 and 2010-19 127 Rates of progress in cause-specific mortality 130 Cost of saving a life over time 132 References 133 Annex 5.1: Interrupted time-series (ITS) analysis methodology 134 Annex 5.2: Methods used to estimate country performance during 2000 to 2019 135 Annex 5.3: Model using to estimate critical income 153 Annex References 154 III Behavior Change and Healthy Longevity 156 Chapter 6 Behavioral aspects of healthy longevity 157 Introduction 158 A behavioral science framework for NCD policies 160 NCD formation: Healthy lifestyle and habit formation 163 NCD detection: Screening 167 NCD management 170 Recommendations for a policy toolkit 172 Conclusion 178 References 179 Annex 6.1 185 Chapter 7 Taxation of harmful products, including tobacco, alcohol and sugar-sweetened beverages, and related topics 186 Introduction: Consumption of unhealthy products and the healthy longevity agenda 187 The consumption of harmful products and their effects 188 Taxation: economic rationales and effectiveness 192 Effectiveness of price policies 193 Related topics 197 Combination of taxes and non-price interventions 199 Challenges from the industry 199 Conclusions 203 References 204 vi CONTENTS COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E IV Gender and Aging 216 Chapter 8 Gender gaps in health and well-being of older adults: A review of the burden of non-communicable diseases and barriers to healthcare for women and men 217 Introduction 218 Gender differentiated health outcomes of aging populations 222 Aging, gender, and demand for services 233 Supply for health and care services for aging populations 236 Women as long-term care providers 242 Discussion 246 References 249 Annex 8.1: Cancer mortality and prevalence by type for older women and men 257 Chapter 9 Gendered responsibilities, elderly care, and labor supply: evidence from four middle-income countries 259 Introduction 260 Data 263 Methodology 264 Results 264 Heterogeneity 275 Robustness check 276 Conclusion 278 References 279 Annex 9.1 280 V Long-term Care 286 Chapter 10 Health and long-term care needs in a context of rapid population aging 287 Introduction 288 Main trends in longevity 289 Aging, health, and the challenges for health care systems 293 Increasing risk of functional dependency and challenges for long-term care systems 314 Final remarks and key policy considerations 328 References 331 Consulted surveys and databases 340 Chapter 11 Demand for and supply of long-term care for older persons in low- and middle- income countries 341 Introduction 342 A demand side view - current LTC landscape in MICs 343 A supply side assessment current LTC landscape in LMICs 356 Discussion and policy considerations 363 References 367 CONTENTS vii VI Social Protection and Jobs 370 Chapter 12 Exploring the labour market outcomes of the risk factors for non-communicable diseases: a systematic review 371 Introduction 372 Methods 373 Results 374 Discussion 379 References 381 Annex 12.1 384 Chapter 13 Productive longevity: what can work in low- and middle-income countries? 385 Introduction 386 A global policy agenda for aging 387 Work in old age – some stylized facts 389 What holds back productive longevity? 395 Productive longevity: what could work in LMICs? 402 Conclusions: some meta-lessons for productive longevity 410 References 411 Annex 13.1 417 Chapter 14 Adequacy pensions and access to healthcare: maintaining human capital during old age 418 Introduction 419 Financial protection 420 Access to health and long-term care services 426 Conclusions 432 References 432 VII Prioritizing Action 434 Chapter 15 Priority setting for NCD control and health systems investments 435 Introduction 436 Identifying candidate NCD interventions 436 A framework for prioritizing NCD interventions 438 Policy implications 443 References 446 Annex 15.1: Methods 448 Chapter 16 Control of non-communicable diseases for enhanced human capital: the case for whole-of-society action: The case for whole-of-society action 450 Introduction 451 Why Focus on NCDs? 451 Impact of Covid-19 Pandemic on NCDs 457 Rationale for a whole-of-society approach to NCDs 458 Engaging non-staff actors (for profit and non-profit) 465 What has been achieved so far, and what have we learned? 468 Opportunities and challenges for a WOS approach: What have we learned? 483 The way forward 486 References 487 viii CONTENTS COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E VIII Monitoring Indicators for Healthy Longevity 491 Chapter 17 Healthy longevity initiative: a performance dashboard for decision-making in low- and middle-income countries 492 Introduction 493 Contextualizing dashboards for performance monitoring and strategic action 494 A common framework for healthy longevity measurement 494 Case studies: Sierra Leone, India, and Colombia 497 Discussion and future directions 503 Limitations 507 Conclusion 508 References 508 Annex 17.1: Examples of dashboards for global goals 511 Annex 17.2: Available data to better understand population ageing 519 Chapter 18 Assessing Human Capital, Non-communicable Diseases, and Healthy Longevity in Low- and Middle-Income Countries: Healthy longevity dashboard and the case for India 520 Introduction 521 Methods 521 Results 526 Discussion 529 Conclusion 530 References 531 Annex 18.1 532 Annex References 533 Appendix 1: Acknowledgements by Chapter 535 Appendix 2: Workshops and Consultations 539 Boxes Box 1.1: Key messages 5 Box 8.1: Gender gaps and the life course 219 Box 8.2: The healthy longevity framework and gender 220 Box 8.3: Understanding long-term care arrangements 237 Box 8.4: Provider behaviors and access to care for older adults 240 Box 8.5: Improving quality of care through standards and practice in the OECD 241 Box 10.1: When is a country “aging”? 290 Box 10.2: Health and socioeconomic status 298 Box 10.3: How is functional dependency measured? 314 Box 10.4: The importance of timely and sound data 330 Box 11.1: Data used for the demand side analysis 343 Box 16.1: Rationale for WoS approach – a sample of arguments 462 Box 16.2: Examples of Public-Private Partnerships for Non-Communicable Diseases 467 Box 16.3: Clínicas del Azúcar: a retail approach to diabetes care 482 Box 16.4: Georgia Healthcare Group: universal health care delivered through the private sector 483 Box 17.1: Case study of a healthy longevity dashboard in Sierra Leone 499 Box 17.2: Case study of a healthy longevity dashboard in India 501 Box 17.3: Case study of a healthy longevity dashboard in Colombia 502 CONTENTS ix Figures Figure 1.1: GDP and GDP per capital gains, average for 10 countries, with an NCD package 8 Figure 1.2: GDP with (perturbed) and without (projected) an NCD package 8 Figure 1.3: Illustration of the theoretical model in terms of individual optimizing behavior 10 Figure 1.4: Estimated health capital per capita (logarithmic scale), for 92 countries, 2011 US4 PPP 11 Figure 1.5: Effect of NCDs on health capital and health investment 12 Figure 1.6: Worsening adult male mortality in former Soviet states, 1950-2000 16 Figure 1.7: Substantial NCD burdens among children in low- and middle-income countries (1.7a: NCD share of total DALYs in <5 year olds; 1.7b: NCD share of total DALYs in 5-14 year olds) 21 Figure 1.8: Trends in the risk of dying between ages 50 and 69 in 25 countries, 1970-2010 25 Figure 1.9: Framework for analyzing health research needs 26 Figure 2.1: Human capital life course trajectory 38 Figure 2.2: Impact channels between NCDs and inclusive growth 42 Figure 2.3: Impact channels between NCDs and human capital across the life cycle 45 Figure 3.1: Frontier mortality rates from 2000, 2019, and 2050 90 Figure 3.2: Avoidable mortality as percentage of total mortality, year 2019 91 Figure 3.3: Avoidable mortality globally and by region, year, and age group 93 Figure 3.4: Gap in life expectancy at birth between observed/projected life expectancy and frontier life expectancy 94 Figure 3.5: Ratio of average VSMU to initial VSMU, for United States age group 0-1 94 Figure 3.6: Value of avoidable mortality as percentage of annual income for years 2000, 2019, 2021, and 2050 (projected) 95 Figure 3.7: Relationship between avoidable mortality rate, GNI per capita, and value of avoidable mortality 96 Figure 3.8: Age-group contributions to the total value of avoidable mortality and population composition by age in year 2019 97 Figure 3.9: Value of avoidable mortality for females and males as percentage of annual income in year 2019 97 Figure 4.1: Mortality frontiers (females and males) for the period 2000-2050 111 Figure 4.2: Economic values of avoidable mortality 117 Figure 4.3: Rates of change (percent per year) of the economic values of avoidable mortality assigned to selected causes of death 118 Figure 5.1: Change in risk of death at start of age range, by age group and country income level, 1990-2019 124 Figure 5.2: Preston curve showing the relationship between life expectancy at birth and income (Preston 1975) 132 Figure 5A.1: Schematic representation of an interrupted time series analysis 134 Figure 5A.2: Mortality rate, global and by country income level, by age and sex, between 2000 and 2019 135 Figure 5A.3: Correlation between performance in mortality decline during 2010-19 among males and females 146 Figure 6.1: Conceptual model of traditional and behavioral determinants of NCDs and healthy aging 160 Figure 6.2: Example of specific biases affecting hypertension formation, detection, and management 162 Figure 8.1: Structure and linkages for women and aging in health review 221 Figure 8.2: Cardiovascular mortality among older adults by regional and income aggregates 223 Figure 8.3: Ratio of female to male CVD mortality by regional and income aggregates 223 Figure 8.4: Cardio-vascular disease by sex for older populations (ages 45 years and above) 224 Figure 8.5: Deaths due to diabetes, ages 45+ (by income groups and regional aggregates) 225 x CONTENTS COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E Figure 8.6: Ratio of female to male mortality due to diabetes by regional and income aggregates 226 Figure 8.7: Gender gaps in diabetes for older populations (ages 45 years and above) 227 Figure 8.8: Deaths due to cancers, ages 45+ (by income groups and regional aggregates) 228 Figure 8.9: Ratio of female to male mortality due to cancers by regional and income aggregates 229 Figure 8.10: Gender gaps in cancers among older populations (ages 45 and above) 230 Figure 8.11: Gender gaps in depressive disorders among older adults 231 Figure 8.12: Gender gaps in self-harm among older adults (ages 45 and above) 232 Figure 8.13: Breakdown of health expenditure per capita in PPP, 2019 238 Figure 8.14: Out of pocket expenditures, 2019 239 Figure 8.15: Share of female LTC workers in select OECD countries (% of total, latest available year) 243 Figure 8A.1: Cancer mortality by type, Ages 45+ years 257 Figure 8A.2: Prevalence of cancer by type, ages 45+ years 258 Figure 9.1: Gross domestic product (GDP) per capita and male-female labor force participation gap, four countries, 2013-2018 261 Figure 9.2: Share of individuals (ages 25 to 59) with a coresident parent and young children in the household 266 Figure 9A.1: Kernel density of long-term care provision by age and country 282 Figure 9A.2: Work hours and care hours gradient 283 Figure 9A.3: Correlation between time spent on long-term care and employment 283 Figure 9A.4: Correlational effect of providing long-term care on individual employment by choice of control group 284 Figure 9A.5: Correlational effect of providing long-term care on annual earnings by choice of control group 284 Figure 9A.6: Correlational effect of providing long-term care on work hours by choice of control group 285 Figure 10.1: Population by broad age group (thousands), estimates 1950-2020 and projections 2025-2100 289 Figure 10.2: Years required or expected for the population share aged 65+ to rise from 7 percent to 14 percent 290 Figure 10.3: Healthy and unhealthy life expectancy at age 65 years, 1990 and 2017 291 Figure 10.4: Burden of disease, by disease type, 1990-2017 (% of total DALYs) 293 Figure 10.5: Percentage of countries that have a national strategy for chronic diseases and their risk factors, by demographic stage 303 Figure 10.6: Percentage of countries with policies that target main chronic diseases risk factors, by type of policies and country’s demographic stage 305 Figure 10.7: Countries that have national dementia plans in place, and percentage of their populations aged 80 or more 310 Figure 10.8: Projected increase in health care expenditures due to demographic effects, 2010- 2060 (in percentage points) 313 Figure 11.1: Having difficulty with any one ADL (Age 65+) 345 Figure 11.2: Needing help with any one ADL (Age 65+) 345 Figure 11.3: Having difficulty with any one IADL (age 65+) 348 Figure 11.4: Needing help with any one IADL (Age 65+) 348 Figure 11.5: Average number of cumulative ADL & IADL difficulties (Age 65+ having difficulty with an ADL or IADL) 351 Figure 11.6: Average number of cumulative ADL & IADL needs (Age 65+ needing help with an ADL or IADL) 351 Figure 11.7: Care mix (Age 65+ having difficulty with an ADL or IADL) 353 Figure 12.1: PRISMA flow diagram in search and selection of records 374 CONTENTS xi Figure 12.2: Proportion of studies by World Bank regions (World Bank 2022) 375 Figure 12.3: Differences in outcomes identified by biological sex 379 Figure 13.1: Demographic changes differ across regions, but aging is coming everywhere 389 Figure 13.2: Labor force participation for older workers falls with a country’s income 390 Figure 13.3: Labor force participation patterns are similar over the life cycle across richer and poorer countries 390 Figure 13.4: In MICs, pension coverage remains low 391 Figure 13.5: Healthy life expectancy has been increasing across the globe 393 Figure 13.6: A policy framework for longer and more productive working lives 395 Figure 13.7: Participation in formal or non-formal training, by age 398 Figure 13.8: Access to training through employers is limited, especially in L/MICs 400 Figure 14.1: Coverage of non-contributory pensions, % population 65+ 422 Figure 14.2: Old age social pension beneficiaries, % population 65+ 422 Figure 14.3: Benefit amount, % GDP per capita 424 Figure 14.4: Expenditure on non-contributory pensions, % GDP (latest year) 425 Figure 14.5: Projected share of 60+ in total population in SAR 425 Figure 14.6: Out-of-pocket expenditure (% of current health expenditure) 430 Figure 14.7: Average share of out of pocket health expenditure by quintile and presence of persons age 65+ 431 Figure 15.1: Cost per DALY averted for 24 clinical interventions, by health system type 439 Figure 15.2: Final prioritization of NCD interventions, by health system type 442 Figure 15.3: Contribution of NCD interventions towards avoidable all-cause mortality 445 Figure 16.1: Trend in global deaths and disease burden by major cause (annual rate, 1990-2019) 452 Figure 16.2: Trend in deaths and disease burden by major cause in LMICs (annual rate, 1990-2019) 453 Figure 16.3: Global deaths attributed to major risk factors, by sex, 2019 454 Figure 16.4: Global disease burden from NCDs by age, 1990 to 2019 455 Figure 16.5: Components of a “whole-of-society” approach 460 Figure 16.6. Framework for multisector causation and impact of non-communicable diseases 463 Figure 16.7: Global policy actions relevant to sugar-sweetened beverages 469 Figure 16.8: Singapore’s Healthy Living Master Plan 474 Figure 16.9: Singapore’s “War on Diabetes” 475 Figure 16.10: Thailand’s ThaiHealth Program 476 Figure 16.11: WBG investment lending on NCDs between FY16 and FY20 478 Figure 17.1: Healthy Longevity Initiative conceptual framework of human capital (HC) across the life course 495 Figure 17.2: Criteria for selecting healthy longevity indicators 496 Figure 17A.1a: Polio transition program monitoring and evaluation dashboard 511 Figure 17A.1b: Polio transition program monitoring and evaluation dashboard 511 Figure 17A.2: Global Malaria Dashboard – Campaign Dashboard 513 Figure 17A.3: Global Malaria Dashboard – Supply Chain Dashboard 514 Figure 17A.4: Page 1 of The Countdown Country Profile: A Tool For Action for Colombia 515 Figure 17A.5: Section of the Child Health and Well-being Dashboard 516 Figure 17A.6: Climate Action Tracker dashboard for Canada 517 Figure 17A.7: Section of the Open SDG Data Hub Country Profile: India 518 Figure 18.1: Criteria and process of indicator selection 523 xii CONTENTS COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E Maps Map 14.1: High shares of informality in developing countries 427 Map 14.2: Global Aging in 2020 428 Map 14.3: Global Aging in 2050 428 Spotlights Spotlight 13.1: Ok, Boomer! The aging brain and implications for work 395 Spotlight 13.2: Gray matter – training for the aging brain 404 Spotlight 13.3: Individual learning accounts to foster adult learning 405 Spotlight 13.4: Generation – facilitating labor market transitions over the working life 405 Spotlight 13.5: Firms with age-sensitive management practices 408 Tables Table 1.1: Determinants of log adult (40q30) mortality and effect of intervention on adult mortality 17 Table 2A.1: Mapping the human capital trajectory and its interactions with NCDs across the life cycle 79 Table 2A.2: Selected policies for human capital and NCDs across the life course 80 Table 2A.3: Human capital promoting policies during working life for accumulation, deployment and protection of HC 82 Table 3.1: Gap between observed and frontier life expectancy at birth in years, 2000, 2019 and 2050, globally and by region 90 Table 3.2: Total mortality and percentage of mortality that is avoidable in 2019, globally and by region, for all ages and by age 91 Table 3.3: Value of avoidable mortality in 2019 globally and by region and age group 95 Table 4.1: Causes of death 107 Table 4.2: Economic values of avoidable mortality as a percentage of annual income by region and cause of death, for the years 2000, 2019, and 2030 112 Table 5.1: Results from interrupted time series analysis of the change in mortality rates) from 2000-10 to 2010-19, by country income, age, and sex 125 Table 5.2: Top 10 and worst 10 performance rankings among 71 LMICs in the rate of decline of mortality, by sex and age, 2000-10 and 2010-19 127 Table 5.3: Diseases and conditions showing country income level and age group with the fastest and slowest progress in mortality reduction, 2000-19 131 Table 5.4: Global maximum survival rates and critical income levels, 1990 to 2019 133 Table 5.5: Top five LMICs with the lowest critical income for ages under 15 years, 15-49 years, and 50-69 years 133 Table 5A.1: Results of the hierarchical model predicting levels of mortality, by age and sex 137 Table 5A.2: Performance rankings of 71 LMICs in the rate of decline of mortality, by age and sex: 2000-10 and 2010-19 138 Table 5A.3: Change in performance in the annual rate of decline in mortality under 15 years, 15-49 years, 50-69 years, and 70-84 years, by sex from 2000-10 decade to 2010-2019 decade 146 Table 5A.4: Deaths avertable globally in 2019 from specific causes of deaths and age groups if rate of decline was same as that from HIV/AIDS (ages 0-14, 15-49) and tuberculosis (ages 50-69, 70+) 154 Table 6A.1: Summary of Behavioral Economics Concepts 185 Table 8.1: Key search teams for the review (not an exhaustive list) 220 Table 8.2: Typology to classify a country’s long-term care services and supports system 237 Table 9.1: Definition of long-term care by country 263 Table 9.2: Co-residency and long-term care provision among individuals ages 40 to 59 265 Table 9.3: Determinants of parental care, three-country pooled sample 267 Table 9.4: Labor supply among individuals ages 40 to 59 268 CONTENTS xiii Table 9.5. Correlational effect on employment of providing long-term care to parents, by gender, four-country pooled sample 271 Table 9.6: Gender difference in the effect on employment of providing long-term care to parents, four-country pooled sample 272 Table 9.7: Correlational effect of providing long-term care to parents on employment, by country 272 Table 9.8: Correlational effect on employment of care hours devoted to parents, pooled sample from three countries 274 Table 9.9: Heterogeneous effect of long-term care provision by education level, pooled sample from four countries 275 Table 9.10: Heterogeneous effect of long-term care provision by presence of children in household, pooled sample from four countries 276 Table 9.11: Sensitivity check: including the number of adult females/males in households 277 Table 9.12: Correlational effect on employment of the intensity of long-term care provision to parents, pooled sample from three countries 278 Table 9A.1: Correlational effect on employment of providing long-term care to parents 280 Table 9A.2: Gender difference in the effect on employment of providing long-term care to parents 281 Table 10.1: Years gained in life expectancy and healthy life expectancy at birth, 1990-2017. Both sexes, by demographic group 291 Table 10.2: Reasons given by adults aged 60 years or older for not accessing health care services, by country income category 298 Table 10.3: Long-term care services, by type 316 Table 11.1: Individual ADLS (Age 65+) 344 Table 11.2: Marginal effects from logistic regressions of having difficulty/needing help with at least one ADL 346 Table 11.3: Individual IADLs (Age 65+) 348 Table 11.4: Marginal effects from logistic regressions of having difficulty/needing help with at least one IADL 349 Table 11.5: Linear regressions of count of cumulative ADL & IADL difficulties/help needs 351 Table 11.6: Marginal effects from logistic models of receiving any care (Age 65+ having difficulty with an ADL or IADL) 354 Table 11.7: Marginal effects from logistic models of receiving formal care (Age 65+ having difficulty with an ADL or IADL) 355 Table 12.1: Total number of studies by countries + characteristics of 8 countries which produced the highest number of studies 376 Table 12.2: Frequencies of risk factors studied in included studies 376 Table 12.3: Frequencies of outcomes identified in included studies 376 Table 12.4: Frequency of studies associated with different combinations of exposures and outcomes 377 Table 12A.1: Search strategy for Ovid (MEDLINE) Sample 384 Table 13A.1: Labor market data for older adults 417 Table 14.1: Design and scope of non-contributory programs targeted to the elderly 420 Table 14.2: Additional dimensions to identify factors to incorporate in the concept of adequacy pensions 426 Table 14.3: Comparison of average shares of OOP health expenditure between households with elderly in the poorest quintile versus households with elderly in the richer quintiles 430 Table 15.1: Interventions considered in this paper 437 Table 15.2: Cost and impact of all NCD interventions by 2050, by world region 438 Table 15.3: Value for money among interventions outside the health sector 440 Table 15.4: Non-cost-effectiveness criteria for prioritizing clinical interventions 441 xiv CONTENTS COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E Table 15.5: Cost and impact of locally tailored high-priority NCD packages, by world region 443 Table 15.6: Total cost per capita for priority NCD interventions by year and country income group 444 Table 15.7: Workforce and facility requirements for implementing the high-priority NCD package by 2050 444 Table 16.1: Bi-directional relationship between selected non-health sectors and NCDs – some illustrations 463 Table 16.2: NCD control and prevention system in Japan 472 Table 16.3: Major NCD Management Initiatives in China 473 Table 16.4: WBG IPF Operations with a >50% focus on NCDs 479 Table 16.5: DPL operations relevant to NCDs supported by the World Bank Group, 2009-2019 482 Table 17.1: Key actions and related domains when developing a healthy longevity dashboard 495 Table 17.2: Healthy longevity dashboard indicators for Sierra Leone, India, and Colombia 497 Table 17.3: Data ecosystem features by domain and information infrastructure maturity level 504 Table 17.4: Proposed healthy longevity indicators by information infrastructure maturity level 505 Table 17.5: Key dashboard users, their purpose, and types of measures needed by users 506 Table 17A.1: Available data to better understand population ageing 519 Table 18.1: Key actions and related domains when developing a healthy longevity dashboard 522 Table 18.2: Indicator domains and indicators in the HLI dashboard for setting with moderate data infrastructure, and data sources for the India HLI dashboard 524 Table 18.3: Indicators of context– India, lower-middle income countries and world in 2019 526 Table 18.4: HLI health indicators (domains 1-6) and scores based on percentile rank approach (P) and z-score approach (Z) for India 527 Table 18.5: HLI social and economic indicators (domains 7-10) and scores based on percentile rank approach (P) and z-score approach (Z) for India 528 Table 18A.1: List of 33 lower-middle income countries with population of more than 7 million in 2020 532 Table 18A.2: Indicator data sources for the HLI India dashboard 532 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E xv Foreword There is no doubt that the world’s population is ageing at a dramatic speed. By 2050 at least a third of the population will be over 60 years of age, and a large proportion of these older adults will be in lower- and mid- dle-income countries (LMICs). Global aging brings new opportunities but also challenges for economies. Many nations are ill-prepared for the magnitude and pace of these demographic shifts. Timely and effective control of noncommunicable diseases (NCDs), the primary cause of adult deaths, is crucial, especially in LMICs – where they already account for over 70 percent of all deaths and a significant portion of disability. While aging is inevitable, healthy longevity is a choice. Over the last two years, the World Bank established the Healthy Longevity Initiative (HLI) to advance technical and operational knowledge about interventions, strategies and measurement promoting healthy longevity to sustain human capital productivity and identify policy priorities in key areas of engagement to reframe the aging policy narrative from one solely focused on risks to one that is also about finding opportunities. The World Bank’s Healthy Longevity Initiative (HLI) complements the work the World Bank has been lead- ing on the Human Capital Project through focusing the spotlight on middle and older age adults, and the main challenge to maintaining (and expanding) their human capital. The initiative is part of a broader effort by the World Bank, in collaboration with the University of Toronto and other prestigious academic and development partners, to spotlight the urgency of, and identify critical policy actions for addressing NCDs worldwide. It’s a call to action for policymakers everywhere to recognize the critical importance of combat- ing NCDs for the sake of our collective future. This compendium, featuring 18 insightful chapters, delves into the intricate relationship between NCDs, human capital, and productivity, drawing on a wealth of analytical work. It is the companion volume to the report, “Unlocking the Power of Healthy Longevity: Demographic Change, Noncommunicable Diseases, and Human Capital”. The compendium presents the comprehensive and cutting-edge background research on key issues related to aging, the economic strain of non-communicable diseases (NCDs), preventable mortal- ity, long-term care, behavioral change, labor market implication, social protection, and holistic government strategies that support the pursuit of a healthy, extended life – the pursuit of which begins in the womb. Promoting ‘healthy, productive and inclusive longevity’ which happens throughout the life course, will re- quire among others a focus on health, labor markets and skills, financial protection, long term care, and fiscal policies. The recommendations across the compendium underscore the necessity for cross-sectoral thinking and action, and the importance of collaboration between governments, civil society, and development part- ners to create positive synergies, given the multidimensional nature of aging. I hope that the research in this compendium and overall recommendations of the HLI will significantly con- tribute to policy dialogue and action to promote healthy longevity. Juan Pablo Uribe Global Director Health, Nutrition, and Population Global Practice World Bank xvi COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E Preface Global demographic transformations are rapidly occurring, marked by a significant increase in noncom- municable diseases (NCDs). These changes present a challenge comparable in scale to threats from climate change or pandemics. The World Bank’s Healthy Longevity Initiative (HLI) conducts analytic research to help countries address demographic and NCD challenges. The HLI originated from the United Nations General Assembly session on the NCD challenge in 2014. Following the launch of the World Bank’s Human Capital Index and project in 2017, which aimed to accelerate investments in essential health and education services, there was a clear need to adopt a life-course approach to improving health and well-being. There was a need to further elabo- rate the link between adult health and the effective use of human capital as a crucial resource for countries. Thus, the HLI was established, beginning with a partnership between the World Bank, World Health Organi- zation, and the University of Toronto. This collaboration produced the first report in 2018, titled Non-Com- municable Diseases and Human Capital Research Initiative. While the COVID pandemic slowed progress on the report, it underscored the connection between pandemics and adults with existing NCDs, highlighting the resulting significant health, social, and economic losses. The HLI focuses on generating and gathering robust analytical materials to support its overall report and encouraging academic and policy research on key components. This Compendium includes 18 detailed background papers written by over 90 authors from various countries, covering six key themes: theoretical base and economic costs; behavior change; financial and social protection; long-term care; gender; and pri- oritizing action. The Compendium papers were further curated by extensive consultations over four years. We hope that it will spur further analytical work and methodological advances at the country level. Latin American coun- tries have begun applying HLI methods to assess the economic value of avoidable mortality. Finally, we emphasize that analyses must be accompanied by expanded country-level action on NCDs, link- ing such investments to nationwide efforts to build human capital. The HLI Compendium Editors Prabhat Jha George Alleyne Paul Isenman Gisela M. Garcia Seemeen Saadat Jeremy Veillard Sameera Altuwaijri August 15, 2024 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E xvii Acknowledgements The Healthy Longevity Initiative (HLI) is a multi-year, multi-disciplinary analytical study to advance tech- nical and operational knowledge about interventions, strategies and measurement promoting healthy lon- gevity and support human capital formation and maintenance through better prevention and control of non-communicable diseases (NCDs). This work was led by Sameera Altuwaijri (Global Lead, Population and Development, Health, Nutrition and Population Global Practice, World Bank), and Prabhat Jha (Professor, Dalla Lana School of Public Health), along with a core team consisting of Sir George Alleyne (Director Emeritus, Pan American Health Organi- zation), Paul Isenman (Former Policy and Operations Manager, World Bank and OECD), Seemeen Saadat (Consultant, Population and Development, Health, Nutrition, and Population Global Practice, World Bank), Jeremy Veillard (Lead Health Specialist, Latin America and the Caribbean, World Bank), Daphne Wu (Re- search Coordinator, Centre for Global Health Research, University of Toronto), Gisela Garcia (Evaluation Officer, IEG Human Development and Corporate Program, World Bank), Meriem Boudjadja (Consultant, Health, Nutrition, and Population, East Asia and the Pacific, World Bank), Debapriya Chakraborty (Re- search Coordinator, Center for Global Health Research, University of Toronto), and Victoria Haldane (Ph.D. Candidate, Institute of Health Policy, Management and Evaluation, University of Toronto). The HLI builds on the Non-Communicable Diseases and Human Capital Research Initiative, which was spearheaded by Jeremy Veillard. The HLI was conducted under the supervision of Juan Pablo Uribe (Global Director, Health, Nutrition, and Population Global Practice) and Monique Vledder (Practice Manager, Health, Nutrition, and Population Global Practice). The work was conducted in collaboration with colleagues from the Social and Protection and Jobs Global Practice who contributed several chapters to this compendium. The team would like to thank Michal Rutkowski (Regional Director, Eastern Europe and Central Asia/former Global Director, So- cial Protection and Jobs Global Practice), as well as to the Human Capital Project team, led by Iffath Sharif (Global Director, Social Protection and Jobs Global Practice/former Manager, Human Capital Project) and Gabriel Demombynes (Manager, Human Capital Project) for their support and contributions. Overall guid- ance for this work was provided by Mamta Murti, Vice President, Human Development, World Bank. This compendium presents the wealth of research conducted under the HLI. In all there are 18 chapters or- ganized by theme. Each chapter was developed by a dedicated team of Bank staff, researchers and academics, and peer reviewed by experts in their respective fields. The HLI team would like to thank all authors and peer reviewers for their contributions to the Initiative: Chapter 1 was coordinated by Sir George Alleyne (PAHO), Timothy Evans (McGill University), Alexander Irwin (Independent Global Health Writer and Researcher), Prabhat Jha (University of Toronto), and Jeremy Veillard (World Bank). It is based on the proceedings of a workshop titled, Non-Communicable Diseases and Human Capital Analytic Work and Key Messaging Workshop, hosted in Toronto by the University of Toronto Dalla Lana School of Public Health on July 9-10, 2019. Chapter 2 was authored by Phillip O’Keefe (University of New South Wales) and Victoria Haldane (Universi- ty of Toronto). Chapter 3 was authored by Angela Y. Chang (University of Southern Denmark), Gretchen A. Stevens (Independent Researcher), Diego S. Cardoso (Purdue University), Bochen Cao (WHO), and Dean T. Jamison (UCSF). Chapter 4 was written by Stéphane Verguet (Harvard T.H. Chan School of Public Health), Sarah Bolongaita (Harvard T.H. Chan School of Public Health), Angela Y. Chang (University of Southern Denmark), Diego S. Cardoso (Purdue University), and Gretchen A. Stevens (Independent Researcher). xviii ACKNOWLEDGEMENTS COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E Chapter 5 was drafted by Daphne C. Wu (University of Toronto), Debapriya Chakraborty (University of Toronto), Ryan Hum (University of Toronto), Prabhat Jha (University of Toronto), and Dean T. Jamison (University of California at San Francisco). Chapter 6 was developed by Ana Maria Rojas (World Bank), Ana Maria Munoz Boudet (World Bank), Ellen Moscoe (World Bank), Julian Jamison (University of Exeter), and Carlos Riumallo Herl (Erasmus University). Chapter 7 was authored Guillermo Paraje (Escuela de Negocios, Universidad Adolfo Ibáñez), Prabhat Jha (University of Toronto), William Savedoff (Social Insight), and Alan Fuchs (World Bank). Daniel Araya provided useful research assistance. Chapter 8 was written by See- meen Saadat (World Bank), Meriem Boudjadja (World Bank), and Sameera Altuwaijri (World Bank). Chap- ter 9 was authored by Roberta Gatti (World Bank), Daniel Halim (World Bank), Allen Hardiman (University of Illinois), and Shuqiao Sun (World Bank). Chapter 10 was composed by Natalia Aranco Araújo (World Bank) and Gisela M. Garcia (World Bank). Chapter 11 was developed by Elena Glinskaya (World Bank), Xiaohui Hou (World Bank), Zhanlian Feng (RTI International), Marco Angrisani (University of Southern California), Guadalupe Suarez (Research Triangle Institute), Jigyasa Sharma (World Bank), Drystan Phillips (University of Southern California) , Jenny Wilkens (University of Southern California), Jinkook Lee (Uni- versity of Southern California), Yeeun Lee Yoo (University of Southern California), Samuel Lau (University of Southern California), Hae Yeun Park (University of Southern California), and Yizhou Chen (University of Southern California). Chapter 12 was written by Debapriya Chakraborty (University of Toronto), Daphne Wu (University of Toronto), and Prabhat Jha (University of Toronto). and David Lightfoot (Unity Health) supported the literature search. Chapter 13 was written by Sara Johansson de Silva (World Bank) and Ind- hira Santos (World Bank). Chapter 14 was authored by Gustavo Demarco (World Bank), Johannes Koettl (World Bank), Miglena Abels (World Bank) and Andrea Petrelli (World Bank), with support from Claudia Rodriguez Alas (World Bank) and Ana Sofia Martinez Cordova (World Bank) on household survey data analysis, and from Danilo Aristizabal (World Bank), who prepared the calculations for the household survey analysis. Chapter 15 was authored by David Watkins (University of Washington), Sali Ahmed (University of Washington), and Sarah Pickersgill (University of Washington). Chapter 16 was drafted by Ramesh Govin- daraj (World Bank) and Sundararajan Srinavasa Gopalan (Independent Researcher) with support from Nora Wu (World Bank) and Charles Dalton (IFC). Chapter 17 was composed by Victoria Haldane (University of Toronto), Gisela M. Garcia (World Bank), Tahir Bockarie (University of Warwick), Daphne Wu University of Toronto), Cristian A Herrera (World Bank), Maria Luisa Latorre Castro (Fundación Universitaria Juan N Corpas), Debapriya Chakraborty (University of Toronto), Beverly Essue (University of Toronto), Prabhat Jha (University of Toronto), Jeremy Veillard (World Bank); and Chapter 18 was developed by Daphne Wu (University of Toronto), Jeremy Veillard (World Bank) Victoria Haldane (University of Toronto), Seemeen Saadat (World Bank) and Prabhat Jha (University of Toronto). Many reviewers offered extensive advice and insightful comments, including (in alphabetical order of last name): Shambhu Acharya (World Health Organization), Tanima Ahmed (World Bank), Francisca Akala (World Bank), Faiza Benhadid (Independent Expert on Gender and Human Rights), Indu Bhushan (Part- nership for Impact), Adriana Blanco (PAHO), Danielle Bloom (World Bank), Mukesh Chawla (World Bank), Sarbani Chakraborty (Access Accelerated), Pedro Conceicao (UNDP), Damien De Walque (World Bank), Gustavo Demarco (World Bank), Gabriel Demom¬bynes (World Bank), Erica Di Ruggiero (University of Toronto), Daniel Dulitzky (World Bank), Patrick Hoang-Vu Eozenou (World Bank), Beverley Essue (Uni- versity of Toronto), Ian Forde (World Bank), Linda Fried (Columbia University Mailman School of Public Health), Cristian Hererra (World Bank), John Giles (World Bank), Elena Glinskaya (World Bank), Vivek Goel (University of Toronto), Amparo Elena Gordillo-Tobar (World Bank), Michele Gragnolati (World Bank), Sue Horton (University of Waterloo), Dean T. Jamison (University of California, San Francisco), Alexey Kulikov (World Health Organization), Christoph Kurowski (World Bank), Victoria Levin (World Bank), Kate Mandeville (World Bank), VR Muraleedharan (Indian Institute of Technology Madras), Rachel Nugent (Research Triangle Institute (RTI) International), Patrick Osewe (Asian Development Bank), Ceren Ozer (World Bank), Patrick Petit (International Monetary Fund), John Pig¬gott (University of New South Wales), Usha Ram (International Institute of Population Sciences), Gonzalo Reyes (World Bank), Hana Ross (University of Cape Town), Akshar Saxena (Nanyang Technological University), Norbert Schady (World Bank), Miriam Schneidman (World Bank), Daniel Sellen (University of Toronto), Owen Smith (World COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E ACKNOWLEDGEMENTS xix Bank), Marc-Francois Smitz (World Bank), Victoria Strokova (World Bank), Jeff Sturchio (Rabin Martin), Cornelis Peter Van Walbeek (University of Cape Town), Michael Weber (World Bank), Derek Yach (Inde- pendent Global Health Consultant), and Feng Zhao (World Bank). All chapters benefited from feedback from members of the core HLI team at various stages and from par- ticipants at the Healthy Longevity Initiative Technical Workshop held in Mexico City in May 2022 hosted by Escuela de Salud Pública De México, World Bank, and University of Toronto, and the Healthy Longevity Initiative’s Author Workshop held in October 2022 at the World Bank in Washington D.C (Annex A.2). A special thanks to Sir George Alleyne and Paul Isenman for their thorough review of all chapters. Please see Annex A.1 for chapter by chapter acknowledgements. The chapters also benefitted from an editorial review by Alexander Irwin. Karim Ezzat provided graphics and typesetting support while Arlene Lucindo Fitz-Patrick (World Bank), Jocelyn Haye (World Bank), and Venus Jaraba (University of Toronto) provided operational support over the course of the project. The team gratefully acknowledges the support of the Access Accelerated Trust Fund and the Centre for Global Health Research, Unity Health Toronto, Dalla Lana School of Public Health in making this work possible. The findings, interpretations, and conclusions expressed in this document are solely those of the authors and do not necessarily reflect the views of the Board of Executive Directors of the World Bank or the govern- ments they represent. Any errors are the sole responsibility of the authors/HLI team. xx COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E Overview Noncommunicable diseases (NCDs) are among the by the World Bank to support the HLI—a collabo- major health and development challenges of our time. rative effort between the World Bank, the Universi- Every year, about 41 million people die due to NCDs. ty of Toronto, and key academic and development This makes up about 74 percent of all deaths globally, partners including the Harvard University and the the majority of which are in low- and middle-income University of Washington. The HLI presents one of countries (LMICs).1 Countless more people live with a growing set of efforts to increase the urgency of NCDs every day. Yet, NCDs are largely treatable and policy response to NCDs across the world. preventable. The risk of developing NCDs and deaths from them can both be lowered with appropriate at- Chapter 1. Enhancing Human Capital and tention to prevention and treatment. Boosting Productivity by Tackling Non- However, weak health systems and limited ac- Communicable Diseases: Results of a cess to affordable care and information, especially in Research Initiative LMICs, contribute to lapses in seeking and receiving appropriate and timely care. While individual respon- Chapter 1 summarizes the proceedings of a technical sibility plays an important role in health maintenance, workshop as part of the project, ‘Non-Communicable addressing NCDs requires a concerted effort on the Diseases and Human Capital Research Initiative’. The part of societies as a whole to invest in better health and project, which was the precursor to the HLI, was led by well-being of individuals across the life course, that is, the World Bank Human Capital Project in partnership in healthy longevity. Reduced NCDs can boost human with the University of Toronto. It focused on the im- capital especially when coupled with education, skills pact of NCDs on the global economic burden of dis- development, and labor market opportunities. ease and selected aspects of human capital such as ed- Strengthening and protecting human capital is ucational outcomes, adult survival, and productivity. essential for countries’ overall welfare and to build The main messages from this research are as follows: inclusive, secure societies. Health, at both the in- dividual and societal levels, is a key contributor to • Rising NCD burdens threaten countries’ human cap- human well-being and a key objective of the devel- ital, a critical determinant of their economic success. opment process, over and above its impact on out- comes such as productivity and growth. Prioritizing • By tackling NCDs with proven, cost-effective in- the health and well-being of individuals throughout terventions, countries can reap substantial eco- their lives, from cradle to old age, not only improves nomic benefits, with some quick wins (within health outcomes but also enhances other aspects of 1–5 years). Gains will come mainly from higher human capital—education, skills, and ability to par- effective labor supply and improved productivity. ticipate in the labor market. However, the links between human capital and • Tackling NCDs in children and their caregivers better health through a reduced burden of NCDs may drive longer-term gains in human capital across the life course—from childhood to adult- and productivity by improving children’s educa- hood—need to be more explicitly explored and artic- tional outcomes. ulated. Filling this analytic gap is a fundamental mo- tivation for the Healthy Longevity Initiative (HLI). • Firms across multiple industries could reap pro- This compendium is a compilation of 18 chap- ductivity benefits from action on NCDs. ters, each exploring a different but related topic in the nexus of NCDs, human capital, and productivi- • Strong, country-owned measurement systems are ty. It is based on a series of analytical work taken up crucial for NCD success and human capital benefits. 1 WHO (World Health Organization). 2023. “Fact Sheet: Noncommunicable Diseases.” https://www.who.int/news-room/fact-sheets/detail/ noncommunicable-diseases. COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E OVERVIEW xxi The evidence presented supports ramping up invest- (LTC), which can lead to budget deficits and crowd- ments in NCD prevention and control that have the ing out of private sector investment. On the revenue potential for countries to bolster their human cap- side, to the extent that NCDs compromise work- ital and higher productivity in relatively short time force participation or productivity at work (from frames. Moreover, the growing appreciation of the higher absenteeism), they will reduce revenues importance of human capital creates a window of op- from labor income and corporate tax revenues. portunity for collaboration between the World Bank, the World Health Organization (WHO), and other • The human capital channel. Growing evidence partners to help countries enhance their human cap- suggests that the negative impacts of NCDs on ital by tackling NCDs. These themes are further de- human capital occur across the life course. NCDs veloped across the chapters in the compendium. compromise both the entire trajectory of human capital and returns to human capital: its forma- Chapter 2. Towards a Framework for tion and ongoing accumulation, its deployment Impact Pathways between NCDs, Human in terms of both duration and productive use, Capital and Healthy Longevity, Economic and its protection and preservation in the face of and Wellbeing Outcomes shocks and late life decline. This chapter lays out the pathways between NCDs, The human capital channel can be further delineated human capital, and the end outcomes of healthy lon- across the life course (figure 1) specifically as follows: gevity, inclusive growth, and well-being. It summariz- es the substantial literature on the impacts of health • In early life, there is strong evidence that NCDs and longevity on growth, using macroeconomic, and poor nutrition compromise cognitive foun- growth accounting, and microeconomic approaches. dations, school attendance, and learning out- Research shows that human capital accounts for comes, all affecting labor market outcomes in almost two-thirds of global wealth and on average adult life. There are also immediate effects from accounts for an increasing share of national wealth the growing burden of childhood NCD mortality as countries grow richer. While there is debate on and morbidity. the extent and causality, analyses of macro-level impacts of NCDs point to a negative relationship • During the stage of human capital deployment, between NCD prevalence and economic growth NCDs can compromise returns to human capi- and conversely a positive contribution to growth tal and its further accumulation, and accelerate from implementing a basic package of interventions depreciation through several mechanisms: (a) for the ‘big four’ NCDs (cardiovascular diseases total loss of labor supply as a result of premature [CVDs], diabetes, respiratory diseases, and cancers) mortality (over 40 percent of NCD deaths occur and mental health conditions. Moreover, the impact before age 70); (b) reduced returns to human of increased longevity on human capital accumu- capital from NCD-based morbidity and disabil- lation and growth differs according to the stage of ity, for example, because of unemployment, early demographic transition. There are three channels withdrawal, absenteeism, or ‘presenteeism’ (being through which NCDs affect inclusive growth: at work but less productive); and (c) failure to invest in further accumulation of human capital • The macroeconomic/savings channel. Despite over adult life, for example, due to reduction in much debate and the challenges related to ac- on-the-job learning and behavioral impacts on counting for non-health factors such as institu- subsequent efforts to accumulate human capital. tions and geography, a growing body of empirical work, including from developing countries, sup- • In the later stages of life, NCDs accelerate the ports a negative relationship between NCD prev- depreciation of human capital, and the NCDs of alence and economic growth. older parents contribute to the underutilization of human capital of their adult children in the labor • The fiscal channel. NCDs also have clear fiscal im- market, particularly of women, due to care de- pacts on the expenditure and revenue sides. High- mands. There may be a ripple effect also to grand- er NCD incidence increases the need for public children to the extent that NCDs in older ages con- spending on health care services and long-term care strain the caring contributions of grandparents. xxii OVERVIEW COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E FIGURE 1 Impact Channels between NCDs and Inclusive Growth Source: Original figure for this publication. Note: HC = Human capital. Health outcomes are also affected by key di- dementias. Moreover, the burden of informal care mensions of equity such as poverty or wealth, gen- for those with serious NCDs falls disproportionately der, and overall well-being. For example, evidence on women—almost 50 percent of women outside the points to a higher incidence of cancer, CVD, and labor market in middle-income countries (MICs) multiple NCDs in households with lower socio- and over 33 percent in low-income countries (LICs) economic status (SES) in LMICs. Such households cite unpaid care duties as the main reason for drop- have a significantly higher prevalence of NCD risk ping out of the labor markets. Based on data from factors such as tobacco and alcohol use, and poor- 53 countries across income levels, the International er nutrition compared to households with higher Labour Organization (ILO) estimates that women’s SES. At the same time, differences in availability unpaid care work would amount to 6.6 percent of and coverage of financial protection and the cumu- global gross domestic product. These themes are ex- lative effects of catastrophic health spending over plored further in chapters 8–11 in the compendium. the life course contribute to variations in outcomes. Consequently, policy responses to prevent and A growing body of evidence that uses the WHO’s control NCDs also need to take a life course approach. multidimensional quality-of-life measure also finds Some of the major policy areas of interest include nu- negative associations between a range of NCDs and trition, public health, and health care services, where quality of life, and that better health is strongly as- the economic and social returns to investments are sociated with higher levels of happiness. high, and affordable packages of essential NCD in- While NCDs dominate the disease burden glob- terventions are available. While cost-effectiveness is a ally for both genders (accounting for two-thirds of key consideration for developing countries in priori- female deaths annually), there are differences in the tizing within limited budgets, it is also important to gender-specific patterns of NCD prevalence and risk integrate other considerations such as equity and fi- factors. For example, male smoking rates (and tobac- nancial protection, as well as feasibility of implemen- co-attributable deaths) are substantially higher than tation given the diverse states of health systems. Out- those in women globally, while women dispropor- side of health, attention to taxation policies to reduce tionately suffer and die from Alzheimer’s and other risky behaviors (most notably tobacco use) while rais- COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E OVERVIEW xxiii ing revenue, policies and interventions that improve lic of Korea. Second, the economic value of avoidable road safety, and expanding coverage of social security mortality in a given year is defined as the proportion of systems hold promise as policies to mitigate and man- annual income one is willing to forego to live that year age NCDs and their social and economic impact. at the frontier survival probabilities. Then empirical estimates of the value per statistical life are calculated Chapter 3. The Economic Value of accounting for the effect of large risk changes by using Avoidable Mortality a logarithmic function to model a nonlinear trade-off relationship between income and risk reduction. This chapter addresses an important question— Figure 2 presents the proportion of avoidable what is the benefit of reducing mortality from mortality globally for 2019. Results show that 69 per- NCDs? It estimates the economic value of avoidable cent, or 40 million deaths globally, were avoidable. mortality by world region, sex, and age, between HICs had the lowest (42 percent) and Sub-Saharan 2000 and 2019, with projection to 2050. The chapter Africa had the highest (91 percent) levels of avoid- introduces two novel methods for defining and es- able deaths (in part due to persisting high and avoid- timating the economic value of avoidable mortality. able burdens from malaria and other infections in the First, a frontier approach is used to compute continent). Globally, more avoidable deaths occurred avoidable mortality by identifying the lowest observed in older adults than children, adolescents, or young- or projected mortality rate. This is established as the er adults. The economic value of avoidable mortality frontier for each age group and year, with avoidable globally in 2019 was 23 percent of the annual income, mortality defined as the difference between the fron- with China having the lowest value at 19 percent and tier and current mortality levels for each country by Sub-Saharan Africa having the highest value at 34 age, sex, and year. Importantly, the frontier lowest percent. The economic value is approximately 37 rates are seen in many countries and not just in select percent higher for males than females, reflecting the high-income countries (HICs) like Japan and Repub- higher avoidable mortality rates of males. FIGURE 2 Avoidable Mortality as a Percentage of Total Mortality, 2019 Source: Original figure for the HLI. These findings have implications for policy and ments worldwide engage in policy dialogues on how planning. They highlight the economic gains from and how much to invest in improving population reducing avoidable mortality in adults. As govern- health, this analysis provides supportive evidence xxiv OVERVIEW COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E on the high economic value placed on improving mates for 2000–2019 for an exhaustive set of 31 causes health, even when considering resource constraints. of death including major NCDs and injuries, the 2022 World Population Prospects population estimates, Chapter 4. The Economic Value Associated and the World Bank’s World Development Indicators, with Avoidable Mortality: A Systematic the chapter first quantifies avoidable mortality, that Assessment by Cause of Death across is, the difference between lowest-achieved mortality World Regions frontiers (the 10th percentile of age-specific mortali- ty rates) and projected mortality trajectories, for each Chapter 4 builds on the analysis in chapter 3 by es- cause of death, for 2000–2050. The value of a statistical timating the economic value of avoidable mortality life approaches is applied to assign economic values to by major disease groups including key NCDs and the estimates of avoidable mortality by cause of death, injuries for six large world regions: China, India, region, and calendar year. These economic values cap- HICs, Eurasia and the Mediterranean, Latin Ameri- ture the percentage of income an individual would be ca and the Caribbean, and Sub-Saharan Africa. willing to forego to live one year under the lowest pos- Using data from the WHO’s Global Health Esti- sible mortality rate for a given cause of death. FIGURE 3 Economic Values of Avoidable Mortality (as % of Annual Income) by Select Causes of Death, Region, and Sex for 2000, 2019, and 2050 Source: Original figure for the HLI. COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E OVERVIEW xxv Results show that the economic implications of The chapter also presents data on the country-level controlling NCDs and injuries would be substantial, performance on mortality decline. A further analy- particularly for CVDs, cancers, and injuries, with sis of the rates of progress in cause-specific mortal- some variations depending on the region and sex. ity, by country income level and age, from 2000 to The economic value associated with CVD avoidable 2019 using average annual rates of reduction (AAR- mortality would be universally large for both fe- R)2 shows the following: males and males, with values spanning 2–8 percent of annual income (for 2019). For cancers, it would • While childhood-cluster diseases (including be large (for both males and females) in HICs and whooping cough, diphtheria, measles, and teta- China, while for injuries, it would be substantial for nus) showed the fastest rate of decline in mortal- males in Latin America and the Caribbean (inten- ity in upper-middle-income countries (AARR of tional injuries) and males in Sub-Saharan Africa 9 percent for ages 0–14 years), progress on NCDs (unintentional injuries). was much slower, with the fastest decline in re- With rapidly aging populations, national health spiratory diseases (AARR of 5 percent) among systems must set difficult priorities toward improv- ages 50–69. ing the healthy longevity of their populations. This chapter provides a systematic monetary assessment • The slowest progress was in tobacco-attributable of the economic value associated with elevated cancers and breast cancer among those of ages 70 NCD and injury mortality, globally and regionally, and older in LICs. and derives an economic metric directly compara- ble to annual incomes (for example, gross national • Roughly 13.5 million deaths due to all caus- income), which can help countries in priority set- es could have been avoided (of which about 34 ting for the health sector as well as other sectors. percent would have been under the age of 70) if efforts to address these had yielded similar rates Chapter 5. Rates of Progress in Mortality of progress as observed for HIV/AIDS and tuber- Decline, 2000–2019 culosis during the same time period. This chapter summarizes the data on global mor- Finally, the chapter assesses the cost associated with tality decline and assesses the monetary cost of sav- saving a life using the concept of ‘critical income’, ing a life at different ages. Using interrupted time defined as ‘the income needed to achieve 80 per- series, it analyzes the change in mortality rates from cent of the global maximum life expectancy’. The 2000–09 to 2010–19, by age and sex, and by country analysis shows that the critical income for child income level. Results show the following: survival fell from US$1,452 to US$800 (at 2017 purchasing power parity [PPP]) between 1992 and • Between 2000 and 2010, mortality rates declined 2017, representing a decline of 46 percent. Howev- significantly across all age groups irrespective of er, it remained relatively constant for adults of ages country income level, except among older men of 15–49 but increased by 26 percent (from US$914 to ages 70–84 in LMICs. US$1,180) for ages 50–69. At the country level, the critical income needed to save a life was the lowest • The rate of mortality decline in the first half of for LICs and in Sub-Saharan Africa. the study period (2000–09) was significantly Technological change and policy attention faster for almost all country income levels than through efforts such as the Millennium Develop- during the latter half, that is, 2010–19 for age co- ment Goals have played an important role in im- horts under 70. proving survival among younger cohorts and re- ducing the cost of saving a life. Similar attention to • For ages 70–84, the rate of decline was signifi- addressing NCDs can yield further gains in reduc- cantly faster during 2010–19 for men in LICs and ing mortality and saving lives. This would require for both men and women in MICs. investments in research and development and glob- al public goods, paired with expanded service deliv- 2 Calculated using standard WHO methodology. xxvi OVERVIEW COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E ery, to ‘bend the cost curve’ for NCDs. can operate independently of other services, health system resources, and other structural factors. Some Chapter 6. Behavioral Aspects of important tools include the following: Healthy Longevity • Small financial incentives. Financial incentives Chapter 6 examines the role of behavioral science seem to promote healthy behaviors in the short as a policy tool for reducing NCDs. Addressing term, but effects disappear when the incentives NCDs to achieve healthy longevity for an aging are removed. Combinations with other behavior- population has become central to global health pol- al insights such as peer and framing effects tend icy goals. New policy tools are needed to effectively to render lasting results. and efficiently tackle health and lifestyle behaviors and habits linked to the development of NCD risk • Commitment contracts. They have also been factors. Behavioral science offers insights into psy- used to encourage other forward-looking be- chological barriers, mental models, biases, and oth- haviors such as weight loss or quitting smoking. er factors that influence decision-making and habit However, results have been mixed. formation. Applying these insights can support cur- rent policy efforts toward healthy longevity. • Saliency and vividness. They have also been The chapter develops a framework to clarify found effective in highlighting the risks of un- relationships between NCD occurrence, detection, healthy behaviors and inflecting those behaviors and management and behavioral determinants at such as through advertising the negative effects the individual, community, and health systems lev- of smoking. els, documenting frequently identified behavioral barriers at three key stages of patients’ NCD trajec- • Plan-making. This can also make adherence tories. It identifies policy lessons from the behav- more salient by focusing on the motivation fac- ioral science literature to address such barriers and, tors and identifying obstacles and planning to ad- together with other policies, reduce NCD incidence dress those that are under the individual’s control. and improve treatment effectiveness. The technique has been shown to reduce drink- More and more health systems, at least in MICs ing and smoking, improve healthy food con- and HICs, have started focusing on NCD prevention sumption, increase physical activity, and support and treatment. However, providing information may weight loss. not by itself be sufficient, because individuals may not act on that information due to a myriad of rea- • Reminders. Electronic reminders are increas- sons ranging from their own assessment of the per- ingly being used by the health system to address ceived need, distrust or dissatisfaction of the health poor medication adherence and appointment system, social norms and behaviors of reference attendance rates. They are especially helpful in groups, and peer pressure. These obstacles can be the short run for medication adherence and have exacerbated by system features such as complicated been shown to improve appointment attendance. processes that can increase people’s ‘hassle percep- tion’ about engaging with the health system, health • Choice architecture. Choice architecture refers insurance, and related institutions and processes. to the practice of influencing choice by ‘organiz- Behavioral interventions that aim to address ing the context in which people make decisions’ the factors behind risky behaviors appear to sup- (Mikic 2020). Removing temptation from one’s port the reduction of NCD lifestyle risk factors. Tra- choice set or making the healthy and unhealthy ditional policy tools based on the rational subject options equally convenient has shown promising model have been moderately effective in changing results for a range of behaviors, from improving unhealthy behaviors. Insights from the behavioral healthy eating habits to suicide prevention. sciences can complement these traditional tools, increasing their efficacy and impact. The behavior- While ongoing management of NCDs, mainly via al science toolkit is large. No single policy tool can medication adherence, is among the most cost-ef- address the complexity of behavioral barriers across fective strategies for NCD control, policy tools the whole journey toward healthy aging, nor should to influence individual health-seeking behavior these be considered stand-alone interventions that can support adherence by addressing barriers that COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E OVERVIEW xxvii currently prevent individuals from entering and es for low-income households, these are largely remaining in care, thereby improving the effective- compensated by medium- and long-term finan- ness of such interventions. cial benefits from better health even leaving aside the intrinsic value of better health. Chapter 7. Taxation of Harmful Products, including Tobacco, Alcohol, and Sugar- • Negative effects of taxation on employment have Sweetened Beverages (SSBs), and been largely exaggerated in industry-friendly re- Related Topics ports and mostly rest on narrow, single-market analyses. When general equilibrium approaches This chapter reviews the evidence on the effect that are considered, it is found that taxes do not de- changes in prices and incomes have on the consump- stroy jobs. Often, they create jobs as fiscal reve- tion of tobacco, alcohol, and SSBs, as a way of assess- nues are spent on labor-intensive services. Simi- ing the consumption impact of raising excise taxes. larly, the effect of tobacco taxation on illicit trade Key findings of the review include the following: is overemphasized. Illicit trade depends on insti- tutional weaknesses and regulatory deficiencies • The large and growing disease burden associated rather than the level of tobacco taxes. with health-damaging products poses challenges for health systems and sustainable development. In essence, reducing consumption of tobacco, alco- Reducing the consumption of such products is hol, and SSBs and the burden of disease associated key for a healthy longevity agenda. with them reduces health care costs, increases the social return on human capital, reduces human suf- • Taxation is one of most cost-effective tools to fering, and lowers poverty, among other benefits. decrease consumption of tobacco and alcohol Taxation is one of the most cost-effective ways to as well as SSBs. It is particularly effective in in- reduce the consumption of harmful products and fluencing early life consumption (for example, increase social welfare. during adolescence or youth) that may have a permanent effect. Chapter 8. Gender Gaps in Health and Well-Being of Older Adults: A Review of • Firms usually control the rate at which taxes are the Burden of NCDs and Barriers to Health passed to prices, and there is consistent evidence Care for Women and Men that in most cases taxes are fully passed to prices (and often over-shifted, that is, the pass through Chapter 8 examines the gender gaps in the health to prices is greater than the taxes). Hence, taxes and well-being of older adults, focusing on the bar- are effective at decreasing consumption. riers in access to health care and LTC. It examines gender gaps in mortality, prevalence, and disabili- • Excise taxes on harmful products increase eco- ty-adjusted life years (DALYs) for major NCDs and nomic efficiency, as they correct consump- mental health using data from the Global Burden tion-related externalities (that is, market prices of Disease 2019 study for older adults in three age without taxes do not reflect all costs borne by groups: 45–59, 60–79, and 80 and above, as well as society) and internalities (that is, individuals’ country income data from the World Development ignoring or not correctly considering harmful Indicators. It also reviews the barriers to demand health effects to themselves when they consume for health care for women (and men) particularly in these products). Hence, taxation increases indi- LMICs and the challenges in providing health care vidual and social welfare. and LTC from a gendered perspective, with special attention to caregivers, particularly women who • When considering the costs to households of provide informal care to older adults. health care, foregone income, and premature mor- Data analysis for CVD and diabetes across tality linked to the consumption of health-dam- countries shows that the burden of disease is similar aging products, taxes on these products are found for women and men ages 45–59, with some excep- to be progressive, as they save valuable resources tions. For CVD, the burden of disease is higher for in low-income households. Though taxation may men in upper-middle and high income countries for have short-term negative financial consequenc- ages 60–79, whereas for ages 80 and above, women xxviii OVERVIEW COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E bear the greater burden. For diabetes, gender gaps among women of ages 4559 compared to men in the begin to appear for ages 60–79, particularly LICs, same age group. However, for older ages, the gender and for 80 years and above, there is once again a gap in cancer prevalence narrows in LICs but grows greater burden of disease for women. Cancers pres- larger in upper-middle and high income countries, ent a more mixed picture with a higher prevalence particularly affecting men of ages 60–79 (figure 4). FIGURE 4 Prevalence of Cancers among Older Adults (%) by Sex and Gross National Income (GNI) Source: Institute for Health Metrics and Evaluation (IHME) Global Burden of Disease (2020), World Development Indicators (2020). Another key area of concern is mental health. daily activities, grows. Evidence from aging econo- Prevalence of depressive disorders is consistent- mies, primarily high- or middle-income Organisa- ly higher for women compared to men for all age tion for Economic Co-operation and Development groups 45 and above, but mortality due to self-harm (OECD) countries, shows considerable diversity in is higher among men for all age groups. The choice the quality and accessibility of LTC services, influ- of method to inflict self-harm and not seeking help enced by governmental policies, social norms, and due to the stigma associated with mental illnesses resource availability. These arrangements, encom- are linked to these outcomes. passing family care, public institutions, private res- While aging signals the onset of NCDs for both idences, and assisted-living environments, reflect women and men, their health outcomes vary. Sever- different care needs. al studies highlight higher male mortality at earlier Delivery of LTC services through the public or ages due to heart disease or diabetes, particularly private sector has a significant effect on the cost of in HICs and MICs. On average, life expectancy at care. HICs often allocate health care costs through birth for women is about 5 years higher than that public funds, while LICs rely more on out-of-pock- for men globally. However, in MICs the difference et (OOP) payments, exacerbating affordability is generally about 8–10 years. Although women challenges, particularly for women especially in the in these countries exhibit similar patterns of NCD absence of robust social protection systems. For ex- prevalence, they tend to live longer and often with a ample, despite efforts to provide social protection lower quality of life. Women also go through meno- for LTC across HICs, OOP expenses remain com- pause as they age, which takes a physical and mental mon, especially in countries that have lower nation- toll and is linked to an increased risk of developing al incomes and less comprehensive coverage. NCDs. Moreover, older women, especially in LICs While informal, family-based care may still with weaker health systems and limited financial be the norm in many LICs, not all families are able protections, are more likely than men to have great- to provide appropriate support, and medical costs er financial vulnerability and often limited mobil- remain a barrier to care for the aging, particularly ity due to social norms—a direct consequence of women. Such unmet LTC needs can have a cata- gender gaps over the life course that affect women’s strophic effect on the quality of life and chances of patterns of labor force participation, income oppor- survival of older adults. tunities, and decision-making power. The increase in demand for LTC has implica- As populations age, the demand for LTC, which tions for women’s employment as well. Evidence includes health care and assistance for performing from aging economies has shown that women will COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E OVERVIEW xxix often reduce their working hours, drop out of the their older parents. Indonesia has the largest pro- labor market temporarily, or retire early to provide portion of women who provide LTC to their par- elder care (explored further in chapters 9–11). The ents, while Poland has the smallest proportion of need for LTC also creates employment opportuni- women providing such care. In addition, women in ties within this labor-intensive sector comprising Indonesia spend a significantly longer time caring both formal and informal caregivers. Here again for their parents. On the other hand, in Colombia women play a significant role, albeit associated with men spend a slightly longer time than women on migratory labor, low wages, and stressful conditions care responsibilities for their parents. contributing to high turnover rates, notably in re- In Colombia and Poland, both men and wom- gions like Europe. en who provide parental care are significantly less The chapter points out key policy and pro- likely to be employed. In Indonesia, while there is grammatic areas for intervention to reduce gender no overall change in the extensive range of labor gaps in health care and LTC for aging populations supply for either gender, there is a significant de- and caregivers. cline in formal employment among both men and women who provide care. However, while men Chapter 9. Gendered Responsibilities, transition from formal to informal employment Elderly Care, and Labor Supply: Evidence when they provide care for their parents, women from Four Middle-Income Countries are more likely to drop out of the workforce. Overall, the analysis finds that providing care Chapter 9 explores the relationship between provid- to older parents is associated with significant reduc- ing care for older parents or parents-in-law and la- tion in the probability of employment, weekly hours bor supply among middle-aged men and women of worked, and annual earnings. This decline is signifi- ages 40–59 in four countries: Colombia, Indonesia, cantly larger among women than men and among Poland, and Arab Republic of Egypt. intensive caregivers, those who provide care for Much of the responsibility for informal care more than 10 hours per week. The analysis shows is often shouldered by women. Time-use surveys that LTC is associated with a 6 percentage point from 64 countries show that 76 percent of the total gender employment gap and a 32 percent gender time spent on unpaid care work is by women. This earnings gap to women’s disadvantage. is three times more than that of men. Caregivers are There is also substantial heterogeneity across also likely to be in their prime working years, which the four countries, with the largest significant de- can affect their participation in the economy and clines in employment in Poland and Colombia for income potential. According to one global estimate, both women and men. In Indonesia, men who pro- 647 million working-age individuals are outside of vide care experience a large decline in work hours the labor force due to family responsibilities and and annual earnings, driven by their transition to 606 million (94 percent) of them are women. casual employment, while in Egypt, labor supply for Across the four countries studied in this chap- men and women was not affected by the presence of ter, an examination of socioeconomic characteris- disabled parents or parents-in-law in the household. tics shows that less-educated women are more likely Although the evidence presented in this chap- to provide care to an elderly parent, likely due to the ter is not causal, these results can contribute to fu- greater opportunity cost of leaving high-paying jobs ture policy dialogue to reduce the gender imbalance by other family members. Older women are more in LTC burdens, particularly in rapidly aging re- likely to provide care to their parents, but this rela- gions where the gender gap in labor market partic- tionship is not significant for men. Being married ipation is large. reduces the likelihood of providing parental care for both men and women. However, the number of Chapter 10. Health Care and LTC Needs in a children ages 0–18 is associated with an increase in Context of Rapid Population Aging the propensity of men providing parental care. While in all four countries, men are more This chapter explores key challenges in health care likely than women to live with their older parents, and LTC as populations age and provides examples women bear the majority of the responsibility to of how countries are responding to them. It was provide care for their parents. In general, women originally developed as a background report for the are at least twice as likely as men to provide care to Independent Evaluation Group evaluation ‘World xxx OVERVIEW COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E Bank Support to Aging Countries’ and adapted for Chapter 11. Demand for and Supply of the HLI. The chapter focuses on developing countries Long-Term Care for Older Persons in Low- that are aging fast and where anticipation and action and Middle-Income Countries are especially important. It documents health care experiences aimed at increasing coverage, adapting Chapter 11 examines the demand for and the cur- systems to new epidemiological profiles, focusing on rent landscape of supply of LTC services in six the prevention of chronic diseases, and promoting MICs: China, India, Malaysia, Mexico, Poland, and healthy aging and discusses mental health policies, a Romania. With rapid population aging, especially comparatively overlooked area until recently. in LMICs, over the next few decades, the currently The evidence presented highlights the slow predominant model of family-based elder care will and uneven progress in addressing LTC needs of not be sufficient to meet the escalating demand for functionally dependent older adults, particularly LTC (which covers both medical care and nonmed- in developing countries. The scope and depth of ical services and support for older adults). the policies vary, and in some cases, little is known The demand-side assessment is based on the about their actual implementation. The role of the analysis of individual-level data sets that are part government in providing and regulating LTC ser- of the International Network of Health and Retire- vices becomes crucial given changes in social norms ment Studies focusing on the measurement of activ- and the lack of affordable and high-quality private ities of daily living (ADL) and instrumental activi- services in many settings. ties of daily life (IADL) among those 65 years and Addressing the gaps in LTC policies and pro- older. The supply-side evaluation is a synthesis of grams is critical for meeting the growing care needs findings from a review of relevant literature for the of older persons. There is a need for holistic strate- past 22 years (2000–2022), which also covers oth- gies that focus on strengthening health care and LTC er countries. It encompasses Asia (where data are systems. This requires increasing coordination and available), Central and South America, Southern integration between the social care and health care Europe and Middle East, and Sub-Saharan Africa. sectors, given the paramount importance of such The demand-side analysis finds differences integration both to achieve a continuum of care across countries in terms of difficulty performing for the older person and to obtain efficiency gains. ADLs and IADLs. Individuals’ SES and gender are Strengthening primary and community care is key correlated with difficulty in performing ADLs or to achieving these goals while containing costs. IADLs. Overall lower education is associated with There is also a need for renewed focus on build- greater difficulty in performing ADLs with coun- ing capacity of human resources that work with try variations. For example, in India, there is a 5–6 older persons, both in the health care and the LTC percentage point difference between those with sec- sectors. This is a priority, given that skilled human ondary and lower or no education, while in Poland resources are an important determinant of the qual- and Romania, the probability of having an ADL ity of care that older people receive. difficulty is about 10 percentage points lower for Besides being accessible and affordable, LTC ser- individuals with secondary education compared to vices should be person-centered encouraging, when- those with a primary education. ever possible, the provision of home care services In China, India, Malaysia, and Mexico, women (aging in place). A person-centered model is crucial were about 7–9 percentage points more likely than for adequate management of chronic diseases, given men to experience at least one ADL difficulty. In the high prevalence of comorbidities. For example, contrast, in Poland, men were 6 percentage points aging in place has been shown to provide the best re- more likely than women to experience difficulty sults in terms of older people’s mental and physical with an ADL. health and to be cost-effective for governments. Age stands out as one of the most important Finally, to solve the challenges brought by pop- predictors of needing help with ADL or IADL. For ulation aging, it is also important to take a life course example, those of ages 80 and over were more likely approach. How people age is, to a large extent, deter- to experience difficulty with IADL in all countries mined by their health earlier in life and the choices compared to those of ages 65–79 years. they made when young. The range of policies should However, the availability of formal LTC is lim- also promote healthy lifestyles, like physical activity ited, and familial (or informal) care remains the and healthy eating, throughout the entire life course. primary form of care for older persons, albeit with COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E OVERVIEW xxxi variations across regions and countries. In some English. The majority (96 percent) of the studies countries, particularly in the Middle East, cultur- were conducted in HICs with 63 percent from Eu- al norms emphasize familial care, particularly by rope and Central Asia. women. Often there is a lack of regulatory frame- High body-mass index (BMI) was the most works and limited public funding. While private frequently reported exposure (reported by 46 per- sector LTC services are increasing, these are gen- cent of the studies), while income was the most erally expensive and clustered around urban areas. studied outcome (reported by 33 percent of stud- The chapter recommends designing a hybrid ies). Of the studies, 77 percent reported significant system of informal and formal LTC services for (p < 0.05) adverse associations between the expo- LMICs; engaging the private sector in developing sures and outcomes. LTC services, markets, and delivery systems in All of the studies included in this review that LMICs; strengthening government stewardship for looked at plausible causal relationships between clear guidance on rules of engagement, quality as- NCD risk factors and labor market outcomes were surance, and regulatory capacity; establishing a sys- from HICs and UMICs (USA, northern European tematic approach to public LTC financing, ideally countries, and Republic of Korea). The main finding following a broad-based social insurance model to of these studies is that individuals with obesity, dia- improve affordability of services; and, drawing on betes, hypertension, depressive disorders, excessive the experiences of HICs, considering multipronged alcohol use, and cigarette use are more likely to have strategies to build and strengthen the LTC work- lower rates of employment, lower income, and high- force and to support family caregivers in LMICs. er rates of sickness absence and disability pension. There is no ideal LTC system that works in all countries, but policy makers in LMICs can learn Chapter 13. Productive Longevity: from the LTC systems in HICs. The WHO recom- What Can Work in Low- and Middle- mends an integrated continuum of LTC framework Income Countries? that supports person-centered, primary health care-driven, and integrated delivery systems. The This chapter focuses on labor market characteristics care continuum should also include palliative care and related policies that can address the key con- and end-of-life care. At present, such integrated straints to productive longevity in LMICs. By 2050, LTC service provision and delivery models large- one in six persons globally will be at least 65 years ly remain a concept in most LMICs. Nevertheless, of age, with nearly four out of five living in LMICs. the growing knowledge base on international LTC Changing demographics with larger cohorts of older research and cross-country learning can help coun- populations will likely require higher public and pri- tries innovate and adapt best practices. vate expenditures for health care and LTC services. There is a need for policies that can sustain welfare Chapter 12. Exploring the Labour levels and ensure that welfare is equitably distributed Market Outcomes of the Risk Factors for across generations and socioeconomic groups. Non-communicable Diseases: Labor market policies that help extend produc- A Systematic Review tive working lives and increase the labor force par- ticipation and productivity of older workers—‘pro- Chapter 12 presents a review of the associations be- ductive longevity’—are part of this important tween labor market outcomes and major risk factors agenda. Several stylized facts about older workers for NCDs (smoking and heavy alcohol consump- have a bearing on policy decisions: tion), key metabolic changes due to the risk factors (overweight and obesity, hypertension, and type 2 • Old-age labor force participation is generally low- diabetes), and major depressive disorders. It also ex- er, the higher a country’s average income level. In amines gender differences across these dimensions. the average LIC, more than half of the 65+ popu- The chapter is based on a systematic review of lation is active in the labor market. cohort and longitudinal studies before July 2022 to establish causality between exposures and out- • Gender gaps among mature workers appear more comes in peer-reviewed literature. The search pa- important for MICs than HICs, with greater em- rameters for resulted in 109 studies that were eli- ployment gaps and women’s earlier exit from the gible for the review. All studies were published in labor force. In LICs, gender gaps are smaller—the xxxii OVERVIEW COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E poorer the country, the higher the chances that ipation of older workers has increased in the past women will also work into old age. few decades. However, the COVID-19 pandemic has resulted in a significant withdrawal of older • Lacking social protection, many mature workers workers from the labor market. in LMICs continue working to secure their live- lihoods. • The changing nature of work may be contributing to more age-friendly jobs. Technology is enabling • Better-off workers are also considering continu- automation of many tasks, but it also raises the ing working later in life, whether to top up re- demand for skills that are (still) complementary tirement income or for other benefits, such as to technology, such as creativity, problem solving, social connections. and socio-emotional competencies. • Old-age labor force participation and education Policies that support older adults remaining healthy, interact differently across countries. While in skilled, and economically active have positive im- HICs, people with higher education may remain pacts across generations. The experience from HICs in the work force, in LICs, those with higher edu- and some MICs shows that the cost of inadequate cation are more likely to retire early because they solutions can be substantial. Unsustainable pen- are less likely to be financially constrained. sion systems or insufficient attention to skills de- velopment for older generations have long-term • Globally, self-employment is more common repercussions. Research on HICs also shows that among older workers than other age groups, with opportunities for voluntary part-time paid work or the highest rate of self-employment among the volunteering activities can contribute to strength- 55–64 age group. ening the physical and mental health at older ages, including for those over 80 years. LICs with less-de- • In advanced economies, the labor market partic- veloped systems can learn from these experiences. FIGURE 5 Labor Force Participation Rate for Ages 65+, by Country Income Group Source: Estimates based on Staudinger et al. 20163. Note: Refers to 2013 data. 3 Staudinger, U.M., et al. 2016. A Global View on the Effects of Work on Health in Later Life. Gerontologist, 56(S2): 281–292. COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E OVERVIEW xxxiii Given high informality and weak enforcement pecially as the evidence is often mixed, pointing to in LMICs, reforms in the formal sector policies, reg- the challenges associated with this policy agenda. ulations, or taxation will only reach a small segment of workers compared to HICs. In such dual labor Chapter 14. Adequacy of Pensions and markets, policy reforms will also need to examine Access to Health Care: Maintaining Human options to avert old-age poverty for those in infor- Capital during Old Age mal markets. This would need to center on strength- ening human capital with health interventions and Chapter 14 discusses the adequacy of pensions strategic, demand-led skills development; support- for health care in old age, particularly for LMICs. ing the businesses operated by mature workers; and Maintaining human capital during old age requires providing social safety nets. access to affordable quality health care as well as Even with constraints, countries can work to adequate pensions to provide financial coverage identify ‘win-win’ policies that benefit all genera- necessary to afford this care. Yet, there is a lack of tions. For example, providing LTC for the elderly adequate social protection in most countries, par- and childcare, thus alleviating the care burden, can ticularly for informal workers, who make up large stimulate labor market participation and the move proportions of working adults in LMICs. to more productive work in both older and younger An examination of OOP payments by households workers, in particular women. Facilitating access to globally shows that (a) households with the elderly have work, for example, with safer transportation systems, 1.3 percentage points higher OOP payments health ex- will have benefits for vulnerable groups in general. penditure compared to households without the elderly; Looking in detail at how gender, skills, location, and (b) lower-income households with the elderly have a 16 other factors affect productive longevity in different percent higher burden of OOP on average compared settings will be essential to develop adequate policies. to higher-income households with the elderly; and (c) This chapter presents a first attempt at high- lower-income households are more likely to be pushed lighting policy areas and mapping out the associated below the poverty line (US$3.2 per day) due to OOP evidence at an aggregate level. Productive longevity health spending (World Bank Atlas of Social Protection does not affect all groups similarly, and a more nu- Indicators of Resilience and Equity [ASPIRE] data). anced analysis is needed. More evidence is needed on While overall OOP payments are the highest what works for productive longevity. Most evidence in LMICs (figure 6), impoverishment due to OOP for what works is based on HICs, which have faced spending occurs even in countries where the entire aging pressures for a longer time. Even in these coun- population is officially covered by a health insurance tries, however, there is scope for further research, es- scheme or by national or subnational health services. FIGURE 6 OOP Expenditure (% of Current Health Expenditure) Source: WHO Data Portal. xxxiv OVERVIEW COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E Even though countries with universal quality trol Priorities Project, specifically, the third edition health care may be able to ensure a minimum stan- (DCP3), published between 2015 and 2018 (Jami- dard of living for the elderly at a lower benefit level son et al. 20184). This chapter takes a broader view (which may not always be sufficient), in countries of NCDs than the Countdown report and considers where the elderly face high OOP health expenses, any DCP3 intervention that can reduce NCD mor- pensions may need to be higher to achieve that tality, including surgical care and mental health care. same minimum living standard. Priority setting starts with an assessment of Determining what level of pensions is ‘ade- cost-effectiveness. The following three additional quate’, what constitutes a ‘minimum level of pro- criteria for identifying high-priority health sector tection’, and who should be eligible are the biggest interventions for NCDs are also considered: equity, design challenges today. Constrained fiscal space, financial risk protection, and implementation fea- multiple competing development objectives, and sibility. Data show considerable variation in inter- the growing number of older people due to de- vention cost-effectiveness across health systems and mographic aging generally means that developing underscore the importance of local analysis. countries would need to afford to provide either a Overall, the cost and impact of scaling up the lower benefit to a larger share of the elderly popu- package of recommended interventions to achieve lation or a higher benefit to a more limited share of 80 percent population coverage in all countries the elderly population. by 2050 or sooner are assessed. Implementing the While there can be different approaches to im- package of high-priority interventions fully could proving coverage, many countries, including LICs, avert up to 150 million deaths (or 2.2 billion DA- have introduced noncontributory mechanisms to LYs) by 2050, at an incremental cost of US$1.3 tril- cover the informal sector and the self-employed. lion. This translates to US$9,300 per death averted However, this also has its limitations. Fiscal con- or US$620 per DALY averted. straints will limit both the size of this kind of pen- These interventions represent a significant pro- sion and the income and age to be eligible to receive portion of what health systems already do to address the old-age benefit. Another major challenge is en- NCDs such as CVD primary prevention, diabetes suring that programs remain affordable over time, management, and mental health care. Roughly half given population aging and the potential for discre- of the interventions would be very cost-effective in tionary increases in the benefit amounts. nearly all settings. Some of the surgical and mental The chapter concludes that including the costs health interventions would be on the borderline of of access to quality health care in the definition of being cost-effective in most settings, whereas others ‘adequate’ pensions does not necessarily imply that (for example, appendicitis management and chronic increased pensions are preferrable to addressing the depression treatment) would be very cost-effective. deficiencies of the health system. It may just reflect For diabetes care, glycemic control by itself is not the additional financial needs pensioners will face cost-effective, but when implemented with CVD if access to universal health care (UHC) of good prevention, it would be cost-effective. There are also quality is not granted. Improving this access is, no several interventions ‘outside the health sector’ that doubt, a more efficient and permanent solution, but are all incredibly cost-beneficial, with tobacco con- it may take longer to implement. trol yielding the highest returns. However, most countries are nowhere near full Chapter 15. Priority Setting for NCD implementation of these interventions. While the Control and Health System Investments proposed package of interventions would be very cost-effective, the budgetary consequences would Chapter 15 focuses on prioritizing health interven- be considerable, with a larger ‘incremental cost’ as tions for NCD control. It builds on the work of the a share of the current health budget for LICs due to NCD Countdown Collaborators and shares a set of the need to build up currently weak health systems modified recommendations from the Disease Con- and underspending on health. The prioritization 4 Jamison, D.T., et al. 2018. “Universal Health Coverage and Intersectoral Action for Health: Key Messages from Disease Control Priorities , 3rd Edition.” The Lancet 391 (10125): 1108–20. COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E OVERVIEW xxxv of interventions also provides a framework for se- encompasses endocrine disorders beyond diabetes, quencing them over time, and intersectoral policies malnutrition (both under- and overnutrition), au- and health interventions that are very cost-effective toimmune diseases, chronic kidney disease, various and can be delivered through primary health care occupational hazards, and injuries, both accidental platforms by nonphysician health workers should and intentional. This definition strengthens the ar- be scaled up as early as possible. gument for a WoS approach, due to the importance There are several important limitations of the of non-health sectors in the causation of and actions analysis presented in this chapter including data against these diseases. availability, modelling assumptions about current A WoS approach is a comprehensive approach coverage of interventions, and the effect of the to planning, design, implementation, and monitor- COVID-19 pandemic on NCD incidence and mor- ing of policies and programs that not only includes tality as well as health system resources in the me- the ‘whole-of-government’ (WoG) but also non- dium term. state actors (for-profit and nonprofit), communities, Yet, the proposed set of recommendations households, and individuals. The WoS approach can help countries achieve significant reductions emphasizes engaging all branches of government in NCD mortality, helping them achieve the Sus- (the judiciary, the legislature, and the executive) tainable Development Goal (SDG) 3.4 target and and political leaders from opposition parties, as the post-SDG targets for NCDs. Implementing the HLI latter play a critical role in supporting or opposing package of interventions has the potential to help government initiatives and their relative success. countries reduce their avoidable mortality across all Similarly, non-state actors have a key role in age groups to about halfway to the observed mortal- preventing, managing, and controlling NCDs, with ity rates in the best-performing countries. In addi- the caveat that while there are strong benefits to pub- tion to reducing adult mortality, these interventions lic-private partnerships, bringing the private sector could improve nonfatal outcomes by reducing the on board requires a clear-eyed approach on the part incidence of disease overall and by reducing the se- of governments, bearing in mind the risks emerging verity of chronic illness, especially for mental and from the profit motive of commercial entities that substance use disorders. The recommendations diverge from public health interests (figure 7). could also help bring a necessary focus and fiscal The chapter highlights several examples of discipline to national NCD strategies and plans and innovative and mutually beneficial public-private be used as benchmarks for future World Bank lend- partnerships for the prevention and control of ing efforts to support NCD program development NCDs though they are limited in scale and scope. in lower-resource countries. It stresses the key role that the nonprofit private sector (nongovernmental and community-based Chapter 16. Control of Non-Communicable organizations) can play, given these organizations’ Diseases for Enhanced Human Capital: The ubiquity in LMICs, notably the poorest, and their Case for Whole-of-Society Action comparative advantage in community mobilization. Chapter 16 makes the case for a whole-of-society Given country experiences and lessons learned, the (WoS) approach to address NCDs, in view of the chapter presents 10 key recommendations (the 10 ‘I’s): two-way relationship of NCDs with certain eco- nomic sectors and makes recommendations for the 1. Involve key stakeholders from the outset, lever- way forward. aging any congruence of interests, to build effec- Building on the WHO’s 5 × 5 framework of tive alliances for change. NCDs, which considers five sets of diseases and five risk factors,5 the chapter proposes a broader defi- 2. Include non-state actors as appropriate and nition of NCDs, including ‘any disease or disabili- within a mutually agreeable policy, operational, ty that is not transmissible through infection’. This and regulatory framework. 5 The five diseases are CVDs, diabetes mellitus, chronic respiratory diseases, cancers, and mental health, while the five risk factors include tobacco use, harmful use of alcohol, unhealthy diets, physical inactivity, and air pollution. xxxvi OVERVIEW COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E FIGURE 7 Components of a WoS Approach Source: Original figure for this publication. 3. Identify clear roles and responsibilities for every 9. Interventions should be prioritized on bur- sector and actor through consensus and with a den-of-disease assessment, cost-effectiveness, clearly identified team lead. and societal value choices. 4. Institutionalize mechanisms for coordination 10. Indicators of success and failure need to be mon- and cooperation. itored continuously for effective decision-mak- ing, with course corrections as necessary. 5. Invest in human and other resources for better coordination and concertation. Chapter 17. HLI: A Performance Dashboard for Decision-Making in Low- and Middle- 6. Innovate continuously for successful and sus- Income Countries tained implementation of a WoS approach to NCDs. This chapter presents a framework for monitoring progress toward healthy longevity. A healthy lon- 7. Information sharing through knowledge-ex- gevity agenda offers an opportunity to promote and change platforms, such as clearing houses and monitor progress toward human capital accumula- South-South platforms, is essential. tion and preservation, health, and well-being across the life course. 8. Incorporate lessons learned from global NCD One way for countries to measure this progress experiences to rapidly integrate best practices is through performance dashboards. Performance and avoid repeating others’ mistakes. dashboards are data visualization tools that bring COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E OVERVIEW xxxvii together representations of indicators into a unified comparable, country-level indicators from interna- interface that allows stakeholders to monitor and tional institutions that can be mapped to the overar- assess performance and progress across selected ching HLI conceptual framework and across different dimensions. The healthy longevity dashboard is an stages of the life course. This approach supports the ongoing effort to develop and refine a suite of indica- selection of indicators that reflect a given country’s tors that bring together relevant data to measure and data infrastructure maturity while ensuring compa- monitor country progress toward healthy longevity. rable and consistent conceptual underpinnings. The chapter first offers a common framework Dashboard indicators are categorized as either for such a dashboard and presents three country contextual or measuring healthy longevity. Contex- case studies—Sierra Leone, India, and Colombia— tual indicators provide crucial information on basic that illustrate a range of possibilities for developing social, demographic, and economic characteristics a country-specific but internationally comparable of a country that enables policy makers and other dashboard that considers differing data and infor- dashboard users to correctly interpret the setting mation system contexts. from which other performance-related indicators The common framework takes a life course are drawn. These indicators are further classified approach and is grounded in the HLI conceptual into 10 broad domains: (a) life expectancy, (b) mor- framework (chapter 2). Indicators cover three key tality at different life stages (under five years, school actions and ten related domains that map the dash- age (ages 5–14), youths and young adults (ages board to the HLI framework allowing for a holistic 15–29), adults (ages 30–69), and older ages (ages and comprehensive view of population health, while 70 and above), (c) child health, (d) adult immuni- drawing attention to the multi-sectoral efforts need- zations, (e) risk factors, (f) access to care for NCDs ed to promote healthy longevity, inclusive growth, and NCD management, (g) education, (h) gender and well-being. equality, (i) labor force, and (j) social protection. The three countries selected represent different Once indicators are selected, a country’s per- contexts in terms of income level, epidemiological formance, in this case, India, is estimated compared and demographic profiles, and different data infra- to other countries, using percentile and Z scores. structure contexts. Sierra Leone’s dashboard illus- In both approaches, the study country is compared trates how countries with nascent data infrastruc- with other countries that fall under the same income ture can bring together available indicators based on strata as the study country and have a population key investments in healthy longevity and NCDs in of more than 7 million (or 0.1 percent of the world a meaningful way. India presents a case where data population). The dashboard for India highlights its are more readily available, including from interna- performance on HLI indicators in comparison to tional sources, driven by national policy objectives, similar LMICs and the world. For example, India and can be used for benchmarking. Colombia offers has a higher proportion of deaths that are attribut- an example of how countries can leverage more ex- able to NCDs and higher UHC coverage compared tensive data sets across multiple sectors. to the average of LMICs but lower than the global The chapter concludes with a discussion on average. Moreover, where gender-disaggregated data what actions are needed to assist country stake- are available, the dashboard shows that males score holders in producing, presenting, using, and insti- better than females on health indicators except in tutionalizing a healthy longevity dashboard tailored harmful alcohol use and prevalence of hypertension. to their unique demographic and epidemiological In addition to focusing on healthy longevity, context and data infrastructure. the novelty of the HLI dashboard lies in its focus on international comparisons. The current HLI Chapter 18. Assessing Human Capital, dashboard relies on data readily available in the Non-communicable Diseases, and Healthy public domain. As such, LMICs can replicate the Longevity in Low- and Middle-Income dashboard for their countries without the need for Countries: Healthy Longevity Dashboard additional data collection. However, relying only on and the Case for India readily available data restricts the range of indica- tors included in the dashboard. This only under- This chapter expands on the development of the HLI scores the need for countries to invest in building performance dashboard for LMICs using India as an up their essential data infrastructure to produce example. The dashboard comprises internationally performance information of sufficient granularity xxxviii OVERVIEW COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E to enable relevant decision-making at the appro- ing the drivers of healthy longevity, improving its priate governance level, with information of high measurement, and producing performance dash- quality and of relevant timeliness and granularity. boards of use to policy makers. Finally, further research is required in understand- COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E xxxix Abbreviations ACS Acute coronary syndrome HLI Healthy Longevity Initiative ADI Alzheimer’s Disease International HLM High Level Meeting ADL Activities of daily living HLMP Healthy Living Master Plan AOR Adjusted odds ratio HMIS Health Management and BCA Benefit-cost analysis Information System BMI Body mass index HNP Health, Nutrition and Population CD Communicable disease (sector of the World Bank) CDC Centers for Disease Control IADL Instrumental activities of daily living and Prevention IARD International Alliance for CEA Cost-effectiveness analysis Responsible Drinking CHARLS China Health and Retirement ICER Incremental cost-effectiveness ratio Longitudinal Study IDU Injection drug use CHE Catastrophic health expenditure IFBA International Food and CI Confidence interval Beverage Association CMPs Communicable, maternal, perinatal, IFC International Finance Corporation and nutritional conditions IHME Institute for Health Metrics COPD Chronic obstructive and Evaluation pulmonary disease ILO International Labour Organisation COVID-19 Coronavirus disease of 2019 INT$ International dollar CRVS Civil Registration and Vital Statistics ITS Interrupted time-series CVD Cardiovascular disease LAC Latin America and the Caribbean DALY Disability-adjusted life year (World Bank regional DCP3 Disease Control Priorities, 3rd Edition vice presidency) DPL Development Policy Lending LASI Longitudinal Aging Study in India DTP3 Diphtheria, tetanus, and pertussis LIC Low-income countries EAP East Asia, and the Pacific LMICs Low- and middle-income countries ECA Eastern Europe and Central Asia LTC Long-term care ECBA Extended cost-benefit analysis MAP Multi-sector Action Plan EU European Union MARS Malaysia Ageing and FCV Fragility, Conflict and Violence Retirement Survey FeEd Female education MCC Multiple chronic conditions GAP Global Action Program MCDA Multi-criteria decision analyses GBD Global Burden of Disease MDD Major depressive disorder GDP Gross domestic product MDGs Millennium Development Goals GDPPC Gross domestic product per capita MHAS Mexican Health and Aging Study GNI Gross national income MIC Middle-income countries HALE Health-adjusted life expectancy/ MNA/ Middle East and North Africa (World healthy life expectancy MENA Bank regional vice presidency) HC Human Capital MOH Ministry of Health HCI Human Capital Index MUP Minimum unit prices HCP Human Capital Project NCD Non-communicable disease HED Heavy episodic drinking NGO Non-governmental organization HIC High-income countries NIH National Institutes for Health HIV/AIDS Human immunodeficiency virus OECD Organization for Economic Co- / Acquired immune deficiency operation and Development syndrome OOP Out of Pocket xl AB B R E V IAT I O NS COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E OPAs Older People’s Associations SIGI Social Institutions and Gender Index PAHO Pan American Health Organization SMU Standardized mortality unit PAICC Programme for the International SSA Sub-Saharan Africa (World Bank Assessment of Adult Competencies  regional vice presidency) PCI Percutaneous coronary intervention SSBs Sugar-sweetened beverages PHC Primary Healthcare STI Sexually transmitted infections PPP Purchasing power parity or Public- UHC Universal health coverage Private Partnership UK The United Kingdom QALY Quality-adjusted life year UMIC Upper-middle-income countries QOL Quality of life UNWPP United Nations World RIP Relative Income Price Population Prospects SAGE Study on Global Ageing and USA The United States of America Adult Health VSL Value per statistical life SAR South Asia Region (World Bank VSMU Value of standardized mortality unit regional vice presidency) WBG World Bank Group SBCC Social and behavioral WHO World Health Organization change campaign WoG Whole-of-government SDG Sustainable Development Goal WoS Whole-of-society SDHs Social determinants of health WPP World Population Prospects SES Socio-economic status YLD Years lived with disability SHARE Survey of Health, Ageing, and Retirement in Europe 1 Theoretical Framework Chapter 1 Enhancing human capital and boosting productivity by tackling non-communicable diseases: results of a research initiative Chapter 2 Towards a framework for impact pathways between non-communicable diseases, human capital and healthy longevity, and wellbeing outcomes 1 2 CHAPTER 1 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E Enhancing Human Capital and Boosting Productivity by Tackling Non-Communicable Diseases Results of a research initiative George Alleyne a, Timothy Evans b, Alexander Irwin c, Prabhat Jha d, Jeremy Veillard e a Director Emeritus, Pan American Health Organization (PAHO) b School of Population and Global Health, McGill University c Independent Global Health Writer and Researcher d Centre for Global Health Research, Unity Health Toronto, University of Toronto, Canada e Health, Nutrition, and Population, Latin America and the Caribbean, World Bank 3 INTRODUCTION AND OVERVIEW Non-communicable diseases (NCDs) (especially cardiovascular disease, diabetes, cancer, chronic respiratory disease, and mental health disorders) increasingly threaten countries’ health achievements and economic future. WHO estimates that 71 percent of all deaths in 2015 were due to NCDs. Over one billion persons suffer from hypertension, which alone is responsible for over 10 million deaths worldwide each year, more than all infectious diseases combined prior to COVID-19. The burden of NCDs continues to rise disproportionately in low-income countries (WHO 2018). NCDs also contribute to 80 percent of the global burden of disability (IHME 2018). Many non-communicable conditions strike working adults during their most economically fruitful years: cutting short lives and careers, undermining productivity, bankrupting families, diverting public resources from more productive uses to cover treatment costs, and preventing society from recouping its investments in workers’ train- ing and skills. In low- and middle-income countries (LICs and MICs), NCDs tend to affect people at younger ages, reducing educational attainment and lifetime earnings. Political mobilization around the NCD threat is in- ment of human capital. The Human Capital Proj- creasing but still not commensurate with the scale ect aims to improve human capital at country level of the challenge. The United Nations political decla- through investment in health and education. The ration on NCDs, adopted in 2011, broke the silence Project’s human capital index (HCI) is an advocacy at top levels. The 2013 WHO Global Action Plan tool that quantifies the contribution of health and established a framework for coordinated efforts education to the productivity of the next generation to counter NCD epidemics nationally and inter- of workers. Countries are using the HCI to assess nationally. In 2015, the Sustainable Development how much income they forego because of human Goals (SDGs) placed NCDs prominently on the capital gaps, and how much faster they can turn global development agenda, with a dedicated NCD these losses into gains if they act now. target (SDG 3.4). Countries’ growing recognition of the impor- Despite increased attention to NCD issues, tance of human capital represents an unprecedent- however, investment and action at country level ed opportunity to accelerate progress on NCDs, continue to fall short. NCD epidemics are grow- particularly in LICs and MICs. NCDs pose a power- ing explosively in many low- and middle-income ful threat to human capital. They damage and deval- countries (LMICs), while influential policy makers, ue it in the short term, mainly through impacts on particularly in ministries of finance, remain poorly adult survival and productivity. They also compro- informed about the implications of NCD burdens mise future human capital creation, in particular by for their countries’ economic development. impacting educational performance. Countries can reap substantial economic rewards via improved NCDs and human capital – human capital, both short- and long-run, by taking clarifying a critical relationship bold action to prevent and control NCDs. Growing concern about human capital creates a Against this backdrop, in 2017, the WHO Direc- window of opportunity for collaboration between the tor-General established a High-Level Commission World Bank, WHO, and other partners to help coun- to advise on means of accelerating action against tries enhance their human capital by tackling NCDs. NCDs. One of the Commission’s recommendations was to seek to incorporate NCDs into global policy The NCDs and Human Capital Research discussions on human capital. Initiative: research to inform policy Countries increasingly recognize the critical importance of human capital—the health, educa- Human capital is emerging as a pivotal concern for tion, and skills of the population—for economic countries and a powerful lens to guide public policy growth and competitiveness. Since 2017, more than and investment. Yet important knowledge gaps on 80 countries have joined the World Bank’s Human human capital persist. Today, most countries are not Capital Project, signaling high momentum for in- clear on the opportunities that exist to strengthen vestment in the creation, protection, and enhance- human capital by tackling NCDs, the rewards such 4 CHAPTER 1 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E efforts may yield, the levels and types of investment The Toronto workshop needed, and the specific policy and implementation and the role of this chapter steps that will work best. To respond, in April 2019, the World Bank, To review emerging results from the NCDs and Hu- in partnership with WHO, launched a time-limit- man Capital Research Initiative and reach consen- ed research initiative on NCDs and human capital. sus on their interpretation, project leaders convened The work was situated within the World Bank’s Hu- a workshop in Toronto on July 9-10, 2019. The event man Capital Project. It engaged leading academic was sponsored by Access Accelerated and hosted researchers and policy experts in a collaboration to by the Centre for Global Health Research of the answer priority questions on NCDs’ human-capital University of Toronto Dalla Lana School of Public impacts and entry points for an effective policy re- Health. Participants included members of the proj- sponse. The effort pursued four main tasks: ect research teams, along with other distinguished scholars and representatives of organizations in- 1. Evaluate the impact of NCDs on education, cluding Access Accelerated, the Pan-American as it relates to human capital Health Organization (PAHO), the United Nations Development Programme (UNDP), WHO, and the 2. Evaluate the impact of NCDs on human capi- World Bank. The detailed workshop agenda appears tal understood as effective labor supply in Annex 1.1 of this chapter. Annex 1.2 presents a full list of workshop participants. 3. Estimate the impact of NCDs on longevity This chapter summarizes the main findings emerging from the five core papers commissioned 4. Consider what measurement systems are by the NCDs and Human Capital Research Initia- needed to buttress the arguments for invest- tive, together with the analysis, interpretation, and ing in human capital at country level and to key messages for policy and practice jointly for- monitor progress mulated by participants at the July 2019 Toronto workshop. All data and analysis in this chapter are The NCDs and Human Capital Research Initiative pre-COVID-19. commissioned original research from five scientific The chapter follows the structure of the Toron- teams at major universities and international or- to workshop. It provides an initial overview and ganizations.1 Research teams began their work in contextualization based on the keynote address by May 2019, shared preliminary results in July 2019, Sir George Alleyne. It summarizes the five scientific and finalized their contributions in the following inputs from project research teams and the discus- months, completing their work in December 2019. sion sparked by each paper in Toronto. The chapter An executive report summarizing preliminary ini- then synthesizes policy lessons and recommenda- tiative findings served as an input to the concluding tions from the workshop debates. Concluding sec- deliberations of the WHO High-Level Commission tions describe the subsequent agenda of the NCDs on NCDs and informed its final recommendations. and Human Capital Research Initiative and outline Methods developed and results obtained by a future research and practice agenda to strengthen the NCDs and Human Capital Research Initiative human capital through action on NCDs. This agen- have subsequently been integrated into the World da has helped to shape the subsequent activities of Bank’s Healthy Longevity Initiative (HLI), launched the Healthy Longevity Initiative. in April 2021. The HLI is jointly led by the World Bank and the University of Toronto, with support from Access Accelerated. Several of the investiga- tors whose findings from the NCDs and Human Capital Research Initiative are summarized in this chapter have gone on to contribute additional orig- inal research to the HLI. 1 Participating research teams came from the Graduate Institute, Geneva; the Harvard T.H. Chan School of Public Health; the University of Toronto Dalla Lana School of Public Health; WHO; and the World Bank Group. 5 BOX 1.1 Key messages • Rising non-communicable disease (NCD) burdens threaten countries’ human capital, a critical determinant of their eco- nomic success. • Policymakers’ growing concern with human capital marks an unprecedented opportunity for the World Bank, the World Health Organization (WHO), and other partners to collaborate in supporting country action to prevent and control NCDs. • By tackling NCDs with proven, cost-effective interventions, including intersectoral measures, countries can reap substantial economic benefits, while improving health. Economic gains from strengthening human capital through NCD control and prevention will come mainly from improved productivity. Some gains can be obtained rapidly (1-5 years). • NCD prevention and control also lay foundations for longer-term improvements in human capital and economic competi- tiveness, notably through gains in young people’s educational performance. New research is uncovering strong associations between NCDs and educational outcomes. • NCDs are major drivers of impoverishment and inequality. Tackling NCDs will improve human capital most among the least well off. This will reduce inequalities, reinforce social stability, and ensure that countries harness talent from their whole populations, an advantage for competitiveness. • Private firms across multiple industries could draw productivity benefits from action on NCDs. Private sector co-leadership can boost the speed and impact of the global NCD fight. • Strong, country-owned measurement systems are crucial for countries to prevent and control NCD epidemics and reap the associated human-capital benefits. • Knowledge gaps hamper countries’ efforts to fight NCDs. NCD and human capital research—including research on preven- tion, the economic burden of disease, and how to measure productivity gains—needs to be aggressively ramped up. RESEARCH INPUTS AND POLICY DISCUSSIONS Non-Communicable Diseases and Human Capital Analytic Work and Key Messaging Workshop: Toronto, July 9-10, 2019 Workshop Day 1, morning session Daniel Dulitzky (World Bank) and Prabhat Jha (University of Toronto) chaired the meeting’s first session. Following their welcoming remarks, Sir George Alleyne (Director Emeritus, Pan American Health Organi- zation) delivered the keynote address. Workshop keynote address: health, which has two principal roots. First, his work health, NCDs, and human capital as a physician and practitioner of public health poli- cy has convinced Dr. Alleyne that the achievement Sir George Alleyne of health equity is possible and technically feasible, whereas equity in other domains of human life is Dr. Alleyne began his talk by paying tribute to three more difficult, perhaps impossible, to attain. This persons whose work has powerfully informed his circumstance gives health a special status. understanding of health, NCDs, and human capi- A second reason for Dr. Alleyne’s enduring tal: Professor Dean Jamison, Professor Prabhat Jha, concern with health is his acceptance of health’s in- and the late Philip Musgrove. Acknowledging these strumental value, as complementary to its intrinsic colleagues’ foundational contributions, Dr. Alleyne or constitutive value. In this connection, Dr. Alleyne then reflected on his own lifelong concern with evoked the conceptual framework developed by Par- 6 CHAPTER 1 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E tha Dasgupta and colleagues (2017). Dasgupta views prevention and control of NCDs represents a potent health as having four major dimensions in relation mechanism for preserving or restoring the productivi- to human agency. Good health is important because ty of human capital.” Furthermore, we have the chance it: (1) enhances human productivity, (2) enables per- to project this message just as human capital is being sons to live longer lives (longevity), (3) enables bet- promoted with renewed vigor by the World Bank, as ter quality of life and well-being, and (4) yields posi- a factor critical for countries’ economic performance. tive externalities at individual and population levels. In closing, Dr. Alleyne signaled a concern of This framework offers a conceptual armature for methodological and philosophical significance for the work of the NCD and Human Capital Research human capital discussions, regarding whether sci- Initiative. If Dasgupta’s basic theorems are valid, Dr. entists can “separate the asset of the individual from Alleyne affirmed, then we should be able to frame an his or her person.” With health and human capital, appropriate argument to convince policy makers of scientists measure, not health itself as the asset, but the salience of NCD prevention and control. the factors (in this case NCDs) which diminish the Dr. Alleyne noted that the Toronto workshop asset; they then correlate these factors with labor and the broader political process of which it is part productivity. Dr. Alleyne underlined his hope that “represent a tide in the NCD affairs which, hopefully, the day comes “when economists find some mea- we can take at the flood.” To justify such optimism, he sure of health which they can use to relate to the cited the ongoing work of the WHO Director-Gener- labor market outcomes they currently estimate.” al’s High-Level Commission on NCDs, of which he was a member. The drafting of the High-Level Com- Highlights of workshop discussion: In the discus- mission’s final report was in progress at the time of the sion that followed, Dr. Alleyne posited that health as Toronto consultation. Dr. Alleyne had received the a state of being similar to happiness might be assessed assurance that the results of the Toronto workshop using a metric derived from qualitative research. Re- would be reflected in the High-Level Commission’s garding the relatively slow progress on NCDs to date, report. He pointed to the World Bank’s Human Capi- he urged that the NCD community frame its argu- tal Project and related technical work on the “Chang- ments more persuasively, and referred particularly ing Wealth of Nations” as further evidence of a rising to the need to address NCDs and their risk factors tide of global concern with health and human capital in children. An opportunity for progress is emerging that signals an exceptional political opportunity. through the confluence of current interests among While much discussion of the Human Capital major institutions: the World Bank in human capital; Project has focused on the foundations of human WHO in NCDs; and UNDP in inequality, along with capital in early childhood, Dr. Alleyne observed: the burgeoning engagement of civil society. My view is that the focus on early childhood is a necessary but not sufficient rationale for an inter- Summary and discussion of research papers est in human capital. I contend that our concern must be for the actual as well as the potential hu- Paper 1: The human capital impacts of investing man capital. Our concern must be for the loss of in interventions to reach the SDG-related targets human capital through premature mortality due for non-communicable diseases: an analysis of to NCDs. Our concern must be for the billions, 10 low- and lower-middle income countries yes, billions of adults who suffer from one or more NCDs which impair their human capital. Seoni Han*, Jean-Louis Arcand*, Jeremy Addison Lauer** (*) The Graduate Institute, Geneva; (**) World Health Orga- In this light, he argued, one of the novel NCD ap- nization, Geneva proaches to be suggested to the WHO Director-Gen- eral is that WHO make common cause with the World Top-line messages: Expanding coverage of NCD Bank at country level, ensuring that investment in hu- interventions in low- and lower-middle income man capital include not only future capital, but also countries can raise countries’ effective labor supply, the quality and quantity of the present stock of human boosting the economy. capital, through prevention and control of NCDs. In this respect, Dr. Alleyne concluded, “We have • NCD interventions increase GDP by about 0.5% the opportunity to demonstrate that investing in annually over 5 years 7 • NCD interventions increase GDP per capita by breast, and cervical cancer. NCD-related inter- about 0.7% annually over 5 years ventions included in the model were: policy and population-wide interventions; periodic outreach • NCD prevention and control are a major determi- campaigns; primary care screening, treatment, and nant of human capital and must figure prominently management; and limited specialized care (e.g., in universal health coverage (UHC) in the near term treatment of ischemic heart disease, COPD treat- ment, and cancer treatment). Cost estimates in- Study background and aims: To support WHO’s cluded the programmatic investments necessary for 2018 Investment Case, the researchers investigated intervention delivery that do not generate economic the impact of scaling up NCD interventions on eco- returns on the selected time horizon. nomic outputs in 10 low- and lower-middle income The paper focuses on two main channels countries (LICs and LMICs) for the period 2019– through which NCD interventions affect a coun- 2023. The countries were: Angola, Bangladesh, Bra- try’s economic outcomes: the cost effects of NCD zil, China, the Democratic Republic of Congo, Ethi- interventions on the accumulation of manufactured opia, India, Indonesia, Nigeria, and Pakistan.2 The capital, and the demographic and health impacts study aimed to calculate the GDP and GDP per capi- of NCD interventions on effective labor supply via ta gains these countries might obtain, if they success- improvements in human capital. This approach fully reached 60 percent population coverage of key foresees that, when an NCD intervention takes NCD interventions, as stipulated under SDG targets. place: (1) a decrease in the stock of physical capital is incurred because of the cost of the intervention4; Methods: The researchers used WHO’s Economic while (2) an increase in labor supply results from a Projections of Illness and Cost of Treatment (EPIC) decline in the mortality and disability due to illness. model, which relates market-valued economic The researchers allow for heterogeneity at output in human-capital terms to the direct and a country level, using country-specific econom- indirect effect on human capital of the burden of ic and disease assumptions. It is necessary to take non-communicable disease averted.3 into account country-specific responses to changes The paper can be seen as presenting develop- in health status following intervention to control ment-accounting mechanisms for the impact of im- certain pathological conditions, because countries proved health status on economic performance in have experienced different stages of epidemiologi- the countries studied, since “human capital” is con- cal transition (Deaton 2006). The model also takes ventionally understood as a stream of future earn- into consideration the fact that the improved health ings (as a component of overall economic output). status of children impacts on the productivity of The authors take changes in overall economic out- their caregivers of working age. put attributable to increases in labor supply due to Analyzing changes in health outcomes and the improved NCD control (i.e., as estimated through consequences of NCD investments for manufac- the “labor share”) as their measure of the improve- tured capital and the labor force, the authors com- ment in human capital due to implementation of a pared gains and losses in terms of economic output package of NCD interventions. for a “business as usual” scenario and an “ambitious The study considered the following NCDs: car- realization of UHC” scenario from 2019 through diovascular diseases, endocrine disease (diabetes), 2023 (Sternberg et al. 2017). Under business as usual, chronic obstructive pulmonary disease (COPD), population coverage rates for relevant interventions asthma, anxiety, depression, bipolar disorder, psy- remain constant at 2015 levels. Under the ambitious chosis, conduct disorder, attention deficit disor- scenario, countries reach their SDG coverage targets der, alcohol dependence, epilepsy, and colorectal, (for NCD services, 60 percent population coverage). 2 Together, these countries account for 67 percent of all expected global deaths for the period 2019-2023 that could be averted by specified advances toward universal health coverage (UHC). For details of the UHC scenarios considered, see Sternberg et al. (2017). 3 The EPIC model is based on a human-capital augmented Solow model and incorporates a recursive production function and exogenous equations of motion for the two production factors – stocks of physical capital and effective labor as modified by human capital. 4 Interventions reduce the funds available for investment because private savings are diverted to pay for interventions. (This is equivalent to out-of-pocket payments.) It is assumed that some portion of the total cost of treatment is funded from domestic savings, with the remainder displacing other consumption expenditures. There is no public sector in this model. 8 CHAPTER 1 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E Key results: The findings suggest that investing in gains of 0.5 percent and average GDP per capita better prevention and control of NCDs would gen- gains of 0.7 percent over baseline for these countries erate average annual gross domestic product (GDP) over the five-year period 2019–2023. FIGURE 1.1 GDP and GDP per capita gains, average for 10 countries, with an NCD package FIGURE 1.2 GDP with (perturbed) and without (projected) an NCD package 9 Discussion, interpretation, and next steps: Among the researchers’ conclusions would be to say that: other practical policy lessons, the results of this “Reaching NCD targets that are embedded within analysis suggest that economic gains from strength- UHC should be expected to produce an increase in ened NCD prevention and control measures could GDP of about 0.5 percent” annually across coun- manifest rapidly, within politically pertinent time- tries over the five-year period under consideration. frames (1-5 years). The model used in this study is flexibly applica- Paper 2: Health capital, economic growth, and ble for different packages of disease conditions and the burden of disease various coverage of health interventions for differ- ent countries. The authors plan to explore for more Jean-Louis Arcand*, Daniele Rinaldo* countries the impacts of investing in prevention and (*) The Graduate Institute, Geneva control of NCDs through UHC. One important as- pect for future work is to incorporate dynamics be- Top-line messages: tween the costs of a health intervention and its con- sequential health outcomes, in order to link the laws • Complementing empirical analyses, theoretical re- of motion of physical capital and labor in the model search in economics may yield new tools that will (that is, how these variables evolve over time). This ultimately enable policymakers to better understand work will follow amelioration of the law of motion the links between health and economic conditions for human capital. • The theoretical model proposed here allows re- Highlights of workshop discussion: Professors searchers to compute a stock of health capital, which Prabhat Jha (University of Toronto) and Sue Hor- can be measured in dollars, and which is compatible ton (University of Waterloo) served as lead dis- with individual optimizing behavior, for all countries cussants for the day’s first two papers. Deferring in the world; this enables investigations and compar- in-depth discussion until after the presentation of isons that were not achievable with previous tools Paper 2 (see below), they asked participants for questions of clarification regarding Paper 1. Fur- • Early results with the model include the observa- ther explanation was sought on two main issues: (1) tion that countries’ health capital per capita tends to what the researchers mean by “physical capital”; and grow significantly faster than GDP per capita (2) whether the NCD interventions included in the model incorporated forms of multi-sectoral action. • The model can be used to run counterfactual ex- Dr. Jeremy Lauer responded. On physical cap- ercises that yield, for example, estimates of the re- ital and its relation to human capital in the model, duction in global effective labor supply caused by he referred participants to the equations included NCD burdens in the World Bank’s recent Human Capital Primer (Flabbi and Gatti 2018). He noted that research gen- Study background and aims: In the field of health erally considers human capital in terms of impact economics, Grossman (1972) is the canonical dy- on national income, such as GDP. Human capital namic model of the demand for health. In this model, is in the labor term of the relevant equations. NCD rational individuals are assumed to divide their in- interventions lead to a diminution of physical cap- come between (i) consumption, (ii) saving, and (iii) ital because of the costs involved, but they simulta- investing in health capital. Their welfare is assumed neously improve the quantity and quality of human to depend solely on consumption: this, therefore, is capital, yielding economic benefits. On the question a world in which health is a pure investment good. of interventions outside the health sector, Dr. Lauer In the study summarized here, Arcand and Rinal- explained that the researchers’ full model included do develop an analytically tractable Grossman-type 188 interventions, some of which go beyond clinical model, which they solve using standard dynamic health care to encompass, for example, public policy programming methods, and which they are then able choices affecting issues like road safety. Meanwhile, to fit to actual data, allowing them to run a number of the focus here clearly remains on the first of the interesting counterfactual exercises. They can do so “three billions” highlighted in WHO’s 2018 Invest- because the theoretical model allows them to com- ment Case: that is, expanding the number of people pute a stock of health capital, which can be measured with access to universal health coverage (UHC). Dr. in dollars, and which is compatible with individual Lauer proposed that an appropriate way to restate optimizing behavior, for all countries in the world. 10 CHAPTER 1 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E Methods: In contrast to standard dynamic macroeco- ate at an increasing rate as an individual ages, leading to nomic models, individuals in Arcand’s and Rinaldo’s finite lifetimes becoming both inevitable and optimal. model are endowed with two stocks of capital: (i) assets The final analytical task is to derive average behavior which generate interest income and (ii) health capital, an in the economy at a given time by aggregating over all important component of human capital, the latter being those individuals who happen to be alive. Remarkably, equated with effective labor supply. In accordance with the expressions for aggregate behavior turn out to be al- human physiology, health capital is assumed to depreci- most as tractable as their individual components. FIGURE 1.3 Illustration of the theoretical model in terms of individual optimizing behavior The basic intuition of the theoretical model in The empirical calibration of the model relies on terms of individual optimizing behavior is illustrated five types of data: (i) life expectancy at birth, which in Figure 1.3. The optimal health capital h∗(t) of an in- corresponds to the optimal finite lifetime of an indi- dividual is high at birth and gradually decreases, as the vidual blessed with perfect foresight and which cor- health depreciation rate increases with age ‘t’. This is responds, in Figure 1.3, to the point where the health also the path followed by effective labor supply, which capital curve intersects the horizontal axis; (ii) health falls to zero when health capital reaches the horizontal expenditures per capita which, ceteris paribus, in- axis (which corresponds to death). Concomitantly, op- crease the stock of health capital; (iii) the difference timal health expenditures m∗(t) initially increase with between life expectancy at 60 and life expectancy at age (in order to smooth the path of health capital), but birth, which allows researchers to obtain a rough mea- eventually fall, as the effect of the increasing depreci- sure of the time-varying component of the deprecia- ation rate of health capital become sufficiently large: tion rate of health capital; (iv) information on the la- this yields an inverse U-shaped pattern over time; op- bor share (measured by wages and salaries divided by timal consumption expenditures c∗(t) follows a sim- value-added), which allows researchers to character- ilar path. Finally, individual assets (which through ize the production technology that produces output the usual national income accounting identity drive using physical capital and effective labor input, and the behavior of the physical capital stock available to which is a key ingredient determining the optimizing firms) increase at a decreasing rate: they cannot fall to behavior of producers; and (v) information on disabil- zero until the time of death, because the individual’s ity-adjusted life years (DALYs) lost to various pathol- ability to finance consumption is increasingly depen- ogies, which allows researchers to parameterize the dent upon interest income, as labor income falls along baseline (invariant) depreciation rate of health capital with her health capital and effective labor supply. and, later, to conduct counterfactual simulations. 11 FIGURE 1.4 Estimated health capital per capita (logarithmic scale), for 92 countries, 2011 US$ PPP Note: The dark blue line corresponds to a nonparametric smooth of the average behavior in the sample. PPP = Purchasing Power Parity. Results: Longevity, investment in health capital through Several interesting empirical regularities emerge. health expenditures, and the depreciation rate of health First, as one would expect, the stock of health capital capital are intimately tied together in the model by the per capita is positively correlated with GDP per capi- fetters of optimizing behavior. Applying standard gen- ta, though the researchers’ estimation procedure does eralized methods of moments techniques, and the re- not anchor the former on the latter – this is an import- strictions implied by the theoretical model, the re- ant point, in that they are not deriving their estimate searchers are able to estimate two key parameters: (i) of health capital by a procedure that is in any manner the efficiency with which health expenditures are trans- linked to an estimate of the value of a statistical life. Sec- lated into increases in health capital and (ii) the sensitiv- ond, health capital per capita tends to grow significant- ity of effective labor supply to changes in health capital. ly faster than GDP per capita. A noticeable example in With these two parameters in hand, they are then able the above figure is China (see the CHN code in Fig- to simulate, using the theoretical model, the path of the ure 1.4), where health capital per capita goes from just stock of health capital per capita for every country in above $10 in 1995 to almost $1,000 in 2013 –a far larger their sample. This is illustrated in Figure 1.4 (note the proportional increase than that of GDP per capita. logarithmic scale on the vertical axis). The paths fol- lowed by health capital per capita are computed by ag- Applying the model: a counterfactual on NCDs. gregating the optimal behavior of all generations of in- What are the implications of this model when it dividuals alive in a given country in a given year, where comes to the impact of NCDs on health capital? the longevity of each generation is also determined en- Figure 1.5 compares the paths of optimal individ- dogenously in the underlying theoretical model. ual health capital and health expenditures with and 12 CHAPTER 1 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E FIGURE 1.5 Effect of NCDs on health capital and health investment without NCDs, where the latter are assumed to in- shop members encouraged Professor Arcand and his crease the rate at which health capital depreciates. team to provide readers with a clearer feel for what When the disease burden is reduced, individual op- the numbers represent and to proceed pedagogically timal health capital unambiguously shifts upwards, in spelling out the implications for policy makers. whereas the inverse U-shaped path of optimal Workshop participants went on to note that, health expenditures shifts to the right: as one would while the empirical analysis undertaken by Han, expect, optimal longevity is also increased. Arcand, and Lauer (Paper 1) could usefully be ex- tended to 2030, there is value in focusing on a five- Highlights of workshop discussion: Professors Pra- year timeframe, which approximates the electoral bhat Jha and Sue Horton again led the conversation. cycles that drive political decisions in many set- The discussants noted broad consistency between the tings. That NCD investments can generate benefits research presented and existing work on topics such in a timeframe of <5 years is an important politi- as the “Portfolios of the Poor.” Health capital growth cal message. On the other hand, some felt that the appears to be faster than GDP growth, which again actual GDP and GDP per capita benefits calculated is consistent with existing work by Dean Jamison, might appear modest to policymakers. Participants Charles Kinney and others. It is thought-provoking discussed whether the model could be adjusted to that the world is growing more unequal in terms of more fully reflect the vast economic burden of men- income, yet may be improving in terms of health cap- tal health conditions, for example, and the benefits ital and its distribution. The discussants and work- that would stem from effectively treating them. Workshop Day 1, afternoon session Erica Di Ruggiero (University of Toronto) and Daniel Dulitzky (World Bank) chaired the first day’s after- noon session. Paper 3: Estimating the longevity benefits of Top-line messages: By aggressively tackling non-com- non-communicable disease mortality reductions municable diseases (NCDs), and in particular reduc- in low- and middle-income countries over 2017- ing mortality from cardiovascular diseases (CVD), 2030: findings from a modeling study low- and middle-income countries can score substan- tial longevity gains that carry high monetary values. Aayush Khadka*, Stéphane Verguet* (*) Department of Global Health and Population, Harvard The study shows: T.H. Chan School of Public Health • Potentially large, sustained longevity gains for 13 controlling NCDs (particularly CVD) and injuries ease-specific mortality rates that correspond to the 90th percentile of the lowest mortality rates of the • Potentially substantial, sustained longevity losses next-higher country income group in 2016; and a (high cost of inaction) if countries fail to tackle “low-performance” trajectory that relies on 2030 NCDs (particularly CVD) and injuries age-sex-disease-specific mortality rates that cor- respond to the 10th percentile of the next-higher This work also establishes a priority-setting paradigm. country income group rates in 2016. It opens the way to a monetized assessment of the glob- All three trajectories are examined for eight al burden of mortality by cause that will enable com- broad disease categories: neoplasms; CVD; chronic paring health-sector and intersectoral investments. respiratory diseases; diabetes and other endocrine diseases; mental disorders; injuries; communicable Study background and aims: To best set priorities diseases; and maternal, neonatal, and nutritional and allocate investments toward the health sector, it diseases. The study estimates per-person changes in is essential to understand the respective contribu- longevity at all ages between base-case, low-perfor- tions of different diseases and conditions to changes mance, and high-performance trajectories as differ- in longevity over time. The researchers estimate the ences in age-sex-year-specific life expectancy per longevity changes induced by various mortality tra- disease category. Using VSL methods, the research- jectories between 2017 and 2030, the end date of the ers derive associated monetary values (in 2011 USD Sustainable Development Goals, for major disease adjusted for Purchasing Power Parity). To assess the groups including NCDs for low-income countries robustness of the estimated monetary values, they (LICs), lower-middle-income countries (LMICs), conduct several sensitivity analyses. and upper-middle-income countries (UMICs). In addition, they quantify the monetary value asso- Key results: The high-performance scenario of CVD ciated with the changes in the mortality schedule mortality control would be associated with impres- that lead to these longevity changes. This effort can sive gains in longevity that carry a high monetary val- be thought of as contributing a “first brick” toward ue for LICs, LMICs, and UMICs. The analysis shows the construction of a money-metric priority setting potentially large per-person period-specific changes framework for health sector investments, building in longevity under the high-performing scenario for on the momentum of “Global Health 2035,” and CVDs for both females and males by 2030; for LICs, drawing from the Copenhagen Consensus exercise. LMICs, and UMICs, the monetary values associat- ed with changes in the age-sex-disease-year-specific Methods: The study’s goal was to estimate mon- mortality schedule for females are estimated at about ey-metric values of longevity changes per disease $1,700, $11,000, and $30,000 in 2030 respectively; for category between 2017 and 2030 in two steps: males, the associated monetary values in 2030 are es- timated at $1,700, $15,000, and $42,000 respectively. • Step 1: estimate changes in longevity (changes in Longevity benefits and the monetary value of the un- mortality rates, hence life years) under different derlying changes in the mortality schedule would be mortality scenarios smaller, albeit still substantial, for other NCDs and injuries considered in the analysis. In both LICs and • Step 2: assign monetary values to the underlying LMICs, large longevity gains via continued reduction changes in the mortality schedule responsible for of mortality from communicable diseases could also changes in longevity by using value of statistical be achieved in the short term. life (VSL) methods Critically, the findings highlight potentially large longevity losses under the low-performance Accordingly, the researchers model survival curves scenario for CVDs across all three income group- for 29 LICs, 50 LMICs, and 50 UMICs under three ings: for females in 2030, the per-person mone- mortality scenarios between 2017 and 2030: a base- tary value associated with the change in mortali- case trajectory that draws from annual age-sex-dis- ty schedule leading to these losses is estimated to ease-country-specific mortality forecasts by the be of about $4,000, $15,000, and $42,000 in LICs, Global Burden of Disease study; a ”high-perfor- LMICs, and UMICs, respectively; for males, the re- mance” trajectory that is defined, for each country spective per-person monetary values are projected income group, as achieving (by 2030) age-sex-dis- to be of about $6,000, $31,000, and $55,000. Lon- 14 CHAPTER 1 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E gevity losses due to injuries and their associated injuries should be viewed with their respective con- monetary value were especially high for males in tributions toward reducing the burden of illness-re- LICs and LMICs, relative to other NCDs: in 2030, lated impoverishment and alleviating poverty. the per-person monetary values are projected to be This research shows that, by systematically around $6,000 and $28,000 in the two respective monetizing the global burden of mortality by cause, country income groupings. one can conceive a priority-setting paradigm to best allocate investments toward the health sector, Discussion, interpretation, and future work: The which can then be evaluated in terms of returns on study finds that large longevity gains, as measured investment and be compared across other sectors by life years and associated monetary values, could outside the health sector. materialize via curbing NCD- and injury-relat- ed mortality. For women, the greatest gains would Highlights of workshop discussion: Professor Dean come from addressing CVDs and neoplasms; men’s Jamison (UCSF) served as main discussant for the longevity gains would be greatest from better control paper. Professor Jamison observed that the work pre- of CVDs, neoplasms, and injuries. Furthermore, this sented could advance the generation of an “economic study points to important setbacks if NCD control burden of disease by time and place,” which would be is not scaled up: substantial longevity losses could highly informative for policy deliberations. The “eco- result, particularly for CVDs and injuries across all nomic burden of disease” would in a sense be com- three country income groups. In addition, largely plementary with the existing epidemiological GBD. due to the unfinished agenda of infectious diseases, Among other benefits, it would help stop the confu- the researchers observe that, in the short term, much sion spawned by the proliferating “Investment Cas- longevity could be gained via communicable disease es” now regularly produced by many disease-specific mortality reduction in LICs and LMICs. programs and constituencies (including for NCDs!). Generally, the analysis stresses that LICs and This work could also, Professor Jamison suggested, LMICs could see large longevity gains in the short be interpreted as a major step towards challenging term while focusing on the unfinished agenda of infec- the models of DALYs and QALYs. tious diseases and maternal and child health. Yet, giv- The idea of a “global economic burden of dis- en the potentially large longevity losses of a low-per- ease” sparked the interest of workshop participants, formance trajectory for major NCDs like CVDs, LICs who explored potential implications for policy and and LMICs should urgently address NCDs and inju- practice. Several participants requested that re- ries, where great sustained benefits lie ahead. UMICs, searchers orient the community in understanding too, should prioritize reducing the burden of mortali- how practitioners can best use these tools to assess ty due to NCDs and injuries, as the analysis highlights and prioritize interventions (including prevention). large potential losses in the case of inaction. Rewards will come from identifying anchoring Nevertheless, the researchers found wide vari- points in these models where policy can engage. ations in the estimated per-person monetary values Participants also asked how the monetized re- under various sensitivity analyses, which highlights sults obtained by Professor Verguet related to the the need to conduct more robust analyses in elicit- calculations of GDP and GDP per capita gains pre- ing mortality risk valuations in LICs, LMICs, and sented earlier by Dr. Lauer and team (Paper 1 in UMICs, in terms of estimating the value of a sta- this chapter). In the course of extended follow-up tistical life. Future work can also build upon this discussions, Professor Verguet explained that it is, study by replicating the approach in other countries in fact, difficult to directly compare the results from and settings (e.g., sub-Saharan Africa, South Asia). the two analyses. Fundamentally, the Khadka and Replications which rely on micro-data or adminis- Verguet (KV) analysis differs from the Han, Arcand, trative data collected in these countries and regions and Lauer (HAL) study in that KV estimate mone- may provide an important point of comparison for tary values associated with changes in mortality risk, the results presented in this analysis. whereas HAL estimate changes in GDP and GDP Future work should also examine within-coun- per capita as a result of changes in the composition try inequalities in the distribution of these longevi- of the skill-augmented labor force and productivity. ty effects along the socioeconomic gradient and age For HAL, GDP and GDP per capita estimates are the groups. Likewise, disease-specific mortality reduc- key outcomes, whereas for KV, gross national income tions including from NCDs, mental disorders, and (GNI) per capita (of which GDP is a component) is 15 an input in their analysis to ultimately estimate the 3. What are the key trends in adult and NCD mortality? monetary value of changes in mortality risk. While not directly comparable, the two meth- 4. Which macro-level policies reduce adult mortality? ods can be thought of as complementary approaches to putting a value on the health gains that countries 5. By how much could stronger NCD interventions can obtain by addressing the NCD epidemic. Both speed SDG progress and reduce premature deaths? can provide potentially motivating information for decision-makers. Both expand the current toolkit of This summary focuses on questions 1, 4, and 5, instruments for NCD analysis and policy dialogue. which have particular salience for engaging policy makers (notably ministers of finance) around NCDs Paper 4: Measurement and determinants of and human capital. Question 2 provoked lively dis- NCDs and their impact on adult survival cussion in the workshop, reflected in the section of this chapter on measurement challenges. Daphne Wu*, Jeremy Veillard**, David Watkins***, Ryan Hum*, Dean T. Jamison#, Prabhat Jha* Economic benefits of reduced adult mortality (*) University of Toronto; (**) World Bank; (***) University Ms. Wu and Professor Jha began by approaching of Washington; (#) University of California at San Francisco from a new angle the question of potential econom- ic payoffs from adult mortality gains, a key theme Top-line messages: of the workshop’s earlier papers. Wu and Jha argued that substantial evidence can be drawn from the • Reducing adult mortality (probably) improves in- existing public-health literature to bolster the case come growth that gains in adult life expectancy can deliver major economic benefits for countries. • Consider 40q30 (mortality from age 30 to 69) as Earlier unpublished analysis of data from the the main metric of adult survival for the Human former states of the Soviet Union provides one illus- Capital Index and the SDGs tration. It is well known that adult male life expec- tancy in the former Soviet states declined sharply • Adult mortality is generally falling worldwide, in the years preceding and after the dissolution of but rising vascular mortality in some settings the USSR, with NCDs (particularly alcohol-related) raises concerns among the drivers of increased mortality. Trends in adult male mortality in the former Soviet region • Costs for reducing adult mortality, at the margin, and in Organization for Economic Co-Operation are rising and Development (OECD) countries have diverged sharply over recent decades (Figure 1.6). • Increasing income and increasing spending on Building on earlier work by Jamison, Lau, and health are effective means of reducing adult NCD Wang (2005), Wu and Jha showed that it is possi- mortality ble to calculate the economic benefits that former Soviet states might have obtained by reducing their • An expanded package of NCD interventions would adult male mortality during the period. In 1990, the go a long way toward achieving the relevant SDGs average 15-year-old male living in the former Soviet republics had a 72 percent likelihood of surviving Study background and aims: The research being to age 60. The corresponding probability in OECD pursued under Paper 4 encompasses the following countries was 86 percent. If adult male life expec- questions: tancy in the former Soviet states had matched the OECD average from 1990 on, the ex-Soviet coun- 1. Does better adult health increase income growth? tries would have enjoyed a substantial boost in eco- nomic growth: growing at 1.4 percent per year, on 2. Can we move from the current 45q15 adult mor- average, over the past three decades, rather than the tality metric to 40q30? 1 percent average rate they actually experienced. 16 CHAPTER 1 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E FIGURE 1.6 Worsening adult male mortality in former Soviet states, 1950-2000 Source: Peto et al. 1994. Determinants of adult mortality fusion and adoption of new technologies, to affect (including macro-level policies) the outcome. The model is given by: Lowering adult mortality can yield substantial eco- ln 40q30i,t = β0 + β1FeEdi,t + β2ln GDPPCi,t + nomic payoffs. But what are the policy levers that β3health_expi,t + β4OOPi,t + β5malariai,t + β6HIVi,t + countries can use to bring mortality down? Wu and β7urbani,t + β8electricityi,t + μi + εi,t Jha explained that a key task for their paper is to shed light on this question by examining the mac- The researchers used data from 163 countries for ro-level determinants of adult mortality. The follow- 2000, 2010, and 2016. They examined the associ- ing notes briefly summarize the methodology that ation between the macro-level variables and adult Wu, Jha, and team have used to analyze macro-level mortality attributed to (1) NCDs, (2) communi- determinants, and the findings obtained to date. cable, maternal, and perinatal conditions (CMPs), and (3) injuries. Methods: The researchers examine the macro-lev- el determinants of adult mortality (mortality at age Results: The researchers found that, other things 30-69, or 40q30), using a random effects regression being equal, female education is not significantly model. In Wu’s and Jha’s model, the macro-level de- associated with all-cause adult mortality, while a 10 terminants of adult mortality are as follows: mean percent increase in income is associated with a 1.5 years of female education (FeEd), GDP per capita percent reduction in all-cause adult mortality, and a (GDPPC), health expenditure as a percentage of a 10 percent decrease in HIV prevalence is associated country’s GDP (health_exp), percentage of health with a 34 percent reduction in adult mortality, re- expenditure that is borne out of pocket (OOP), spectively (Table 1.1). malaria incidence per 1,000 population at risk (ma- The finding that there is no significant associ- laria), HIV prevalence as a percentage of the popu- ation between female education and adult mortali- lation age 15-49 (HIV), urban percentage (urban), ty contrasts with the results of Pradhan and others and access to electricity (electricity). The model al- (2017), as they did not control for the impact of oth- lows for differences across countries, such as the dif- er macro-level variables. Female education, however, 17 is significantly associated with CMP- and injury-at- tributable mortality by 3.1 percent. An increase in tributable mortality: An additional year of female health expenditure of 2.5 percent of GDP would education decreases CMP- and injury-attributable reduce all-cause mortality by 3.8 percent, NCD-at- mortality by 6 percent and 2.9 percent, respectively. tributable mortality by 8.5 percent, and injury-at- Income is significantly associated with all four tributable mortality by 5.7 percent. Lower out-of- categories of mortality: A 10 percent increase in pocket health expenditure (OOP) is associated with income reduces both all-cause adult mortality and a reduction in adult mortality: a 10 percent decrease NCD-attributable mortality by 1.5 percent, CMP-at- in OOP reduces all-cause adult mortality by 2 per- tributable mortality by 2.5 percent, and injury-at- cent, though it does not affect NCD mortality. TABLE 1.1 Determinants of log adult (40q30) mortality and effect of intervention on adult mortality Coefficient Intervention Reduction in adult mortality (%) All-cause adult mortality Female education -0.004 ln GDP per capita -0.145* 10% increase 1.45% Health expenditure -0.015* 2.5% GDP increase 3.78% OOP 0.002* 10% decrease 1.95% Malaria 0.000(1) HIV 0.034* 10% decrease 34.13% Urban 0.000 Electricity -0.002* 10% increase 2.09% NCD-attributable adult mortality Female education 0.000 ln GDP per capita -0.153* 10% increase 1.53% Health expenditure -0.033* 2.5% GDP increase 8.45% OOP 0.001 Urban -0.003* 10% increase 2.62% Electricity 0.005* 10% decrease 4.82% CMP-attributable adult mortality Female education -0.058* 1 additional year 5.95% ln GDP per capita -0.248* 10% increase 2.48% Health expenditure -0.020† 2.5% GDP increase 5.05% OOP -0.002 malaria inc 0.000(2) HIV prev 0.090* 10% decrease 94.69% Urban 0.001 Electricity -0.011* 10% increase 10.89% Injury-attributable adult mortality Female education -0.028* 1 additional year 2.86% ln GDP per capita -0.309* 10% increase 3.09% Health expenditure -0.022* 2.5% GDP increase 5.65% OOP -0.001 Urban 0.004 Electricity 0.005* 10% decrease 4.74% Note: *p<0.05. Effect of intervention only indicated for variables with p<0.05. 18 CHAPTER 1 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E Assessing implications for policy and program- accounts for 60 percent of the total EUHC cost, pre- ming. Initial analysis suggests that effective means vention and treatment of NCDs would account for to reduce adult mortality from NCDs include: about US$15 billion (or 52 percent) in low-income increasing income and increasing spending on countries and about US$55 billion (or 56 percent) in health. Of note, the researchers also found that a 10 lower middle-income countries. percent decrease in out-of-pocket health expendi- With the implementation of EUHC, about 1.2 ture (OOP) reduced 40q30 by 2 percent. million deaths and 2.9 million deaths due to all caus- Further work is underway to deepen this es could be averted in low- and lower-middle-in- analysis and clarify policy and program-deliv- come countries, respectively. The number of deaths ery options. One clear conclusion from the in- averted due to NCDs, CMPs, and injuries would be vestigation to date is the importance of robust about 0.53 million, 0.65 million, and 0.06 million health spending, guided by strong evidence. This respectively in low-income countries, and about 1.9 approach is embodied in Disease Control Prior- million, 0.94 million, and 0.1 million respectively in ities 3 (DCP3), a key practical resource for coun- lower-middle-income countries. In terms of prog- try policy makers, their advisers, and partners. ress towards the 40q30 target, low-income countries DCP3 defined an Essential Universal Health Cov- would achieve greater progress in all-cause mortality erage (EUHC) package, which consists of a list of reduction, as well as in all of the three areas of con- 218 interventions that were deemed to provide dition-specific mortality, compared to lower-mid- good value for money in multiple settings, ad- dle-income countries (80 vs. 60 percent in all-cause dress a significant disease burden, and be feasible mortality reduction, 88 vs. 70 percent in mortality to implement in low- and middle-income countries reduction due to NCDs, 86 vs. 63 percent in mortal- (LICs and MICs) (Watkins et al. 2018). The pack- ity reduction due to CMPs, and 46 vs. 19 percent in age covers nearly all major health conditions from mortality reduction due to injuries). birth to death, including maternal and child health Similar analyses done in partnership with the conditions, infectious diseases, NCDs, injuries, and Asian Development Bank for urban areas of Indi- health services such as surgery and rehabilitation. an and Bangladesh also suggest significant benefits It even goes beyond direct health services to cover (Wu et al. 2020). the health system interventions used for monitoring and surveillance. Highlights of group discussion. The principal Watkins and colleagues estimated that, for 80 discussants for the paper were Adriana Blanco percent population coverage, the EUHC package (PAHO) and Jeremy Veillard (World Bank). A key would cost a total of US$65 billion in low-income topic of interest concerned the balance to be found countries and US$270 billion in lower middle-in- between focusing on adult mortality and morbidi- come countries. Based on a total population of 0.90 ty, in relation to NCDs. In terms of mortality as a billion and 2.7 billion in 2017, respectively, this proposed best measure of NCD impacts, some par- would mean US$72 per capita in low-income coun- ticipants remained skeptical, in light of the magni- tries and US$101 per capita in lower middle-in- tude of burdens imposed by non-fatal diseases such come countries. as mental health conditions, diabetes, obesity, and Of the total service delivery cost, which ac- others. For many, disability remains a crucial mea- counts for 60 percent of the total EUHC cost (while sure for communicating the importance of NCD the remaining 40 percent consists of costs of ancil- epidemics and the human and economic damage lary services and programs to support health ser- they inflict. Participants cited, for example, the ris- vices), prevention and treatment of NCDs would ing tide of neurodegenerative disorders (e.g., Alz- account for about US$17 billion (or 41 percent) in heimer) and mental health conditions. In the near low-income countries and about US$86 billion (or future, in some settings, resources on the order of 51 percent) in lower middle-income countries. 5-7 percent of GDP may have to be channeled to- Based on an estimated US$18 billion and US$73 wards public-sector management of these burdens. billion in current spending, the additional cost to Costs of such magnitude should catch decision cover 80 percent of the population in low- and lower makers’ attention. Some participants noted that the middle-income countries (“incremental cost”) would human capital framing enables this discussion to be be about US$46 and US$160 billion, respectively. Of positioned in the context of how we can most intel- the total incremental cost for service delivery, which ligently invest in people. 19 In considering how governments and societies the discussion by observing that the challenge re- can respond, the promise of technology to alter the mains of turning the NCD story into a narrative current equations has not been adequately grasped. that demands urgent action, a “narrative of crisis.” Large-scale new investment in NCD interventions He noted that education experts at the World Bank and technologies is needed. In this respect, some and elsewhere have been successful in communicat- argued, digital health could mark a revolution, in- ing around the idea of an education crisis. The hu- cluding in mental health and NCD care. man capital framing may help the NCD community As co-chair, Dr. Jeremy Veillard summed up achieve comparable impact. Workshop Day 2, morning session Tim Evans (World Bank/McGill University) and Rachel Nugent (RTI) co-chaired the first session of Day 2. The session began with a review of main results from the first day of the workshop, summarized by Sir George Alleyne and Professor Dean Jamison. Recapping results from Day 1 Health perspective as DALYs, which minimize the importance of disabil- Sir George Alleyne ity—a key facet of NCD impacts on human capital. Prabhat Jha and Daphne Wu, in Paper 4, Dr. Alleyne prefaced his comments by noting that, had argued persuasively for adopting the 40q30 while many aspects of the previous day’s rich dis- measure—rather than or in addition to 45q15—to cussion would merit attention, he would focus on assess the effect of NCDs on adult mortality. The the workshop’s core theme: links between NCDs subsequent conversation had again highlighted the and human capital. Through this lens, notable con- importance of considering disability in setting pri- tributions emerged from each research paper. orities for NCD prevention and control. Paper 1, presented by Jeremy Lauer, brought a Dr. Alleyne noted that, at various points powerful message on the GDP gains that countries throughout the day, the issue of a whole-of-govern- could achieve by scaling up NCD interventions. ment approach to NCDs arose, and the point was The paper specified key avenues of intervention made that a human capital perspective makes the on NCDs, from primary care services to popula- fiscal demands of dealing with NCDs easier to ac- tion-level and policy measures. Bringing relevant cept, because it opens a clear line of sight from NCD interventions to 60 percent coverage would lead to spending to economic gain. a decrease in physical capital concomitantly with an increase in the effective labor supply. The net result Economics perspective for countries would be an average annual increase Dean Jamison of 0.5 percent in GDP and 0.7 percent in GDP per capita over five years. Professor Jamison reported three main reactions to From Paper 2, presented by Jean-Louis Arcand, the research findings and discussion on Day 1. the most critical message for Dr. Alleyne was the First, on the economic benefits of NCD action, demonstration of a valuation of health capital which he observed that the community is now close to be- had two possible measurements – dollars or years. ing able to assemble a robust account of the economic The key contribution of this approach is the ability benefits of the gains in NCD prevention and control to quantify the impact of NCDs on health capital and and the dollar values that will be achieved. The liter- to validate that health capital grows faster than GDP. ature on this issue is generally consistent, and work- Stéphane Verguet, in Paper 3, had proposed a shop contributors have that literature well in hand. novel monetized assessment of the longevity benefits The intersectoral dimension of NCD challeng- of interventions to decrease NCDs. Dr. Alleyne noted es is a second key topic, where consensus within the that much of the workshop discussion of this paper community appears more elusive. On this topic, im- turned around the merit of a conjoined measure such portant choices (including for the workshop group) 20 CHAPTER 1 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E remain outstanding. Pressing problems around inter- 1. The prevalence of NCDs among children is sectoral action include at least two sub-issues: non-trivial and rising. Globally, NCDs account- ed for 10 percent of all deaths in children under 5 1. Intersectoral policies to manage behavioral and years of age in 1990 and 18 percent among children environmental risks. What policies are effective, aged 5 to 14 years. By 2017, these figures had in- and how would we know? What avenues of re- creased to 14 percent and 23 percent, respectively. search should be prioritized, to provide the answers policy makers need? Professor Jamison cited the 2. Yet much of the existing research evaluating the control of obesity as an example: evidence-based impacts of poor health on educational attainment policy recommendations are generally lacking; the and subsequent labor market outcomes has fo- obesity problem still belongs more to the research cused on issues of nutrition and early childhood agenda than to the domain where policies can be development (low birthweight, stunting, and psy- chosen from a set of proven options, with well-un- chosocial stimulation) and common acute condi- derstood costs and benefits. In general, Professor tions (parasitic infections and deworming). Jamison argued, the economics of risk factors lags far behind the economics of health. 3. Preliminary analysis of survey data from the UK and India shows that NCDs in childhood are 2. The financial impacts of NCD epidemics outside correlated with lower test scores and lower ed- the health sector. An example concerns the costs ucational attainment. The impacts of NCDs in of long-term care. Professor Jamison noted the childhood appear to accumulate into adult life. case of the Netherlands, where an aging population Experiencing NCDs in childhood is correlated with increasingly complex NCD comorbidities is with lower employment and earnings. propelling a cost explosion in long-term care. 4. Education does appear to mitigate some of the As a third and final point, Professor Jamison evoked adverse associations between NCDs and labor the perspective of ministries of finance on the fis- market outcomes. Nevertheless, productivity cal consequences of NCD burdens. He noted what losses in adults who had NCDs as children appear appears as an inevitable trend in many countries, to be large and warrant attention. whereby NCD-related services that had been infor- mally provided in households are being shifted onto 5. These preliminary associations underscore the the public sector (relevant diseases include mental need to pay attention to NCDs in children, iden- health conditions and neurodegenerative disor- tifying interventions to mitigate their adverse ders). The fiscal consequences of this shift are likely effects. The observed associations also highlight to be very substantial in many settings. the need for further research to causally identify As a result, Professor Jamison noted that his and quantify the impacts of NCDs on education personal takeaway from the first day’s discussion and subsequent labor market outcomes. was the opportunity for a “pithy focus on the crisis.” The major social problem we are facing, he argued, Study background and aims: The overarching ob- is the fiscal consequence for government of man- jective of the research is to understand the relation- aging the growing burden of NCDs—linked to the ships between NCDs in children and their parents, fiscal opportunities that may arise from better man- educational attainment, and wealth accumulation. agement of these diseases. Education positively impacts health, and good health and nutrition can improve both the quantity Paper 5: NCDs, education, and human capital: and quality of education (Behrman 1996; Pradhan how do NCDs interact with education to et al. 2018). There is, without doubt, reverse cau- affect human capital accumulation? sality between health, education, and wealth, and this has generated substantial research attempting Sanam Roder-DeWan*, Ojaswi Pandey**, Aakash Mohpal** to disentangle causal impacts (Strauss and Thom- (*) Harvard University (**) World Bank Group as 1998; Glewwe and Miguel 2008). However, the literature has mostly focused on in utero and ear- Top-line messages: The research is ongoing and cur- ly childhood conditions, nutrition in infants and rently points to five main preliminary results: children, communicable diseases such as parasit- 21 FIGURE 1.7 Substantial NCD burdens among children in low- and middle-income countries Note: 15.5% of DALYs for children <5 years globally (7a) and 42.4% for children ages 5 to 14 years (7b) are from NCDs. Even in LICs, NCDs accounted for 13.2% of DALYs for children <5 years old and 32.5% for 5- to 14-year-olds. In HICs, the corresponding figures are 44.4% and 72.3%. ic infections, vision problems in adolescents, and effects of NCDs on labor market outcomes. A final energy supplementation in adults. NCDs con- piece of analysis considers the impacts of NCDs stitute 62 percent of DALYs worldwide and 42.4 in adults (parents) on the education of children in percent of DALYs among those aged 5-14 years. households. Since NCDs impose a significant eco- However, little direct evidence exists evaluating nomic burden on households in the form of work- the impact of NCDs on educational attainment and place absenteeism, reduced employment, and lower wealth accumulation (Figures 1.7a and 1.7b). incomes, it is possible that children who grow up This study looks specifically at the impact of in households where adults are chronically ill face NCDs (of child and parent) on educational attain- greater challenges in accumulating education. ment in childhood and on the ability to accumulate Given that among the countries and data sets wealth, maintain savings, and avoid poverty in adult- selected, NCDs in children are reported only in the hood. The paper is broadly premised on a life-course UK and India, the first part of the analysis is restrict- approach that links early life experiences, starting in ed to these data. The UK data is a cohort study of the preconception phase, to a variety of outcomes 17,000 children born in 1958 who were followed to throughout life (WHO 2013; Baird et al. 2017). The age 55 (National Childhood Development Study). researchers hypothesize that an NCD shock in child- The India data is a nationally representative panel hood creates cumulative disadvantage both through household survey with two rounds – 2005 and 2012 lower educational attainment and through addition- (Indian Human Development Survey). Data from al, independent health pathways that compromise Mexico and Indonesia are panel surveys and are used adult capacity to accumulate human capital. only to assess the impacts of NCDs in parents on the educational accumulation of the children. Currently, Methods: To test this hypothesis, the ongoing re- analysis is well advanced for India and the UK; anal- search began with a literature review on the poten- ysis for the other countries is in earlier stages. tial pathways through which NCDs and education Research assessing the effects of health on ed- may jointly affect human capital accumulation. The ucation faces formidable econometric identification study is now proceeding to analyze survey data from challenges (Behrman 1996; Glewwe and Miguel four countries (India, Indonesia, Mexico, and the 2008). This study attempts to address the method- UK) to investigate the prevalence of NCDs among ological challenges by using a panel data household children and how NCDs in children may affect fixed effects estimation strategy. The panel data helps school enrollment, educational attainment, and la- remove time-invariant individual and household bor market outcomes. Extensions of the work also level unobservable characteristics, while the house- look at whether education can mitigate the adverse hold fixed effects ensure that all the identifying vari- 22 CHAPTER 1 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E ation comes from siblings in the household (which 11 percent lower scores at age 11. In India, having to some extent controls for genetic variation). Nev- NCDs is correlated with a reduction of 0.75 stan- ertheless, while these methods provide an improve- dard deviation in overall test scores. ment over analysis of cross-sectional data, as is well attested in the literature, such methods do not over- • Having NCDs in childhood is associated with low- come all econometric identification issues. Address- er educational attainment in both the UK and In- ing all identification issues is beyond the scope of the dia. In the UK, NCDs in childhood are associated paper. Therefore, the results should be interpreted as with 1.7 fewer years of completed education at age correlations and not as causal parameters. 23 (compared to the average of 8.6 years, statistical- ly significant at 1 percent). In India, the effects are Key results: While analyses are still in progress, sug- even larger, considering the average duration of ed- gestive early findings have emerged. These include: ucation is 7.9 years. Having NCDs at ages 15-24 in India is correlated with 1.3 to 1.8 fewer years of ed- • In the UK, the prevalence of NCDs in children ucation seven years later (measured at ages 22-31). aged 7 years in 1965 was high. 6.6 percent of chil- dren suffered from chronic respiratory illnesses, • NCDs also have a significant detrimental effect 2.4 percent had a heart disease, and 15 percent on labor market participation and outcomes. In suffered from neurological, developmental, or the UK, people who had NCDs as children are 3.3 mental health disorders. Finally, 13.7 percent of percentage points less likely to be employed at age children reported having visual difficulties, and 23 and 4.9 percentage points less likely to be em- 4.8 percent had hearing difficulties. ployed at age 50. Education does appear to play a mitigating role. The coefficient on the interac- • In India, the figures are much lower. In 2012, tion term of having NCDs as children and years 0.12 percent of children aged 0-14 years reported of schooling is positive, suggesting that children having chronic respiratory diseases, 0.34 percent with NCDs who acquire more education are less suffered from neurological, developmental, and likely to be unemployed. mental health disorders, 0.30 percent had vision difficulties, and 1.63 percent had hearing diffi- • Results from India are similar to those observed culties. The low prevalence of NCDs may partly in the UK. Youth aged 15-24 who had NCDs are 8 reflect poor access and diagnostic quality in the to 11 percentage points less likely to be employed health sector, as well as the nature of the survey, seven years later (at ages 22-31). Adults aged 25- which relies on recall data (as opposed to direct 44 are also 7 percentage points less likely to be measurement in the UK). employed seven years later (at ages 32-51). • NCDs in childhood are strongly associated with • NCDs in childhood are associated with lower school absenteeism in the UK. Children with adult height in both countries studied. Children NCDs report 0.9 percentage points lower atten- with NCDs in the UK grow up to be 0.64cm dance rates at age 7 and 0.07 percentage points shorter at age 23 than their counterparts (the lower attendance at age 11, compared to their mean is 169.7cm). Children in India with NCDs counterparts. In India, most of the effects are seen are 0.66 to 2.57cm shorter, compared to their on the extensive margin. NCDs in childhood are counterparts (the mean is 142.6cm). correlated with a 15-percentage-point lower school enrollment rate (estimations with household fixed • These emerging correlations suggest that NCDs effects). Conditional on going to school, there is no in childhood and adolescence may substantial- significant correlation with absenteeism. ly lower subsequent adult productivity through their impact on education. • In both countries, having NCDs is associated with lower educational performance. In the UK, Discussion and future work: Results to date under- NCDs are correlated with 11 percent lower math score the urgent need to take NCDs in childhood scores at age 7, and with 16 percent lower math seriously, as they could have a large detrimental scores at age 11. Similarly, NCDs are correlated impact on human capital. Preliminary findings, with 10 percent lower reading scores at age 7 and although correlational, suggest that policies to mit- 23 igate the potential adverse impacts of NCDs in chil- the data from Mexico and Indonesia will be analyzed. dren on educational outcomes could substantially improve future productivity, when today’s children • Consider specific categories of NCDs, as opposed reach adolescence and adulthood. The research also to grouping them all together. looks at intergenerational impacts of NCDs and the potential role of education in this relationship. Spe- • Survey the literature for additional methods to cifically, it seeks to understand if education can mit- overcome the endogeneity problem and use best igate the transmission of adverse impacts of NCDs available practice for panel data. in adults to children in the household. This compo- nent of the research is ongoing. Highlights from workshop discussion: The lead discussant for the paper was Professor Sue Horton Limitations and caveats: While the preliminary (University of Waterloo). Professor Horton urged the findings point to the possibility of large adverse researchers to be circumspect in formulating results educational and labor market impacts of NCDs in and not “overpromise,” as some of the language used children, these findings are subject to several cave- in presenting early findings had appeared to suggest ats and need to be interpreted with caution. a demonstration of causality. (The researchers have subsequently worked to address these concerns.) • First, the estimates are correlational and do not rep- Professor Horton suggested that the researchers resent causal parameters. The panel data household might experiment with alternative models to incor- fixed effects approach in India alleviates this con- porate additional factors potentially affecting out- cern to some extent, but still leaves room for crit- comes, considering patterns among siblings, exog- icism (see Glewwe and Miguel (2008) for a review enous shocks, and other factors. Professor Horton of this literature). In addition, given the low preva- recalled the important work done at Sweden’s Kar- lence of NCDs, we are also left with little variation. olinska Institute on multi-directional interactions between parents and children with regard to NCDs. • Part of the reason for the low prevalence of NCDs Some participants emphasized the relevance of in the India data is probably that the data are the life-course literature, with its attention to critical self-reported. Self-reported data are subject to re- windows of vulnerability – certain periods that are call bias and measurement error. It is also possible more sensitive to shocks. Colleagues endorsed the that the reported positive cases are those of the value of carefully separating out the different kinds most-affected children, which could lead to large of NCDs, noting that the policy implications for and upward-biased estimates. distinct disease types may be very different. (The re- searchers have again taken this guidance on board.) • While data in both countries have high follow-up Finally, some participants urged that the researchers rates, we have not yet attempted to account for ultimately not decline to offer their own causal in- attrition over survey rounds, which could lead to terpretations, even though, with the available data, biased estimates. Preliminary analysis of the data they cannot establish the magnitude of impacts. from both the UK and India shows that those with NCDs are more likely to drop out of the survey, which will lead to downward-biased estimates. Key messages from research papers and workshop discussions • So far, we have grouped all NCDs into a single cat- egory. However, the biomedical literature suggests With the facilitation of the co-chairs, workshop partic- that different NCDs could have very different effects ipants reflected on the learning that had emerged over on educational attainment (for example, mental dis- the course of the event, both from the research papers orders vs. type I diabetes). More work is needed to and presentations and from group discussions. Each parse out the effects of specific NCDs of interest. participant was invited to formulate key learning points in writing. The ideas were then reviewed, grouped un- Future work: der a series of emerging thematic headings, and further discussed. Key topics and messages that garnered sub- • Assess the impacts of NCDs in parents on the edu- stantial consensus included the following: cational accumulation of children. This is also where 24 CHAPTER 1 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E Enhancing human capital through NCD control The power of a life-course approach will yield economic benefits for countries The focus on specific needs and interventions for Countries can reap substantial economic benefits adult survival underscores the power of an approach by strengthening their human-capital base through to health tuned to the stages of the life course. This action on NCDs. New research for this initiative includes gestation, infancy, childhood, and adoles- suggests the magnitude of possible gains and shows cence, with their specific windows of vulnerability that countries may obtain benefits within relative- and opportunity for developmental processes, as ly short timeframes (1-5 years). Key mechanisms well as the diverse stages of adult life. A life-course driving the gains include improvements in effective perspective is implicit in the analysis of human cap- labor supply and longevity. ital and further enhances human capital’s value as a frame for analysis and action on NCDs. Better human capital means higher productivity, short- and long-term Intergenerational effects Most economic benefits of improved NCD preven- As they affect multiple stages of the life course, so tion and control will stem from productivity gains. NCD impacts also cascade from one generation to Some benefits will be felt rapidly, as survival and the next. By definition, non-communicable diseases productivity rise among current workers. This ini- are not transmitted by infectious “bugs” from person tiative is finding evidence that, by reducing NCDs in to person. But NCDs’ negative economic and social children and caregivers, countries will also improve effects can spread within households and permeate human capital and boost productivity in future gen- neighborhoods and communities (Jan et al. 2018). erations, through improved educational outcomes. Indeed, one reason why NCDs are so important, in addition to their direct impact on income and well- The price of inaction will be high being, is that adults who are compromised by NCDs Successfully controlling NCDs will bring countries may have trouble executing their intergenerational major fiscal rewards through human capital gains. responsibilities to children and the elderly. In this Conversely, if countries fail to adequately protect way, NCDs cause human capital losses that perpet- their human capital from NCD epidemics, the uate themselves across generations. economic consequences will be severe—including soaring health expenditures and foregone govern- Increased efficacy through action across sectors ment revenues. Khadka’s and Verguet’s “poor-per- DCP3 and other research efforts have highlighted the forming” scenario has helped to quantify this threat. promise of intersectoral action in NCD prevention The up-front investments needed to manage NCD and control. NCD prevention strategies, in particular, burdens may be substantial, but they are dwarfed are vital to build and protect human capital. Priorities by the costs that loom, if necessary action is post- include improving child and adult nutrition, promot- poned (Bertram et al. 2018; Nugent et al. 2018). ing healthy mobility, and using excise tax measures to discourage consumption of health-damaging prod- Tackling NCDs can reduce poverty and inequality ucts, most prominently tobacco. For these and many and promote social stability other approaches, collaboration across government Analyzing 283 studies, Niessen and colleagues sectors is essential for high-impact action on NCDs. (2018) found overwhelming support for a positive Where feasible, a whole-of-government approach association between low incomes, low socioeco- may be the most effective way to make progress. nomic and/or educational status, and NCDs, in- Alignment and pooling of know-how among WHO, cluding: tobacco use, obesity, hypertension, cancer, the World Bank, and other partners at country level and diabetes. Global health equity gaps are likely to may facilitate the delivery of successful intersectoral widen further as the share of premature death and policies and the evaluation of their impacts. disability caused by NCDs rises in LICs and LMICs. Tackling NCDs is an effective way to improve hu- Toward an “adult survival revolution”? man capital among the least well off, reduce in- As this initiative’s commissioned studies are dis- equalities, reinforce social stability, and ensure that seminated and enrich the NCD and human capital countries fully harness talent from their whole pop- knowledge base, they may help to drive what some ulations, an advantage for competitiveness. workshop participants termed an urgently needed 25 “adult survival revolution.” UNICEF under James P. ular through prevention and control of NCDs. Grant spearheaded the Child Survival Revolution in The idea of a revolution in adult survival and the 1980s and 90s, with remarkable impact. Now, wellbeing could capture the imagination and cre- some call for a similar global campaign to focus sus- ativity of leaders around the world, and the human tained action on improving adult survival, in partic- capital lens could enable this transformation. Workshop Day 2, afternoon session The workshop’s final session was co-chaired by Sir trends, Wu and Jha argued that the main measure George Alleyne (PAHO) and Daniel Dulitzky of adult survival now included in the human capital (World Bank). Main discussion topics included mea- index—the likelihood that a current 15-year-old will surement issues linked to NCDs and human capital; die before age 60 (“45q15”)—may be suboptimal. directions for a future research agenda; and the im- The researchers cited evidence that many countries plications of project findings for key stakeholders, in- have registered sustained reductions in the likeli- cluding national governments, multilateral agencies, hood of adults’ dying between ages 50 and 69 (Figure civil society, and the private sector. 1.8). Global life expectancy is now almost 70, such that a large global population of older adults needs Refining measurement tools and strategies to be taken into account in analyzing health and hu- man capital—beyond what the 45q15 measure can The case for alternate metrics of adult mortality. accommodate. On this basis, Wu and Jha proposed, The analysis of adult mortality from Wu, Jha, and possible future versions of the HCI might consider colleagues in Paper 4 provided an entry point for complementing 45q15 with the 40q30 metric: the measurement discussions. Given global mortality likelihood of dying between ages 30 and 69. FIGURE 1.8 Trends in the risk of dying between ages 50 and 69 in 25 countries, 1970-2010 Source: Norheim et al. (2015). 26 CHAPTER 1 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E The HCI continues to evolve. Workshop par- Strengthening measurement systems at country ticipants appreciated that the HCI is not set in stone level. The discussion confirmed broad agreement among and that future enhancements may be possible and workshop participants on key principles connected with desirable. Some argued that, given the acknowledged NCD measurement in countries. NCD data systems difficulties in measuring human capital, a valuable should be country-owned and country-led, and reliance activity for the World Bank would be convening a on modeling should decrease. An important goal is to series of broadly inclusive discussions among glob- nurture NCD measurement systems at country level that al experts and stakeholders to systematically work are fully integrated with national statistical systems. through human capital measurement issues, partic- ularly as they relate to NCDs. Toward a future research agenda It was acknowledged that, regarding NCDs in the HCI, current concerns are less about the relevance of Act now, and keep learning. Participants agreed NCDs than the problem of obtaining a robust, single that countries do not need to await extensive further metric of NCDs, and the possibility of this being reliably research before ramping up their action on NCDs. sourced in all countries. Some participants emphasized Through resources such as the DCP3 and the WHO that the HCI is an advocacy tool, not a perfected scientif- Best Buys, an accessible evidence base exists that ic instrument. It is a means of organizing and commu- can guide foundational policy, programming, and nicating information that continues to evolve. The cur- implementation choices. Ultimately, however, more rent version, if used strategically, can help get countries research of several types will be needed, to ensure and key institutions thinking about NCD investments. that countries and partners get the best results as A new metric—the economic burden of disease. they work to strengthen human capital through Participants observed that research contributions pre- NCD control. sented to the workshop were helping establish the con- A systematic approach to setting research ceptual and methodologic foundations of what could be priorities. To chart directions for research, proven a major advance in global health measurement: a rigor- tools can be adopted, notably the evaluation frame- ous, comprehensive assessment of the global “economic work originally developed by Jamison and colleagues burden of disease.” Such a measure would complement for WHO’s Ad Hoc Committee on Health Research the established epidemiologic burden of disease, and Relating to Future Intervention Options (1996) and would provide crucial information to policy makers. subsequently refined (Jamison 2009) (Figure 1.9). FIGURE 1.9 Framework for analyzing health research needs Source: WHO Ad Hoc Committee (1996). For a revised version, see Jamison (2009). 27 The framework summarizes a systematic pro- better track and quantify these impacts and map cess that can enable analysts to identify and evaluate their pathways. The focus is on implementation research lines potentially suited to optimize existing research and research into the communication interventions, along with those that may yield new mechanisms to ensure reception by policy makers. interventions to tackle the segments of disease bur- dens that cannot be averted with current tools. • Put a price tag on passivity. Additional research Emerging research directions. While awaiting can sharpen assessments of the cost of inaction or of systematic analyses, participants anticipated possi- maintaining the status quo. Important to this effort ble broad directions for a future research agenda. will be improved measurement tools to capture the Not intended to be exhaustive, the discussion aimed impact of non-fatal outcomes and long-term care. to surface some areas of early promise for subse- quent systematic assessment: • Continue the quest for better metrics. Collaborative work should move forward to establish robust mea- • Measure human-capital gains from NCD action surement of the global economic burden of disease. at country level. Studies sponsored by this initia- In parallel, WHO and partners could lead a consen- tive plausibly predict substantial economic benefits sus process to construct improved NCD metrics. for countries that effectively prevent and control major NCDs. These analyses call for further em- • Close the research gap in mental health. Mental pirical verification. In partnership with countries, health conditions and effective strategies to address research should proceed to analyze the impact of them have been chronically neglected in research. NCD prevention and control on Human Capital This applies in particular to mental health challenges Index results in a set of countries with reliable data. in low-income settings. The human capital lens can help spur a new wave of research to: (1) better un- • Clarify NCD effects on productivity. The negative derstand the health and economic impacts of mental impacts of NCDs on productivity offer a potent health conditions (including their reflection in coun- line of argument for engaging ministers of finance tries’ HCI performance); and (2) accelerate the design, and private sector partners. Targeted research can deployment, and evaluation of effective interventions. RECOMMENDATIONS FOR KEY STAKEHOLDERS During the workshop’s concluding discussions, with the facilitation of Co-Chairs Sir George Alleyne and Daniel Dulitzky, participants framed a series of action recommendations for stakeholders in NCDs and human capital. National governments Act on the evidence that investments in NCD aware- ness, prevention, and management can: Develop a comprehensive national strategy to re- duce the impact of NCDs and the disability they • Increase GDP and productivity cause on human capital and the labor market, in- corporating: • Reduce health spending growth • A whole-of-government approach • Decrease health inequity • Tailored actions across the life course • Increase social security • An equity lens Commit to build decentralized data systems for country-level monitoring of NCDs and tracking of • Link to educational outcomes their human-capital impacts 28 CHAPTER 1 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E WHO and World Bank (joint action) incorporating life-course perspectives, with a fo- cus on NCDs • Expand the scope of investment cases to assess the impact of NCD control on human capital • Strengthen the analytical underpinnings of multi-sectoral lending operations by systemat- • Launch a new agenda of institutional collabora- ically assessing NCD impacts on human capital tion on NCDs and human capital, driven from and economic growth the highest levels of both organizations and an- chored in joint work in countries • Support countries to develop robust data systems for tracking and measuring progress on NCDs • Move swiftly to identify a group of countries that and integrating NCD data into broader health want to be forerunners in the NCD/human cap- management information systems ital effort; work with countries to develop initial plans, mobilize resources, and jumpstart action Private sector • Engage additional partners and jointly establish • Collaborate in documenting the effects of NCDs a TDR/CGIAR-type research program for NCD on human capital and productivity at the firm interventions, to “bend the cost curve” level, including issues such as absentee- and pre- senteeism. Collaborate with academic researchers WHO to design and test interventions to protect and en- hance workers’ human capital against NCD risks • Adopt and share ownership of the human cap- ital index (HCI). Ensure that all WHO country • Reward “bottom-up” innovation from within representatives are fully briefed on the concept of firms that generates promising strategies to pro- human capital and can inform ministers of health tect human capital and boost productivity in the on how prevention and control of NCDs affect a context of NCDs country’s human capital • Individual business leaders, firms, and industry • Disease divisions should use a common template or regional alliances: participate in multisectoral to demonstrate how their work contributes to initiatives to reduce NCD impacts on human countries’ human capital capital, including Human Capital Project efforts at country level • Highlight NCD prevention and control in univer- sal health coverage (UHC). Advocate impact on Civil society human capital as a criterion for including condi- tions in countries’ essential UHC benefits pack- • Advocate for a sustained NCD/human capital fo- ages. Ensure countries know that this criterion cus in the work of WHO, the World Bank, and argues powerfully for incorporating NCDs and other multilateral agencies, as well as in national related services in benefits packages government policies World Bank • Use social media to popularize the concept of preventing and controlling NCDs to protect and • Present the investment case for NCD action in enhance human capital. Engage people living Human Capital Project work at country level with NCDs as leading voices in this effort • Help countries to “know their NCD burden”; link • Utilize data on human capital in calls for account- with projections and impact on labor productivi- ability at country level. ty and economic growth • In developing a next iteration of the HCI, host consultations to surface the best approaches for 29 CONCLUSIONS AND WAY FORWARD Rising burdens of NCDs threaten countries’ human capital—their most precious resource for economic develop- ment in a fiercely competitive world, where cognitive skills, productivity, and adaptability are the watchwords of success. NCDs compromise countries’ present stock of human capital—notably through their impact on adult life expectancy and productivity. Research also suggests that NCDs in children or among their caregivers jeopardize future human capital formation, through potential negative impacts on educational performance. This under- mines the talent and productivity of rising generations, denying them the chance to achieve. This also means that the stakes for countries in suc- conditions. Through the Working Group 2 re- cessful NCD prevention and control are high, and port, early results from this initiative informed the the potential rewards immense. New research sum- High-Level Commission’s final recommendations marized in this chapter shows that tackling NCDs to the WHO Director-General, and the messages now can bring countries strong economic gains in he carried to the September 2019 United Nations relatively short timeframes, for example by boost- High-Level Meeting on UHC. ing effective labor supply and longevity. Additional Investigators completed their analyses and research for this initiative is finding evidence that ef- write-ups for the NCDs and Human Capital Re- fective NCD action can also drive improvements in search Initiative in the closing months of 2019. Re- future productivity and economic growth through sults were later published in peer-reviewed journals better educational outcomes. As these findings are and/or incorporated into ongoing research pro- confirmed and shared, more and more countries grams, notably at the University of Toronto, WHO, will be able to seize the opportunities. and the World Bank. The research summarized in The human capital lens holds power to trans- this chapter has informed the World Bank’s Healthy form global action on NCDs. Viewed through this Longevity Initiative (HLI). Several of the investiga- lens, in particular by finance ministries, NCDs tors whose work is presented in these pages have cease to appear as a parochial worry for the health extended their inquiries and produced additional sector. They show themselves for what they are: a scientific outputs under the HLI. potent threat to countries’ economic future—and An unprecedented window exists for the World an opportunity for countries to reap exceptional Bank, WHO, and other partners to join forces in economic benefits while improving health. supporting country efforts to build and protect hu- Following the Toronto workshop described man capital through NCD prevention and control. in this chapter, the NCDs and Human Capital Re- The human capital agenda offers a clear rationale for search Initiative entered its concluding phase. In such partnership and provides a natural platform July 2019, a summary of its findings was submit- for collaboration at country level and globally. The ted to the WHO Director-General’s High-Level deadly interaction between NCDs and COVID-19, Commission on NCDs. Initiative results and rec- and the health and economic losses that people and ommendations were reflected in the final report countries have suffered as a result, make this agen- of the High-Level Commission’s Working Group da more urgent than ever. Joining forces to advance 2, on universal health coverage (UHC). One of the country gains in human capital, the World Bank, Working Group’s top policy recommendation urged WHO, and partners can power an adult survival countries to adopt “impact on human capital” as a revolution and youth capability transformation that criterion in setting essential health benefits pack- will drive gains in productivity, longevity, and qual- ages. This means giving an important place to the ity of life where they are needed most. prevention and control of NCDs and mental health 30 CHAPTER 1 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E REFERENCES Initiative commissioned studies 6. Deaton, Angus. 2006. “Global patterns of income and health: 1. Arcand, J.-L., and D. Rinaldo. 2019. “Health capital, economic facts, interpretations, and policies.” Technical report. National Bu- growth and the burden of disease.” Draft paper and PowerPoint reau of Economic Research, Cambridge, MA. presentation. Non-Communicable Diseases and Human Capital 7. Flabbi, L. and R. Gatti. 2018. “A primer on human capital.” Policy Analytic Work and Key Messaging Workshop. Toronto, University Research Working Papers. World Bank, Washington, DC. of Toronto Dalla Lana School of Public Health, July 9-10, 2019. 8. Glewwe, P., and E. Miguel. 2008. “The Impact of Child Health and 2. Han, S., J.A. Lauer, and J.-L. Arcand. 2019. “The human capital im- Nutrition on Education in Less Developed Countries.” Handbook pacts of investing in interventions to reach the SDG-related tar- of Development Economics 4. gets for non-communicable diseases: an analysis of 10 low- and 9. Grossman, M. 1972. “On the Concept of Health Capital and the lower middle-income countries.” Draft paper and PowerPoint Demand for Health.” Journal of Political Economy 80 (2): 223-255. presentation. Non-Communicable Diseases and Human Capital 10. IHME. 2018. Global Burden of Disease 2017. Seattle, WA: Institute Analytic Work and Key Messaging Workshop. Toronto, University for Health Metrics and Evaluation. http://www.healthdata.org/ of Toronto Dalla Lana School of Public Health, July 9-10, 2019. gbd/gbd-2017-resources. 3. Khadka, A., and S. Verguet. 2019. “Estimating the longevity 11. Jamison, D.T. 2009. “Cost effectiveness analysis: concepts and benefits of non-communicable disease mortality reductions in applications.” In The Oxford Textbook of Public Health, 5th edition, low-income countries over 2017-2030: findings from a model- vol. 2, edited by R. Detels, R. Beaglehole, M. Lansang, and M. Gul- ing study.” Draft paper and PowerPoint presentation. Non-Com- liford, 767-782. Oxford: Oxford University Press. municable Diseases and Human Capital Analytic Work and Key 12. Jamison, D.T., L.J. Lau, and J. Wang. 2005. “Health’s Messaging Workshop. Toronto, University of Toronto Dalla Lana contribution to economic growth in an environment of partially School of Public Health, July 9-10, 2019. endogenous technical progress.” In Health and economic 4. Roder-DeWan, S., O. Pandey, and A. Mohpal. 2019. “NCDs, educa- growth: findings and policy implications, edited by G. Lopez- tion, and human capital: how do NCDs interact with education to Casanovas, B. Rivera, and affect human capital accumulation?” Draft paper and PowerPoint L. Currais, 67-91. Cambridge, MA: MIT Press. presentation. Non-Communicable Diseases and Human Capital 13. Jan, S., T.-L. Laba, B. Essue, et al. 2018. “Action to address the Analytic Work and Key Messaging Workshop. Toronto, University household economic burden of non-communicable diseases.” of Toronto Dalla Lana School of Public Health, July 9-10, 2019. Lancet 391: 2047–58. 5. Wu, D., J. Veillard, D. Watkins, R. Hum, D.T. Jamison, and P. Jha. 14. Norheim OF, Jha P, Admasu K, Godal T, Hum RJ, Kruk ME, Gó- 2019. “Measurement and determinants of NCDs and impact mez-Dantés O, Mathers CD, Pan H, Sepúlveda J, Suraweera W, on adult survival.” Draft paper and PowerPoint presentation. Verguet S, Woldemariam AT, Yamey G, Jamison DT, Peto R. 2015. Non-Communicable Diseases and Human Capital Analytic Work “Avoiding 40% of the premature deaths in each country, and Key Messaging Workshop. Toronto, University of Toronto 2010-30: review of national mortality trends to help quantify the Dalla Lana School of Public Health, July 9-10, 2019. UN sustainable development goal for health.” Lancet 385(9964): 239-doi: 10.1016/S0140-6736(14)61591-9. Epub 2014 Sep 18. Additional references 15. Nugent, R., M.Y. Bertram, S. Jan, et al. 2018. “Investing in non-com- municable disease prevention and management to advance the 1. Ad Hoc Committee on Health Research Relating to Future Inter- Sustainable Development Goals.”Lancet 391: 2029–35. vention Options. 1996. “Investing in health research and devel- 16. Pradhan, E., E.M. Suzuki, S. Martínez, M. Schäferhoff, and D. Jamison. opment.” Geneva: World Health Organization. 2017. “The effects of education quantity and quality on child and 2. Baird, J., C. Jacob, M. Barker, et al. 2017. “Developmental Ori- adult mortality: their magnitude and their value.” In Dis-ease Control gins of Health and Disease: A Lifecourse Approach to the Pre- Priorities, 3rd edition, volume 8, Child and adolescent health and vention of Non-Communicable Diseases.” Healthcare 5(1): doi: development, edited by D. Bundy, N. de Silva, S.E. Hor-ton, D.T. 10.3390/healthcare5010014. Jamison, and G.C. Patton. Washington, DC: World Bank. 3. Behrman, J. 1996. “The Impact of Health and Nutrition on Educa- 17. Pradhan, E., E. Suzuki, S. Martinez, M. Schäferhoff, and D. Jamison. tion.” The World Bank Research Observer 11(1). 2018. “The Effects of Education Quantity and Quality on Child and 4. Bertram, M.Y., K. Sweeny, J.A. Lauer, et al. 2018. “Investing in Adult Mortality: Their Magnitude and Value.” In Optimizing Education non-communicable diseases: estimation of the economic and so- Outcomes: High-Return Investments in School Health for Increased Par- cial benefits of scaled-up systems and services.”Lancet 391: 2071–78. ticipation and Learning, edited by D. Bundy, N. de Silva, S.E. Horton, 5. Dasgupta, P., R. Rodin, and F. Chaloupka. 2017. “Economic evalu- D.T. Jamison, and G.C. Patton. Washington, DC: World Bank. ation of childhood obesity: conceptual framework.” WHO Com- 18. Strauss, J., and D. Thomas. 1998. “Health, Nutrition and Economic mission on Ending Childhood Obesity. Working paper. World Development.” Journal of Economic Literature 36(2). Health Organization, Geneva. 19. Watkins, D.A., D.T. Jamison, A. Mills, R. Atun, K. Danforth, et al. 31 2018. “Universal Health Coverage and Essential Packages of Health Organization. Care.” In Disease Control Priorities, 3rd edition, volume 9, Disease 21. WHO. 2018. Time to deliver: report of the WHO Independent Control Priorities, Improving Health and Reducing Poverty, edited High-level Commission on Noncommunicable Diseases. Geneva: by D. T. Jamison, H. Gelband, S. Horton, P. Jha, R. Laxminarayan, World Health Organization. C. N. Mock, and R. Nugent. Washington, DC: World Bank. 22. Wu, D., E. P. Banzon, H. Gelband, et al. 2019. “Health-care invest- 20. WHO. 2013. Global Action Plan for the Prevention and Control ments for the urban populations, Bangladesh and India” Bulletin of Noncommunicable Diseases: 2013-2020. Geneva: World of the World Health Organization 98(1): 19–29. ANNEX 1.1 Toronto workshop agenda AGENDA Non-Communicable Diseases University of Toronto Monday, July 8th - Arrival of Participants and Human Capital Dalla Lana School of Public Health (DLSPH) Hotel: Hilton Toronto, Analytic Work and Key Messaging Workshop - Health Sciences Building, 155 College Street, Room 145 Richmond Street West July 9-10, 2019 HS208 (2nd Floor) Toronto ON M5H 2L2; Telephone: 416.869.3456 Time Description DAY 1: Tuesday, July 9 th AM Session: 8:15-9:00 Breakfast and Registration Co-Chairs: Daniel Dulitzky (WB) and Prabhat Jha (U of T) 9:00-9:15 Welcome and Overview of meeting objectives 9:15-9:45 Keynote Speaker: Sir George Alleyne (PAHO) Health, NCDs and Human Capital 9:45-10:00 Questions and Answers (All) 10:00-10:45 Paper 1: NCDs and labour market returns – Jeremy A. Lauer (WHO) Paper 2: NCDs and labour market returns, cross country regression 10:45-11:30 analyses – Jean-Louis Arcand (Graduate Institute, Geneva) 11:30-12:00 Discussants for Papers 1 and 2: Prabhat Jha and Sue Horton (U of Waterloo) 12:00-12:30 Questions and Answers (All) 12:30-1:30 Lunch and Coffee PM Session: Co-Chairs: Erica Di Ruggiero (U of T) and Daniel Dulitzky 13:30-14:15 Paper 3: NCDs and longevity – A long view: Stephane Verguet (Harvard) 14:15-14:45 Discussant: Dean Jamison (UCSF) 14:45-15:15 Questions and Answers (All) Paper 4: NCD mechanisms to impact human capital – 15:15-16:00 Daphne Wu (U of T), Dean T. Jamison & Prabhat Jha 16:00-16:30 Discussants: Alexey Kulikov (WHO) and Jeremy Veillard (WB) 16:30-17:30 Questions and Answers (All) Evening Reception and Dinner – Location: U of T Faculty Club 18:00-20:00 41 Willocks Street Vivek Goel, Vice President U of T; Remarks on Human Capital and Knowledge Generation Day 2: Wed. July 10th AM Session: 7:30-8:30 Breakfast Key Messages from Day 1: Sir George Alleyne (Health perspective) 08:30-09:00 & Dean Jamison (Economics perspective) Co-Chairs: Tim Evans (McGill, by phone) and Rachel Nugent (RTI) 9:00-9:45 Paper 5: NCDs and education effects – Aakash Mohpal (WB) 9:45-10:15 Discussants: Sue Horton 10:15-10:45 Questions and Answers (All) 32 CHAPTER 1 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E Time Description Day 2: Wed. July 10th AM Session: 7:30-8:30 Breakfast (continued) Toron Coffee Break 10:45-11:00 11:00-12:30 Key Messages and narrative from research papers: Alexander (Alec) Irwin (WB) and All 12:30-1:30 Lunch PM Session: Co-Chairs: Sir George Alleyne and Daniel Dulitzky Implications of research findings for WB operations: 13:30-14:15 Miriam Schneidman, Jeremy Veillard and Ayo Akala (WB) 14:15-15:00 Implications of research findings for WHO: Alexey Kulikov and Adriana Blanco (PAHO) 15:00-15:30 Next steps and timelines for analytical work and dissemination including Fall 2019 UN Meetings 15:30-16:00 Conclusions and final thoughts: Daniel Dulitzky, Prabhat Jha, Sir George Alleyne 16:00 Meeting concludes, departure for most participants 16:00-17:00 Writing workshop for researchers and writer (researchers and Alec Irwin only). ANNEX 1.2 Toronto workshop, list of participants Francisca Akala World Bank George Alleyne PAHO Jean-Louis Arcand The Graduate Institute, Geneva Adriana Blanco PAHO Sarbani Chakraborty Access Accelerated Pedro Conceicao UNDP Erica Di Ruggiero University of Toronto Daniel Dulitzky World Bank Timothy Evans World Bank / McGill University Vivek Goel University of Toronto Sue Horton University of Waterloo Alexander Irwin Independent consultant Dean Jamison University of California at San Francisco Prabhat Jha University of Toronto Alexey Kulikov WHO Jeremy Lauer WHO Aakash Mohpal World Bank Rachel Nugent Research Triangle Institute Miriam Schneidman World Bank Daniel Sellen University of Toronto Jeremy Veillard World Bank Stephane Verguet Harvard T.H. Chan School of Public Health Daphne Wu University of Toronto 2 CHAPTER 2 33 Towards a Framework for Impact Pathways between NCDs, Human Capital and Healthy Longevity, Economic and Wellbeing Outcomes Philip O’Keefe a and Victoria Haldane b a Australian Research Council Centre of Excellence for Population Ageing Research, University of New South Wales b Institute of Health Policy, Management and Evaluation, University of Toronto 34 CHAPTER 2 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E INTRODUCTION This chapter aims to outline a framework for examining the impact pathways between non-communicable dis- eases (NCDs), human capital (HC), and end outcomes - primarily healthy longevity, inclusive growth, and well- being. It supports the discussion of the various dimensions and their interactions with evidence from a selection of low- and middle-income countries and summarizes policies which may influence those pathways. While the pathways and their linkages with policy are complex, the framework aims to be as simple as possible, but suffi- ciently open to accommodate multiple factors. The emphasis to date of the Human Capital Proj- implications for their well-being, productivity, and ect (HCP) and Index (HCI) has been largely on the human capital trajectories (De Pue et al. 2021). years up to age 18. The Healthy Longevity Initiative The structure of the chapter is as follows. First, (HLI) complements this work by taking a whole life it outlines the major factors which would need to be course approach, with an emphasis on adulthood considered in any framework, including defining and, particularly, later years of life. The HLI also key concepts and highlighting some cross-cutting aims to complement the HCP by developing the themes. It then outlines a trajectory of human capital links between human capital and a wider definition across the life course, before briefly summarizing lit- of wellbeing above and beyond the health and work erature on human capital and growth. It then discuss- domains. As such, the HLI can be considered in part es the relationship between human capital and both an effort to extend the HCP across the life course, growth and country wealth before a more detailed something the HCP itself is aiming for as it evolves.1 consideration of the key channels through which In that context, the role of NCDs as important NCDs may impact inclusive growth. A discussion influencers of human capital outcomes in adult life of distributional and gender considerations follows. becomes more prominent than a sharp focus only It then briefly discusses the impacts of health, and on the earlier years of life. more specifically NCDs, on wellbeing, emphasizing While already an important social policy pri- the intrinsic importance of good health above and ority, the need for raising the profile of the healthy beyond its instrumental value. In the final section, longevity agenda has been made more pressing by there is a summary of policies and interventions that the disproportionate and ongoing impacts of the are likely to help minimize and manage NCDs and COVID-19 pandemic on people living with NCDs. their impact on HC accumulation, deployment, and Indeed, not only do those with NCDs such as dia- protection, and thus positively influence the end betes, hypertension, and cardiovascular diseases, and outcomes of concern. The chapter does not aim to go Chronic Obstructive Pulmonary Disease (COPD) into depth in specific subject areas, which are cov- have greater risks of morbidity and mortality from ered thoroughly in the background research on HLI COVID-19, but they are also impacted by disruptions presented in this compendium, including on eco- in health service delivery and ongoing care. Further, nomic burden of NCDs, whole-of-government and public health and social measures, while necessary whole-of-society approaches to NCD prevention to mitigate transmission of the virus, may disadvan- and management, gender, and behavioral aspects of tage or exacerbate inequities for those with NCDs or NCD policy, and labor market policies to promote older age groups. For example, physical distancing longer productive working lives. or ‘shielding’ of older adults or those with NCDs has 1   The HCP has recently introduced Human Capital Complementary Indicators (HCCI) which are a wider set of HD indicators beyond the HCI components that take greater account of adult outcomes. HCCI vary by country but typically include labor force participation rates by gender, youth unemployment/NEET rates, and life expectancy at birth. https://www.worldbank.org/en/publication/human-capital. CHAPTER 2 35 FACTORS IN DEVELOPING A FRAMEWORK, SOME DEFINITIONS, AND CROSS-CUTTING THEMES In thinking about impacts of NCDs and human capital on the end outcomes of healthy longevity, inclusive growth and wellbeing, the discussion aims to incorporate several dimensions. These include structural factors, which may relate to levels of development, stage of demographic transition, and cultural and social norms. Overlaid on those big picture contextual factors is the policy and institutional environment which will play an important role in shaping the accumulation, deployment and protection of human capital, the extent to which NCDs may compromise human capital, and how human capital feeds through to healthy longevity, inclusive growth, and wellbeing. Behavioral factors also play an important intermediating role between structural and policy factors and end outcomes. In addition, as important as the link between NCDs and HC is, the channels of influence of NCDs on the end outcomes considered in this chapter are not only through HC (see Figure 2.2). For the individual, the life course perspective on how these dimensions come together over time (and across genera- tions) is also an important element of the discussion. In considering these factors, it is useful first to define a few terms. The core concepts and their definitions for the purposes of our research are: • Healthy longevity is produced across the life such as “learning poverty”3. The health element of course and means avoiding death and serious human capital is commonly analyzed through life disability in middle age, enabling a high level of expectancy, or by some combination of mortality mental and social functioning through middle and morbidity indicators. In terms of measuring and older ages, and includes a socially connected returns to human capital, present value of future and reasonably pain-free, short period of time be- earnings is the most commonly used measure. This fore death. means that the standard measure of returns will have in-built bias towards men in the large majori- • For NCDs, we focus primarily on cardiovascular ty of developing societies due to lower female par- disease and diabetes, respiratory diseases, can- ticipation in market work and gender wage gaps. cers, and mental health. • Wellbeing as used in the chapter is a subjective • Human capital as understood within the HCP is, concept from psychology with three distinct di- “…the knowledge, skills, and health that people mensions: evaluative wellbeing (or life satisfac- accumulate throughout their lives, enabling them tion), hedonic - or affective - wellbeing (feelings to realize their potential as productive members such as anger, happiness, sadness etc.), and eude- of society.”2 The education/skills element of HC monic wellbeing (sense of meaning or purpose) (often called knowledge capital) is usually proxied (Steptoe, Deaton, and Stone 2015). It is a key end by years of schooling (quantity), or where data are goal of good health and human capital, as it is of available educational outcomes (quality) repre- growth and the overall development process. sented usually by test scores, and skills by years of work experience to proxy on-the-job skill acquisi- While the HCP, like many other sources and consis- tion. Increasingly, some quality-adjusted measure tent with the foundational work of Gary Becker, in- which captures actual learning is preferred (e.g., cludes health within the definition of human capital, Hanushek and Woessman 2020), reflected for ex- this is not universally the case. The OECD for exam- ample in the HCP use of quality-adjusted years of ple defines human capital as “…the knowledge, skills, schooling based on harmonized international test competencies and attributes embodied in individu- scores (Angrist et al. 2021) and other measures als that facilitate the creation of personal, social and 2   Health is both an element of human capital as well as an input to producing other forms of human capital (Bleakley 2010), and the same is true for education and skills as influencers of the health element of human capital (Grossman 2015; Heckman et al. 2016). 3   This is a recent concept from the World Bank and UNESCO Institute for Statistics. “Learning poverty means being unable to read and under- stand a simple text by age 10. The indicator brings together schooling and learning indicators: it begins with the share of children who have not achieved minimum reading proficiency (as measured in schools) and is adjusted by the proportion of children who are out of school (and are assumed not to be able to read proficiently).” https://www.worldbank.org/en/topic/education/brief/what-is-learning-poverty. 36 CHAPTER 2 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E economic well-being” (OECD 2019).4 While health This is separable from the well-known socio-eco- could be included in “attributes”, it is at best hidden nomic gradients for HC overall. Poorer people on in such a definition, which is more akin to knowl- average have higher incidence of most common edge capital. Beyond the definitional, a considerable NCDs (Dalstra et al. 2005; Oshio and Kan 2019), amount of literature speaks to only the knowledge lower educational attainment and on-job skills and skills dimension of human capital when discuss- acquisition, and returns to human capital that di- ing the concept and its economic implications, effec- verge further from their peers over their shorter tively eliding knowledge capital with human capital. life cycle (Deaton and Paxson 1998). While most evidence on socioeconomic status (SES) gradi- Cross-cutting themes: In the discussion that fol- ents is from developed countries, increased ev- lows, there are several cross-cutting themes which idence from developing countries suggests that recur and are useful to highlight at the outset: sharp socioeconomic gradients can also be seen for both prevalence and risk factors in develop- • First, that there are multi-directional relationships ing countries. In fact, both income and education between NCDs and HC, and between them and gradients for NCDs have been found to be even healthy longevity, inclusive growth, and wellbe- more pronounced in low-income countries than ing. Crudely, HC and NCDs have a two-way im- in middle-income countries (Hosseinpoor et al. pact on each other, and human capital elements 2012; Malta et al. 2013; Williams et al. 2018; Al- may be affected by levels and rates of growth, len et al. 2017; Sommer et al. 2015). Apart from healthy longevity, and wellbeing. Given these SES, an important dimension of inequality when complex relationships, there has been sustained considering NCDs, human capital and healthy debate on the extent to which relationships be- longevity is gender (see section on inequality). tween these variables are causal or only cor- relates, though progress has been made on clar- • A fourth insight is that the economic impacts of ifying some dimensions of the question. improved health will vary by the stage of demo- graphic transition, which in turn is closely related • A second insight is that, in addition to the individ- to the level of economic development. Reductions ual trajectory of human capital and NCDs, there in NCD prevalence will help improve longevity in are important inter-generational dimensions. The all countries, but there are variations across stages interactions between the NCD and human capital of demographic transition and levels of develop- profiles of parents and their children are significant. ment at what stage that translates to lower fertili- This applies not just in transmission of HC and di- ty, increased HC investment in children, and oth- rect well-being effects of health from parents to er factors. The differential impacts across stages children when the latter are young. It also applies of demographic transition are observed in overall as parents age and their co-morbidities and dis- growth, savings rate effects, and other factors.5 ability require increased care from adult children, impacting the returns to their human capital, and • Finally, the structural and institutional environ- their mental health and self-reported quality of life ment of countries matters for the scale (and even (Lambert et al. 2017; Yiengprugsawan et al. 2022). direction) of the relationships between human capital, NCDs, healthy longevity, growth, and • A third important consideration is distributional. wellbeing. This includes institutional factors of Within countries, there are clear socioeconom- a more foundational nature, such as culture and ic gradients for incidence of NCDs, with their attitudes, but also underlying institutions such as impact on both HC and directly on well-being. legal systems or protection of property rights (Ac- 4   Measurement of all these attributes will be incomplete, especially when looking at a wide range of developing countries. As a result, as OECD notes, “…there is no comparable and consistent measure across countries reflecting all these elements available.” (OECD 2019). 5   The stages of demographic transition used are from the Global Monitoring Report 2015 of World Bank. They are: (i) pre-dividend countries with high-fertility (all LICs); (ii) early-dividend countries where fertility rates have started to fall and changing age structures are conducive to growth (a mixture of LICs and LMICs); (iii) late-dividend countries which have had rapid fertility decline and working age population share will shrink over coming decades (mostly UMICs); and (iv) post-dividend countries with high elderly shares and fertility rates below replacement for several decades (mostly HICs). CHAPTER 2 37 emoglu, Gallego, and Robinson 2014). These are on human wellbeing should not be lost sight of. examples of the structural and social factors in the Indeed, wellbeing is a key underlying objective outer circle of Figure 2.1. But it also includes more of the development, including economic growth, proximate policies and institutions such as social process. The wellbeing impacts of health are dis- security systems and labor market policies and in- cussed later in this chapter. stitutions, which have differing impacts on the re- turns to human capital across and within countries The life course trajectory of human capital and the degree of protection provided to a given stock of human capital. Another is the built en- Given the life course emphasis of the HLI, it is use- vironment and the extent to which that facilitates ful to think of a “normal” human capital trajectory or constrains labor force and social participation across the life course (Figure 2.1). The broad dimen- for those with functional limitations. It may also sions of the HC trajectory are formation (also called include wider issues not traditionally considered accumulation), deployment (which in this chapter in the social sectors such as climate change. Apart refers both to utilization of HC and extending the from the aggregate influence of structural factors, years over which HC is deployed), and protection. they also play a role in promoting and sustaining Like other forms of capital, HC is also subject to de- inequalities and discrimination in accumulation, preciation, a process which may be accelerated by deployment, and protection of human capital, e.g., NCDs or structural factors such as rapid techno- along gender or racial lines. logical change. The formation of human capital is most intensive during childhood and adolescence, • A final important point is that there are signif- but also includes ongoing accumulation of HC in icant wellbeing benefits from investments in adulthood (for example through on-job skill ac- health capital which do not show up in standard quisition). Deployment is considered here first as measures of returns to HC, which rely primarily realizing returns to HC but extending the period on labor earnings. Such non-pecuniary benefits, of deployment – for example through minimizing starting with the intrinsic value of being alive and NCDs, or policies to extend productive working in good health, have direct bearing on individu- lives – is also important. Protecting HC may take al and societal wellbeing which are independent various forms, the most pertinent for the HLI be- outcomes of concern. These have received less at- ing preventing or reducing the burden of NCDs, tention in the literature, though this is changing but also including policies such as social protection in recent years (Steptoe 2019). A starting point of programs which facilitate HC formation or prevent the HLI and this chapter is that health at the indi- and mitigate the depreciation of HC. Protective pol- vidual level, and the (increasingly NCD-driven) icies may also permit a longer period of deployment burden of disease at societal level, are themselves of human capital, with resultant economic and oth- important measures of wellbeing, over and above er benefits. The final stage of the life cycle is one of their impacts on other outcomes such as produc- declining HC accumulation, and more rapid depre- tivity and growth. While a good deal of the dis- ciation of human capital, though even at that stage cussion that follows looks at evidence on the eco- interventions to delay rapid depreciation have an nomic impacts of NCDs, this fundamental point important role to play. 38 CHAPTER 2 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E FIGURE 2.1 Human capital life course trajectory School to work & Mid career to end of Early years Younger elderly Older elderly early career working life (0-14 years old) (70-79 years old) (80+ years old) (15-35 years old) (36-69 years old) Source: Original figure for this publication Good health, and particularly reducing the in- erated basis due to illness and mortality over the life cidence and consequences of NCDs, will be a key course.6 Recent work aims to calculate dollar values driver of slowing human capital depreciation across of national health capital, finding that health capital this cycle. Economic benefits include both standard per capita has increased faster than GDP per capita returns to human capital, which focus on earnings, globally (Arcand and Rinaldo 2019). and the non-market contributions of older people, Whether looking at human capital narrowly as including their indirect contributions through facil- knowledge capital, or more broadly also to include itating the accumulation and deployment of human health dimensions, NCDs can negatively impact hu- capital by other family members, e.g., through pro- man capital at all stages of the trajectory. Conversely vision of care to grandchildren. their prevention, early detection and effective man- The human capital life course trajectory is more agement can help maximize the accumulation and commonly thought of with respect to knowledge deployment of HC and avoid or delay its depreci- capital, with poor health impacting its accumula- ation. The impact channels and evidence on them tion, deployment, and depreciation. However, econ- are discussed in the following sections of the chap- omists also describe health capital in similar terms, ter. In short, NCDs in childhood and adolescence with health as a form of investment of labor and (both NCDs of children themselves and the NCDs commodities in healthcare with returns in the labor of parents impacting their children) will negatively market and beyond (Mushkin 1964; Becker 1964) impact human capital formation, with consequent and viewed as “a durable capital stock that produces negative impacts on its deployment in adult life. an output of healthy time” (Grossman 1972; Arrow, NCDs that manifest in adulthood will not impact Dasgupta, and Mumford 2014). It depreciates with initial HC formation (or “HC stock at entry” to the age, either with natural decline or on a more accel- labor market), but may impact HC deployment, 6   Health is also viewed as interacting with other forms of capital. One dimension is the role of social capital (or social connectedness) as an important determinant of health (Turner 2003), and another strand of research focuses on the role of cultural capital in shaping the interactions of people with the healthcare system (Shim 2010). CHAPTER 2 39 may compromise its further accumulation during distinguishes developed and developing countries is working life, and will accelerate its depreciation. that under-development and/or under-coverage of During working age, the NCDs of older parents may public pension systems in poorer countries will usu- also negatively impact HC deployment of working ally result in a period of HC deployment which ex- age adult children through care responsibilities that tends till much closer to death, as people are obliged lessen labor market participation and opportunities to work until they lack the capacity to do so. This for on-job skill acquisition. NCDs of working age is the case even though the physical environments parents may also have additional negative effects on and labor policies that developed countries offer for HC accumulation for partners, children, and grand- people with NCD-related conditions and disabil- children. They will also compromise the ability of ities may allow for longer productive working lives older parents to provide care to others. And later in than are possible in less enabling environments. A life, NCDs are likely increasingly to accelerate health third caveat is that the characterization of older age (and knowledge) capital depreciation. This depreci- as largely a period of decline is overly simplistic, ation may in turn be delayed or mitigated through a increasingly so as healthy years of life expectancy range of medical/technological interventions, good rise and people at any given older age tend to have nutrition, social connectedness, and other factors, better average cognitive performance over time, as so that human capital preservation is a further im- healthy longevity rises.8 A growing body of evidence portant dimension to consider as life progresses.7 also finds that frailty and cognitive decline are more The various impacts of NCDs across the life course malleable at older ages than historically assumed. will ultimately compromise economic growth, In addition, a facilitating external environment can through the human capital channel and macro and contribute to higher functioning at any given level of fiscal channels. They will also compromise individ- morbidity. Related to this point, a fourth qualifica- ual wellbeing, both directly and indirectly. tion is that most estimates of HC returns fail to val- The stylized trajectory of Figure 2.1 comes with ue non-market contributions, which will truncate a number of caveats, some relevant to comparisons or dilute the HC deployment period, failing to value across countries and others to groups of people activities such as informal care provision, volunteer within countries. First, in developing economies the work, etc. A final point, related to the fourth, is that “normal” life course may be considerably shorter, this “normal” pattern has a gender bias to the extent with the stages of younger and older old age absent that women are unable to engage fully in market ac- or more truncated. Global life expectancy at birth tivities during their adult lives. This last bias shows was just over 73 years in 2019 and had risen by six up in most estimates of human capital wealth, with years in the first two decades of the century (WHO World Bank for example estimating that women 2019) but in many countries it is considerably lower account for less than 40 percent of global human – s low as the early-mid 50s for some LICs. At the capital wealth due to lower labor force participation, same time, global life expectancy at 60 was 83 years lower hours of work when participating, and gender in 2019 – a highly relevant complementary indicator wage gaps (World Bank 2018). Gender aspects are when discussing longevity – and healthy years of life discussed in section on inequality. expectancy have grown steadily over much of the world (Salomon et al. 2012). A second caveat which 7   There is also growing frontier research into not only human capital preservation but active reversal of the ageing processes which contrib- ute to health and cognitive decline (e.g., stem cell and methylation research). While important work, it is not considered here due to the likely lack of current relevance for the developing world. 8   The US National Academies of Medicine’s Global Roadmap for Healthy Longevity has a useful summary on the evolution of health and cognitive performance at older ages (NAM 2022). 40 CHAPTER 2 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E HUMAN CAPITAL, GROWTH, AND COUNTRY WEALTH There is a huge literature on the determinants of economic growth and the role of human capital in the growth pro- cess. Within this, there is ongoing debate between those who view human capital – largely with a focus on the skills dimension - as foundational for economic growth and those who view it as proximate (North and Thomas 1973). While some authors argue that human capital itself is foundational in the growth process (Easterlin 1981; Glaeser et al. 2004; Gennaili et al. 2013), others argue that the foundations of growth are institutional (or found in other struc- tural factors such as geography or culture), and that human capital is a proximate determinant of growth which is determined by institutions (e.g., Acemoglu, Gallego and Robinson, 2014).9 Estimates of the level of human capital’s contribution to growth vary significantly between the two approaches, with those arguing that human capital is a proximate cause finding much lower returns to human capital (and at levels which are aligned with findings from microeconomic studies).10 Nonetheless, both approaches find significant impacts of higher human capital on growth. It is also important to emphasize the limitations of focusing solely on growth in country income or wealth, which neglect wider dimensions of wellbeing which are important in considering the benefits of good health and education. Related to the large literature on human capital and wealth between 1990 and 2014 was highest in LICs growth is research on the cross-country shares of and UMICs, at around 80 percent in LICs as against factors of production in country wealth. This is a 60 percent or lower in other income groupings. useful complement to a focus on economic growth The World Bank estimate of human capital in na- by looking at changes in the underlying asset bases tional wealth is consistent with other global estimates of countries, including accumulation, investment, which seek to look beyond economic growth alone. depreciation, and depletion of different elements One example is the estimates in the various Inclusive of national wealth. As a result, it focuses squarely Wealth Reports (IWR) from UNEP which also aim to on the sustainability of growth, and also has a rela- look at all sources of wealth in an economy, including tionship to other work under HLI on the economic inter-generational dimensions. Not dissimilar to the burden of disease. A recent example of the research CWON estimates, the 2018 IWR estimates that hu- on the role of human capital in country wealth is man capital accounted for 59 percent of average na- the Changing Wealth of Nations (CWON) from the tional wealth over the period 1990-2014. Of that to- World Bank (Lange, Wodon, and Carey 2018, and tal, 33 percent was estimated to be education-induced World Bank 2021).11 The importance of human capital and 26 percent to be health-induced capital. capital in the wealth of countries is striking and A final important point to note before getting rises sharply across levels of development. Globally into the discussion of human capital, NCDs, healthy in 2014, human capital accounted for 64 percent of longevity and growth is that there will be significant global wealth. In addition, the correlation of life ex- wellbeing benefits from investments in knowledge pectancy to the per capita growth in human capital is and health capital which do not show up in stan- stronger than any other factor (Lange, Wodon, and dard measures of returns to HC which rely primar- Carey 2018).12 But the variation across levels of de- ily on labor earnings. These will have direct bearing velopment is striking, with the human capital share on individual and societal wellbeing which is an end of wealth for LICs, LMICs, UMICs and HICs at 41, outcome of concern. The wellbeing impacts are dis- 51, 58, and 70 percent respectively, emphasizing the cussed in the section on welfare effects. increasing importance of human capital in national The following sections review the literature first wealth as countries grow richer. At the same time, the on the relationship between NCDs, longevity and contribution of human capital to growth in national growth, and then between education and skills and 9   The debate helps clarify that human capital, like many other things, has important underlying institutional determinants. And those insti- tutional determinants at a given point in time have their own determinants, including lagged human capital, as e.g., the effect of educa- tion on political institutions that affect future human capital. This is the societal counterpart of the intra-family inter-generational above. 10  Acemoglu et al. argue that the much higher estimates on returns to human capital in studies which treat it as foundational are driven by omitted variable bias, with the human capital variable reflecting some of the effect of institutions or proxying some unknown variables. 11  The components of national wealth used in the report are produced capital and urban land, natural capital, human capital, and net foreign assets. 12   The other factors modelled include schooling levels, labor force share and growth rate, population growth, government spending, public investment, trade, and inflation. CHAPTER 2 41 growth. The channels for NCD impacts on growth and growth. At the same time, cross-country work are then discussed, notably the savings channel, fiscal by Kotschy and Bloom (2023) points to the negative channel, and HC channel. The following section then impacts of societal ageing on growth rates as the de- summarizes evidence of the reverse impact of educa- mographic transition progresses. tion and work on health and health behaviors, bring- There is growing evidence that the impact of im- ing out the bidirectional influences of NCDs and HC. proved societal health (proxied by increased longev- ity) on human capital accumulation and ultimately The relationship of NCDs, growth differs to according to what stage countries longevity and growth are in the demographic transition to lower fertility rates, and again post-transition according to whether There is a substantial literature on the impacts of they have reached a very high level of longevity (Cer- health on levels and rates of growth, using macroeco- vellati and Sunde 2009, 2015; Bhargava et al. 2001; nomic, growth accounting, and microeconomic ap- Hansen 2013; Desbordes 2011). The relationship be- proaches (Jack and Lewis 2009). Two caveats to bear tween life expectancy and GDP per capita exhibits in mind in assessing the literature on the relation- an inverse U-shape and countries tend to follow that ship between NCDs and growth are, first, that much curve over time. Bloom, Kunh, and Prettner (2018) of the work has been done on overall health rather summarizes the intuition of this literature as follows: than NCDs per se, and second that a good deal of “In pre-transition economies, greater longevity is not the research does not control thoroughly for under- associated with greater educational attainment or a lying non-health determinants of growth (the latter reduction in the birth rate. As such, increased surviv- applying similarly on the relationships between pre- al rates translate into a higher net rate of reproduc- ventable mortality or several disability and growth). tion”, which tends to be a drag on growth. In contrast, While the first is a clear limitation, the indicator “… in post-transition economies, greater longevity most often used in macro-level work is life expec- is associated with increases in various measures of tancy. While rising longevity is driven by a combi- education and consequently reductions in fertility”. nation of prevention and control of communicable However, at very high levels of life expectancy, the and non-communicable diseases, increasingly the relationship to growth seems to become negative burden of mortality is NCD-dominated, so that the again, as large amounts are spent “unproductively” gap between the economic impact of overall health in economic growth terms using expensive medical status and NCD-only health impacts is falling rap- technology on people beyond working years. idly over time and is already low in MICs and above. A second approach to exploring the causal nex- The other point to note is that the relationship may us between health and growth has been growth ac- be bidirectional. On the second caveat of failure to counting, which informs the macro approach with control for non-health determinants of growth, im- microeconomic estimates of the impact of health on portant examples would be institutional or structural individual productivity and may provide a firmer determinants (see above re Easterly and Acemoglu methodological foundation for understanding the critiques), failure to control for is likely to upwardly impacts of improved health on growth. The work of bias estimates of the contribution of health to growth. Weil (2005) and others (Ashraf, Lester, and Weil 2009 In work taking a macroeconomic approach, Hsieh and Klenow 2010) find positive, but consider- there has been considerable debate on the direc- ably lower, impacts of health on growth than some of tion of causality between better health and higher the macro work such as Bloom and others. The HCI country income, and to what extent health causes builds on the literature on the relationship between growth (Lewis and Jack (2009) have a useful sum- health and productivity through incorporation of mary of the diverse strands of literature in this de- the adult survival rate from age 15 to 60 as one of its bate). There are real challenges in demonstrating components and estimates that a 10 percent increase the channels and strength of the health to country in adult survival to age 60 results in a 6.5 percent in- income relationship, with technical challenges and crease in labor productivity (Kraay, 2018). variation in identification strategies, and differences The third approach of microeconomic studies in the scale and even direction of estimated impacts looks at household level health factors and their ef- of health on growth. Nonetheless, a summary of the fects on income. This is based on a clear relationship literature by Bloom, Kunh, and Prettner (2018) con- between individual health and incomes but does not cludes that “...the bulk of research identifies a pos- capture externalities or general equilibrium effects. itive, if often weak, causal link…” between health Microeconomic studies find positive effects of good 42 CHAPTER 2 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E health and nutrition on growth, most notably in ear- icant impacts on overall economic growth and em- ly childhood with high economic rates of return on ployment through negative effects on labor force par- investments. Nonetheless, microeconomic estimates ticipation (Png et al. 2016). Global studies of diabetes find impacts are “1-2 orders of magnitude smaller also point to negative impacts on growth; however, than cross-country studies” (Bleakley 2010). the drivers of these vary between HICs where nega- More directly, analyses on the macro-level im- tive effects on labor force participation drive impact, pacts of NCDs support a negative relationship be- as compared to MICs and LICs where premature tween NCD prevalence and economic growth. A mortality from diabetes drives the negative effects on 2010 study by Stuckler et al. calculating gaps in Mil- growth (Bommer et al. 2017). Studies of the impact of lennium Development Goals (MDG) achievement cancer on growth have similarly shown a negative ef- estimated that each 10 percent increase in NCD-re- fect through impacts on workforce participation and lated death in a population was associated with a 0.5 productivity (Beaglehole et al. 2011). Estimating the percent reduction in annual economic growth. Put impact of common mental illness on growth is chal- in a positive light, Han et al. (2019) in a study of 10 lenging particularly in LMICs, in large part due to a LIC and LMIC countries that a basic package of NCD paucity of data. However, there is growing evidence interventions for the “big four” plus mental health of the negative effect of mental illness on work force would result in GDP being 0.5 percent higher annu- participation and productivity, and in turn GDP ally (or 0.7 percent higher per capita) over a five-year (Knapp and Wong 2020; Kilian and Becker 2007). time horizon. Looking at specific NCD categories, in HICs, Suhrcke and Urban (2010) and Hyclak, Impact channels between NCDs, human Skeels, and Taylor (2016) find a negative causal effect capital and inclusive growth of cardiovascular disease on subsequent economic growth for 1960–2000 and 2000–2012 respectively. Looking beyond the overall NCD to growth rela- Bloom, et al. (2014) project substantial negative im- tionship above, this section reviews in more depth pacts on output in China and India from five major the evidence on the diverse impact channels of NCDs, primarily through negative effects on labor NCDs on productivity and growth. It starts with supply and capital accumulation. Similarly, research the savings and fiscal channels, then looks in depth from Singapore has projected that the growing prev- at the NCD-HC-productivity and growth linkages alence and increasingly early onset of type 2 diabetes across the life course. Figure 2.2 illustrates these amongst the working-age population will have signif- channels and their constituent elements. FIGURE 2.2 Impact channels between NCDs and inclusive growth Source: Original figure for this publication CHAPTER 2 43 i. The savings and investment channel tio for a group of 69 largely developing countries, though how different institutional and policy factors NCDs may impact savings behavior, which in turn mediate impacts is not well understood.15 In terms will impact investments in physical capital, a key of net effects on growth, any potential negative effect contributor to economic growth. The net impact on savings from increased societal longevity could on savings of increased longevity comes down to be offset by increased labor supply, extending the whether behavioral or compositional effects domi- pay-off period for earlier investments in human cap- nate. The behavioral effect is that increased longev- ital and reducing the period of dissaving.16 ity encourages higher savings at all ages as people An additional – though less direct and more anticipate longer post-work lives and hence periods speculative - channel is the impact of increased lon- of dissaving (Bloom, Canning, and Graham 2003; gevity on technological progress, which has increas- Kinugasa and Mason 2007; Lee, Mason, and Miller ingly been recognized as an important source of 2000; Doshi 1994).13 This beneficial impact on sav- growth since the work of Romer and others (Romer ings will be compounded in the earlier stages of de- 1990; Aghion and Howitt 1998). In such models, in- mographic transition by favorable dependency ra- creased longevity positively impacts technological tios. On the other hand, as the population stabilizes progress (and ultimately long-run growth) through with a higher share of elderly people, the positive be- higher savings which in turn drive down equilibri- havioral effect of people saving more across the life um interest rates and increase incentives to invest in cycle may be offset - and eventually dominated - by R&D. An additional way of looking at the impact on the higher share of people in the overall population R&D investment, linked to how addressing NCDs who are dissaving at older ages (the compositional increases human capital, is that parents who invest effect). Apart from these general effects, there may more in the education and health of their children also be more specific effects of NCDs on household help create a deeper human capital stock which is savings behavior. Intuitively, if lower prevalence ultimately a key input to the R&D sector (Baldanzi, of NCDs increases productive working years, and Bucci, and Prettner 2017). working years are ones of positive net savings, this effect may itself contribute to higher aggregate sav- ii. The fiscal channel ings. In contrast, there is evidence of the dissaving and financial stress impact of smoking (Greenhalgh While much of the discussion on the impact of et al. 2022), but also that smoking in LMICs lowers health and education on growth and productivity household level investments in human capital, on focuses on individual or household level effects, average across studies lowering education spending NCDs may also have fiscal impacts, both on the by around 8 percent (Do and Bautista 2015). expenditure and revenue sides. Higher NCD inci- For pre-demographic dividend developing dence is likely most obviously to increase the need countries, the net positive savings effects of in- for public spending on healthcare services, and creased longevity can be substantial and can persist also aged care services in countries where those for 50 years or more, as the experience of East Asia are financed in part by governments. How much demonstrates (Horioka and Terada-Hagiwara 2012; healthcare spending on NCDs increases relative to Li, Zhang, and Zhang 2007).14 The empirical analysis outcomes will be a function of the efficiency of that of Bloom, Canning and Graham (2003) suggests that spending (on which there has been major efforts to the positive effect on savings of increased longevity refine “best buys” and cost-effectiveness of inter- dominates the deterioration in the dependency ra- ventions by WHO and others). Those cost-effective 13   Though Bloom, Canning, and Graham (2003) find that the increased savings effect may be non-linear across the life cycle and was most pronounced up to age 65 in the early 2000s, suggesting that the positive impact of increased longevity on savings rates will be more pronounced in countries with initially lower life expectancy. The life cycle hypothesis underlying this effect may also be diluted where there are adequate and high coverage public retirement income systems. 14   Though the caveat on non-health determinants of growth and savings probably applies unusually strongly to East Asia. 15   There is also a gender dimension to this given the disproportionate burden on women for care of older frail parents and/or children with NCDs. To the extent this compromises their ability to undertake market work, it would result in lower income and ultimately savings. 16   This effect may be constrained in more formalized economies by mandated retirement ages or other incentives for retirement at particular ages. 44 CHAPTER 2 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E interventions, in particular to take steps from early iii. The human capital channel in the life cycle to reduce the incidence of NCDs, represent an important effort to bend the cost curve There has long been an intuition that greater longev- downward. To the extent that NCDs nonetheless in- ity leads to investment in more years of schooling in crease public expenditure and budget deficits, that the anticipation of a longer dividend period on the results in increases in public debt and requires high- initial years of educational investment – the so-called er interest payments to those purchasing the gov- “horizon effect” (Ben-Porath 1967). Increased lon- ernment bonds that finance the deficit (i.e., higher gevity is of course driven by a combination of com- bond yields). It may also lead to crowding-out of municable and non-communicable diseases, but in- private sector investment. While fiscal deficits need creasingly the burden of disease is NCD-dominated. not be negative for growth if debt finances produc- While micro-level analysis has found a robust causal tivity-enhancing investments in physical or human relationship between increased life expectancy and capital (and are a standard part of a Keynesian arse- higher human capital (e.g., Jayachandran and Lle- nal), the experience of 2022 has demonstrated that ras-Muney 2009), macro analysis differs between persistent rapid increases in fiscal deficits and resul- those who confirmed a similar positive causal rela- tant public debt may have serious consequences for tionship at the macro-level (Lorentzen, McMillan growth, inflation, and capital markets. and Wacziarg 2008; Bloom, Canning and Sevilla The other fiscal channel is on the revenue side. 2004) and those who argue that increased population The budgetary impact of a package of NCD pre- growth in response to increased longevity results in a vention programs can be significantly mitigated by negative causal relationship once changes in per cap- including taxes on tobacco products, alcohol, and ita income levels are taken into account (Acemoglu sugar-sweetened beverages (SSB), which produce and Johnson 2007). In any event, global life expec- revenues as well as longer term contributions to tancy has been on a steady increase for decades, from health and to productivity and growth (Summan et around 51 years at birth in 1960 to 73 years by 2019 al. 2020). As discussed in more detail below, tobac- (with a small drop due to the COVID-19 pandemic co taxes, in particular, have been shown to be highly after that). It is instructive to look at the impact of effective in reducing smoking, so reducing illness NCDs on human capital across the life course. This and death from a range of respiratory and cardi- allows us to understand how NCDs impact different ac diseases and cancers. There is also increasingly aspects of the human capital trajectory (formation/ strong evidence for the revenue and health benefits accumulation, deployment, depreciation, and deple- of taxing alcohol and SSB (The Task Force on Fiscal tion, with protection a common element across the Policy for Health 2019). For the longer term, to the life course) and with differing intensity. Annex Table extent that NCDs compromise workforce participa- 2A.1 is a stylized mapping of life course stages, hu- tion, that will reduce revenues from labor income, man capital trajectory, NCD patterns and impacts of either entirely where people fail to enter or with- NCD patterns on human capital across the life course. draw prematurely from the labor force, or at the The following sections discuss these relationships margin where they are working fewer hours over across the life cycle in more detail. They are summa- time. To the extent that NCDs compromise pro- rized below in Figure 2.3 across the NCD and human ductivity while at work rather than participation or capital life trajectories. Two important points to note hours worked, the negative revenue impact is likely are: (i) there are important interactions of NCD status to come through the firm profits and corporate tax and HC outcomes across generations in addition to channel rather than labor taxes. the direct impacts of NCD on the human capital of in- dividuals; and (ii) there is often a two-way relationship between NCDs and human capital which can make causal channels hard to determine with confidence. CHAPTER 2 45 FIGURE 2.3 Impact channels between NCDs and human capital across the life cycle Source: Original figure for this publication a. Gestation, early years, and childhood cluding cardiovascular and renal diseases, metabolic (Conception - 14 years) disorders including non-alcoholic fatty liver disease, metabolic syndrome and type 2 diabetes and chronic For both NCDs and human capital formation and lung disease in adulthood (Armengaud et al. 2021; its trajectory over the life course, pregnancy and Kruger and Levitt 2017). Distinct from IUGR, low early years of life are critical, both in the short-term birthweight is associated with an increased risk of and the long-term. While for most people NCDs NCDs in childhood and in adult life, including for only manifest in adulthood, pregnancy and infancy the latter diabetes, hypertension, cardiovascular dis- crucially influence the risks of NCDs in childhood ease, and neurological disorders.17 In addition, “… and adult life, with attendant impacts on human the associations between low birthweight and NCDs capital accumulation deployment and depreciation can be transmitted across generations, even in the (Heidari-Beni 2019; Shrimpton and Rokx 2012). absence of further adverse exposures, such as mater- Starting prior to birth, inter-uterine growth re- nal malnutrition” (Nyirenda and Byass 2019). striction (IUGR) affects 10–15 percent of all preg- In the early years after birth (particularly up to nancies worldwide, and may be of maternal, placen- age 3 but for this discussion 0-5 years old), NCDs tal, or fetal origin. Whatever the causes of IUGR, it and HC exhibit a mixture of short- and longer-term is increasingly clear that it makes affected infants impacts and inter-relationships. The first short-term more susceptible to a host of NCDs later in life, in- issue is the incidence of childhood NCDs. This is a 17   The relationship between low birthweight and diseases in adulthood is known as the Developmental Origins of Health and Disease (DOHaD) hypothesis (Hanson and Gluckman 2014). 46 CHAPTER 2 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E significant and growing health burden in itself, with Children with NCDs were also found to be less like- around 20 percent of childhood deaths and around ly to be employed as adults and to work less when 42 percent of DALYs due to NCDs (WHO 2016; Pro- they are employed, thus influencing future earn- imos and Klein 2012; Gore et al. 2011; Patterson et al. ings and on-job human capital accumulation. This 2019 re childhood and adolescent diabetes; Chaten- echoes work on negative impacts on educational oud et al. 2010 and Bertuccio et al. 2020 re childhood indicators for specific NCDs such as diabetes, asth- cancers; McCoy et al. 2005 re asthma). This represents ma and epilepsy (Wodrich and Cunningham 2008; an enormous loss of human potential globally. Meng et al. 2013), and of mental health conditions Childhood under-nutrition has major well-doc- such as ADHD (Currie and Stabile 2006). umented impacts on both NCDs across the life course Apart from the health and nutritional status and human capital formation and deployment. Over of children themselves, there are important associ- the long-term, early under-nutrition significantly ations between parental health and HC status and increases susceptibility to NCDs in adulthood, in- the health, cognitive and educational outcomes of cluding CVD, type 2 diabetes, obesity, and hyperten- children. There is an important gender dimension sion (Karokochuk et al. 2018; Shrimpton and Rockx of improved adult health in women and strong 2012; Black et al. 2013; Victora et al. 2008). It also has inter-generational spillover effects on children’s strong negative impacts on neurological develop- health, as well as an interaction of child health ment and cognitive performance, compromising the with education (Bhalotra and Rawlins 2011; Field foundations for learning in childhood and potential- et al. 2013). There is a well-developed literature on ly over the life course (Nyaradi et al. 2013). the correlations between overall maternal health Increasingly, interventions for promoting bet- and various child development outcomes, includ- ter childhood nutrition need to focus on the dou- ing birth weight, survival, cognitive development, ble burden of malnutrition, i.e., not only undernu- schooling performance, adult health, and produc- trition but also childhood overweight and obesity tivity (Bloom, Kuhn and Prettner 2015). While the which has become a growing epidemic in develop- impacts of paternal health on children’s outcomes ing countries due to changes in food systems, life- are less studied, there are obvious channels such as styles, and other factors (Abdullah 2015; Popkin loss of household income on children’s dietary di- et al. 2020). Childhood obesity is a direct cause of versity and nutrition, as well as documented nega- childhood NCDs and co-morbidities which usually tive impacts on children’s health from paternal risk manifest in adulthood (Pizzi and Vroman 2013; Fin- factors such as smoking (Oldereid et al. 2018). kelstein, Graham, and Malhotra 2014). It may also While the literature on the impact of NCDs contribute to poor socialization and poor learning. specifically on child health and human development Like under-nutrition, it is also a strong predictor outcomes is somewhat thinner, available studies on of both obesity and a range of NCDs in adulthood, parental NCDs and child outcomes are consistent in with associated health costs and lost productivity. finding negative impacts of maternal NCDs on child Finally, there are inter-generational effects with pa- health and human capital development (see Onar- rental obesity linked to obesity of children (WHO heim et al. 2016 for a systematic review). Jayachan- 2015 for a summary of evidence). dran and Lleras-Muney (2009) also provide evidence As children move into school years (around from Sri Lanka that increases in female adult surviv- 5-14 years old for this discussion), the interactions al rates led to differential increases in the education between nutritional status and the educational di- of girls over boys. The picture on mental health is mension of human capital formation become more somewhat more mixed, though meta-analysis of evident and are well documented. The literature on studies from developing countries finds that mater- childhood NCDs and educational performance is nal depression is associated with early childhood un- thinner but shows clear negative effects. Work for derweight and stunting in early years – and thus with this project finds that in India and the UK NCDs are NCD likelihood in adult life (Surkan et al. 2011), and correlated with lower school enrolment, for some in high income countries has negative effects for off- countries attendance, and strongly impact educa- spring health, behavior, and psychosocial function- tional attainment, the key human capital outcome. ing (Propper, Rigg, and Burgess 2007).18 For more 18   Paternal mental health during pregnancy, early years and adolescence also appears to affect the development of children, but contextual mediators strongly affect the strength of the association (Sweeney and Macbeth, 2016). CHAPTER 2 47 acute mental illnesses, the negative effects on child thus returns on a given stock of human capital), as mortality and morbidity are clear (Webb et al. 2005). well as further human capital accumulation during There is also a longer-run macro channel of impact working life. Reductions in NCD-related mortality through fertility and education, with the transition and morbidity in adulthood will thus raise the re- from a low equilibrium “poverty trap” of high fertil- turns to the productivity of educational investments ity and low educational investment in pre-dividend (Mushkin 1964; Bleakley 2010; Bouncekkine, de countries to a new situation of declining fertility and la Croix and Licandro 2002). NCDs in adulthood increased educational investments in early dividend may also accelerate depreciation of human capital countries and beyond (Bloom, Kuhn and Prettner in different ways. Some of these impacts are fairly 2015; Onarheim et al. 2016; Victora et al. 2008). self-evident and others less so. Mushkin categorizes In the formative years of life, the role of social the three impacts of poor health on labor product as assistance in both promoting human capital accu- due to death (resulting in loss of workers altogeth- mulation and protecting it has increased markedly in er), disability (lost work time) and debility (loss of the developing world over recent decades. In terms of productive capacity at work). The relative likelihood promoting human capital accumulation, there is solid and intensity of different impacts will vary over the and growing evidence of the positive impacts of social life course for the individual and over time for the assistance transfer programs (both food and increas- society as the average age of workforces increases. ingly cash) on a range of human development/capital Figure 3 summarizes some of the key channels in outcomes. This includes vaccination, use of maternal which NCDs may negatively impact accumulation, and child health services, child and maternal dietary deployment, and depreciation of knowledge capi- diversity and other nutritional outcomes, and school tal across the HC trajectory. These are discussed in attendance (Bastagli, Hagen-Zanker, and Sturge 2016; turn in this section. Annex 2.1 provides a more de- O’Keefe et al. 2022).19 These beneficial impacts have tailed mapping of the human capital trajectory and been found in both unconditional and conditional its interactions with NCDs across the life cycle. transfer programs.20 Given that such programs are of- This is also a domain where structural factors ten targeted on the basis of need, the benefits in most and labor market policies and institutions will mat- cases accrue in a progressive manner, though exclu- ter a great deal. For example, a health condition sion errors in targeting often limit programs having which may be completely disabling so far as ability as widespread an impact as intended. In terms of pro- to work and earn in one society may have more lim- tecting human capital, transfer programs have simi- ited or even negligible impacts in societies which larly helped to reduce negative coping mechanisms have accommodative workplaces and labor practic- such as school drop-out and child labor. es, or in occupations which are less physically de- manding. To the extent that NCDs impact physical b. Impacts of NCDs during school-to-work capacity, their impacts may for example be expect- transition and working ages ed to be more acute in low-income countries with (15 – 65 or 69 years) greater prevalence of manual labor in agriculture. In addition to the evidence on childhood NCDs and • The most obvious impact is the labor supply and their impacts, NCDs which manifest in adulthood human capital loss from premature mortali- negatively impact labor force participation (and ty due to NCDs.21 Given that over 40 percent of 19   See also Handa et al. (2018), Baird et al. (2014), García and Saavedra (2017), Petrosino et al. (2014), Snilstveit et al. (2016) for cash transfers. 20   There is long-running debate on the pros and cons of conditioning transfer programs. While there is evidence that conditioning has in some countries resulted in somewhat better human development outcomes than UCTs, in many developing countries there are issues of supply side availability of services, administrative capacity to monitor compliance with co-responsibilities and other factors which caution against strict enforcement of compliance. These supply-side challenges have led to increased interest in “cash-plus” transfer pro- grams where the transfer is accompanied by social and behavioral change programs and other complementary interventions, but where participation of beneficiaries is nudged rather than strictly enforced. 21   WHO estimates around 42 percent of NCD deaths occur before 70 years old, an age at which substantial shares of the labor force continue to work in low-income countries in particular. Of these, an estimated 1·7 million (4 percent of NCD deaths) occurred in people younger than 30 years of age, and 15·2 million (38 percent) in people aged between 30 years and 70 years (Bennett et al. 2018). For the EU, OECD estimates that premature death due to NCDs corresponds to a rate of about 200 per 100,000 population in the age group 15-64 (OECD, 2016). See also Sweeney et al. (2015) for estimates of the economic cost of premature NCD mortality for 12 largely developing countries. 48 CHAPTER 2 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E NCD deaths occur before age 70, they represent stance abuse, with strong relationships between a major loss in total labor supply and thus human poor mental health and unemployment. Across capital globally. The relative importance of this im- the OECD, people with mild to moderate mental pact channel is increasing over time as the average illness, such as anxiety or depression, are twice as age of labor forces rises. The economic impact of likely to be unemployed, and that can rise to as premature mortality is also projected to increase high as seven times with severe mental illnesses. more quickly over time relative to impacts from The stigma around mental health has been shown morbidity (Sweeny, Rasmussen, and Sheehan to impact hiring and job loss, and to involve a 2015). The value of that lost labor output (HC de- self-stigma element which constrains job search ployment) will vary according to labor force par- (Brouwters 2020). Substance abuse both increas- ticipation rates and un- and under-employment es the chances of being unemployed and is more rates (with significant gender differentials also) likely to occur among those who are unemployed, but is an obvious drag on productive potential creating a vicious cycle (Henkel 2011). Apart and a reduction in returns to the human capital of from mental health and substance abuse, the workers who die prematurely from NCDs. There is limited number of studies from developing coun- also an economic impact of not being able to pro- tries also show negative labor market impacts of vide care for children, partners, and older parents. specific NCDs such as diabetes (e.g., in Mexico, there was a 7.7 and 6.3 percent reduction in prob- • The next channel is reduced returns to human ability of being employed for men and women capital from NCD-based morbidity and disabili- respectively for those with self-reported diabetes, ty resulting in reduced labor input and/or worker though no significant impact on hours worked or productivity. Available estimates suggest that this wages, Seuring et al. 2019). is a notably higher economic impact than pre- Variations in approach and accounting make mature mortality from NCDs, though this is less rigorous comparison or aggregation challeng- likely to be the case in some of the poorest coun- ing, and the limited developing country evidence tries with the highest premature mortality rates points to significant diversity in the impacts of and lower productivity workforces (e.g., Sweeny, different NCDs on labor force productivity across Rasmussen, and Sheehan 2015). Reduced returns countries. The economic impacts of presenteeism to HC may take the form of complete non-partic- in particular have proven difficult to measure and ipation in the labor force (or early withdrawal), compare, with differing methodologies, variable unemployment, lower hours worked, absentee- impacts according to welfare regimes and other ism and “presenteeism” (being at work but less factors leading to highly variable estimates of im- productive due to NCDs), delayed return to work, pacts (Mattke et al. 2007; Cancelliere et al. 2011).23 and, for formal sector workers, sick leave. These A literature review concludes that “the biggest impacts may be measured in terms of DALY totals problem remains the lack of an established and or shares lost, or monetized output losses. While validated method to derive monetary estimates evidence is largely from developed countries, the of the cost of lost productivity” (Garrow 2016), economic losses through this channel appear to and Schultz et al. 2009 caution against assigning be substantial.22 There also appear to be gender dif- dollar amounts to presenteeism losses. Accepting ferences in how specific NCDs impact labor mar- these substantial caveats, one study for 12 mostly ket outcomes, with for example high BMI having developing countries finds the economic impacts more negative impact on women’s labor market of NCDs due to absenteeism and presenteeism in outcomes than for men, and diabetes having 2020 to be between 0.9 and 1.6 percent of GDP higher negative impacts for men (Chakraborty, annually for absenteeism, and between 2.6 and Wu, and Jha 2024). Focusing first on non-partici- 3.7 percent of GDP annually for presenteeism pation and unemployment, perhaps the best doc- (Sweeny, Rasmussen, and Sheehan 2015). umented NCD group is mental health and sub- The gender dimension of these impacts is also 22   One systematic review found that g less than 4 percent of studies from developing countries (Chaker et al. 2015). 23   Mattke et al. (2007) describe three broad approaches to measuring presenteeism: assessment of perceived impairment by employees; comparative productivity, performance, and efficiency measures; and estimates of unproductive time while at work. CHAPTER 2 49 important to note. Globally in developing coun- sistent health shock is due to the indirect effects tries, labor force participation rates of women are of reduced human capital accumulation after the significantly lower in paid employment than those shock (Capatina et al. 2020). Just as interestingly, of men. The participation gap is exacerbated by the the same analysis finds that two thirds of the im- gender wage gap. One accounting effect of this is pact of health shocks on inequality of subsequent that reduction in measured returns to human cap- lifetime earnings is due to the behavioral effects ital from NCDs among women overall will be sys- rather than the direct impact of the health shock tematically lower than for men. At the same time, itself.25 Both direct and indirect impacts would Bloom et al. 2015 show that healthier women are be expected to have a compounded effect where more likely to be in formal sector employment, with NCDs induce multiple health shocks over time. higher earnings and productivity than either infor- mal work or no market work. An important second • NCDs can also contribute to accelerated depre- order effect is that women in formal employment ciation of human capital. This may be primary/ tend to have fewer and better-educated children, so direct, where NCDs in the working age person re- the economic and social benefits of averting NCDs sult in long spells of unemployment or withdrawal are inter-generational. The gender inequalities in from the labor force. While not easy to measure, different dimensions of human capital are discussed carefully done work finds evidence of skill de- in detail in section on inequality below. preciation from periods out of work (Dinerstein, Megalokonomou, and Yannelis et al. 2020; Laureys • A third channel is the reduction in further accu- 2021). With respect to health capital, the impact mulation of human capital during working life of NCDs and poor health is obvious. Alternatively, from both the direct impact of NCDs on labor the impact may be secondary/indirect, whereby input and quality of labor input (and hence work absenteeism and presenteeism accelerate the nor- experience and on-the-job learning), and the in- mal depreciation in human capital that has been direct behavioral impacts on subsequent effort to observed to happen over time for healthy workers. accumulate human capital and changes to saving and other behaviors (Capatina et al. 2020; Smith • Depending on the severity of NCDs during normal 2004; Hokayem and Ziliak 2014). Firstly, with working age, there may be negative spillover effects respect to the direct impact, substantial human on other adult family members and children. For capital accumulation occurs on the job, through children, this is discussed above, and there are learning-by-doing and/or structured training and clear negative spillovers of parental NCDs on chil- skill upgrading (see Program for International As- dren’s human capital development. For prime age sessment of Adult Competencies (PIACC) publi- adults reducing work or withdrawing from the cations of various years24; Ma, Nakab, and Vidart labor force due to NCDs, the net impacts are less 2022). Secondly, the indirect effect on further hu- clear. There may be an added worker effect where man capital accumulation has been less studied other adult(s) in the household who were not in but appears to be substantial. The logic is that those the labor force prior to the NCD shock take on experiencing poor health or significant health market work to offset the loss of income from the shocks will anticipate reduced future labor sup- withdrawal of the breadwinner (Lundberg 1985) ply and thus lower future returns to their current but this depends on care needs and options. human capital investments, disincentivizing such investments. The extent to which health shocks c. Later life (66 or 70 years – death) from NCDs impact earnings over the life course is also likely to vary according to the chronicity The later stages of life present a quite distinct tra- and the unpredictability of symptoms of different jectory of both NCDs and HC, and their inter-re- NCDs (Heckman and MaCurdy 1980). Estimates lationships. From the NCD perspective, prevalence for the US show that around 40 percent of the loss increases, and co-morbidities are more frequent. in remaining lifetime earnings from a major per- ADL/IADLs are increasingly compromised for 24   Full database of publications available at: https://www.oecd.org/skills/piaac/publications/ 25   Though the analysis factors in access to social insurance in the US, which may have differential impacts on low skill workers than high skilled. 50 CHAPTER 2 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E many and may become disabling conditions be- is significant potential for lower returns to the exist- yond a certain point. The risk of age-related NCDs ing stock of human capital of adult family members such as dementia also increases sharply, as well as and stalled accumulation of further human capital. depression related to social isolation. From the hu- The impacts can be direct (care needs and time for man capital perspective, the educational/skills el- the older parent) and/or indirect (older parents no ement of human capital accumulation decelerates longer able to care for grandchildren and the care (and eventually stops) and depreciation accelerates, responsibility falling again to the “sandwich genera- while returns on human capital according to stan- tion” parent). These social dynamics are well under- dard measures diminish or cease with withdrawal stood anecdotally, but for developing countries have from the labor force (though this reflects the lim- only in recent years been better researched with itations of standards measures of returns to human the spread of HRS-style household surveys which capital rather than implying that older people out- quantify time and opportunity costs and benefits side the workforce make no contribution to their of informal care provision (e.g., CHARLS in China, households or society).26 Protection of the health SAGE surveys27 in India and other countries).28 The dimension of HC becomes ever more important, gender dimensions of care burden and the resulting aiming for compression of morbidity. under-utilization of women’s human capital is dis- The inequality dimension is also on average cussed more in section on inequality below. more acute, as the cumulative deficits of a lifetime for those at lower ends of the socio-economic gradient Impacts of human capital and work on come home to roost in the shape of higher co-mor- NCDs, overall health and health behaviors bidities and shorter life expectancy and lower in- comes (Deaton and Paxson 1998; OECD 2017), but There is more debate about the extent of causality also commonly having to “work till one drops” out in the other direction – from human capital (accu- of financial necessity, lack of old age financial protec- mulation, deployment and depreciation) to NCDs. tion from formal pension schemes and inadequacy of There is less evidence on the effects of education social assistance benefits, and lack of affordable for- on NCDs specifically in developing countries, but mal aged care or other unpaid care options. In poorer more evidence of education gradients and impacts countries, a positive factor is high multi-generation- on general health and NCD-related risk behaviors. al co-residence, which facilitates mutual caregiving The excellent summary by Grossman (2015) con- across generations, though tends to decline steadily cludes, “...many studies suggest that years of formal as countries get richer (Evans and Palacios 2015). schooling completed is the most important correlate The inter-generational element of the human of good health. There is much less consensus as to capital and NCD interaction is crucial at this life whether this correlation reflects causality from more stage. The most common negative impact of advanc- schooling to better health” (Heckman and Karapa- ing NCDs is diversion of adult children/children- kula (2019); Heckman, Humphries and Veramendi in-law from market work in order to provide care. (2016); Conti, Heckman and Urzua (2010); Cutler The obligation for care falls disproportionately on and Llera-Muney (2006). He reviews the literature women and has diverse impacts on their labor mar- which addresses these challenges in various ways: ket participation depending on prior status and the by introducing omitted variables; by examining en- extent to which care demands have major impacts vironmental factors (in twin studies); and through on capacity for market work or are more modest use of instrumental variables. As a starting point, and manageable (Moreira da Silva 2019). But there there are consistent education gradients with respect 26   The standard measures of economic output have well-known limitations in this regard, with non-market contributions of older people (such as caring for grandchildren or volunteering) not being valued either directly or implicitly in terms of their positive impacts on the labor force participation of adult children. The HLI program will examine this in the context of women’s non-market work and its valua- tion. The insights from this will also be relevant to the non-market contributions of older persons. 27   WHO Study on Global Ageing and Adult Health (SAGE). https://www.who.int/data/data-collection-tools/study-on-global-ageing-and- adult-health. 28   ILO estimates that on average women around the world perform 4 hours and 25 minutes of unpaid care work every day compared with 1 hour and 23 minutes for men (Pozzan and Cattaneo 2020). CHAPTER 2 51 to NCD risk behaviors such as smoking, including mental health, and self-rated health (Schuring et al. in developing countries such as China and Kenya 2012; Van der Noordt 2014; Modini et al. 2016). (Jin et al. 2022; Donfouet et al. 2021). Inclusion of Finally, there is an expanding literature – large- omitted variables does not significantly reduce the ly from richer countries - on the health effects of size of impacts of schooling on self-reported health retirement or continued work. The findings across and NCD risk behaviors such as smoking (e.g., Con- studies are not definitive (Van der Heide et al. 2013), ti and Heckman 2010; Conti et al. 2010; Conti and but on balance meta-analysis finds positive effects Hansmann 2013; Van der Pol 2011; Savelyev 2014). on mortality, cognition, and some other health con- Twin studies have mixed findings, while instrumen- ditions of extending working lives beyond normal tal variable analysis for developing countries also retirement ages (Sewdas et al. 2020; Rohwedder and has mixed findings on a causal relationship but with Willis 2010), confirming country level studies (e.g., variation across different measures of health and Wu et al. 2016; Kachan et al. 2015). Other studies health behaviors and for some studies gender (Clark have linked working past retirement age with a re- and Royer, 2013; Braakmann 2011; Fabrice and duced risk of dementia and heart attack. Studies are Jones 2011; Buckles et al. 2013; Atella and Kopinska not unanimous, though. Meta-analysis suggests that 2014). To take the important example of smoking, retirement may result in improvements in mental it would be hard to argue that less smoking causes health, though was contradictory on impacts on more education rather than the other way around. general and physical health (Van der Heide et al. For conditions such as obesity, the relationship in 2013). The review also found little analysis of poten- developing countries may be more complex. tially variable impacts across types of work (white There is also a literature on how income, versus blue-collar) or on variation according to wealth, and employment shocks (all elements of whether retirement was voluntary or not. Yet again, returns to human capital) affect adult health, pri- two-way causality is likely to be a factor: those who marily NCD conditions (Capatina et al. 2020). The work longer may be healthier; but those who are literature finds that job losses result in negative im- healthier have both the capacity and desire to work pacts on health in the longer run, including worse more (for income and non-income reasons). health behaviors, self-reported health and mental health, and mortality (Smith 1999 and 2004; Black Protecting human capital et al. 2015; Eliason and Storrie 2009; Adda, Banks, and von Gaudecker 2009). Looking at NCD risk As shown in Figure 2, in addition to the life course factors, a systematic review of low and lower-mid- impacts of NCDs on human capital accumulation, dle income countries found that in most studies low deployment and depreciation, a further important SES groups had higher tobacco and alcohol use, aspect of the human capital trajectory is protection and consumed less fruit, vegetables, fish, and fiber of human capital across the life cycle. This will pri- than high SES groups (Allen et al. 2017). Reviews of marily be in the face of various types of shocks, in- health literature finds similarly strong relationships cluding idiosyncratic and covariate health shocks, between unemployment and various indicators of external shocks such as weather-related or eco- mental health (mixed symptoms of distress, depres- nomic crises, or employment-related shocks such sion, anxiety, psychosomatic symptoms, subjective as job loss. It also includes consumption smoothing well-being, and self-esteem), as well as general and across the life course into older age through pen- physical health (McKee-Ryan et al. 2005; Paul et al. sion or social assistance systems. The main focus of 2009; Norström et al. 2014), with the effects appear- discussion below is on health shocks, in particular ing to be causal.29 Those who are unemployed are NCD-induced health shocks, and financial protec- also more likely to engage in NCD risk behaviors tion in the face of them. such as over-consumption of alcohol, tobacco con- While the range of measures of catastrophic sumption and substance abuse (Henkel 2011). Con- health expenditure (CHE) and the non-health ele- versely, being in work, or returning to work, appears ments included (or often not included) in their cal- to have a protective effect for depression, general culation makes cross-study comparisons challeng- 29   As with other areas, there may also be reverse causality. 52 CHAPTER 2 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E ing, nearly all studies show worryingly high levels is also evidence that risk factors such as smoking can of CHE.30 In many developing countries, illness and be significant contributors to CHE and driving fami- death is one of the most common drivers of pov- lies into poverty (Verguet et al. 2015; Global Tobacco erty (McIntyre et al. 2006), and health and weath- Economics Consortium 2018). Where systems are er-related shocks are recurrent and their impacts built around insurance, those who are uninsured are cumulative across the life course (e.g., for Africa, 2-7 time more likely to experience CHE (Jan et al. see Nikoloski, Christiaensen, and Hill 2018). One 2018), though similar differences might also be ex- study by Wagstaff et al. (2018) using spending data pected where systems are general revenue financed from 133 countries found that over 800 million but fail to cover segments of the population. people in 2010 incurred CHE, and that the share The figures above are also likely to underes- of total households incurring spending of at least timate the impacts of NCDs on human capital, as 10 percent of household consumption rose steadi- many households may avoid or delay seeking care ly to almost 12 percent in 2010, and approached 15 due to the financial burden, or services may simply percent in Latin America and the Caribbean (LAC) not be available. In addition, lost labor income from region.31 At the individual country level, there was time spent accessing care may not be properly ac- a wide range in incidence of CHE, from under one counted for, nor in many studies costs such as trans- percent in some richer countries to over 40 percent. port for accessing care (e.g., see Jan et al. 2018 on Alternative estimates for developing countries using studies which include transport costs, which in India a different measure of CHE (40 percent of capacity for example is the major cost item of seeking care). to pay) find that the incidence of CHE was 30 per- In terms of financial protection against CHE, cent of households in LMICs, 17 percent in UMICs the push towards universal health coverage (UHC) and 15 percent in LICs (Haakenstad et al. 2022).32 across the developing world has been a crucial de- While the literature on the household financial velopment of recent years. However, the realiza- burden of seeking and receiving NCD care in de- tion of UHC continues to encounter challenges, in veloping countries is expanding steadily, it remains terms of the proportion of population covered by more limited for LICs and for specific conditions services, the share of health expenditures which such as COPD (Jaspers et al. 2015; Eze et al. 2022; are prepaid, and the range of NCD-related health Kankeu et al. 2013; Jan et al. 2018, Essue et al. 2017). services included in UHC packages (Wagstaff and Nonetheless, the evidence is consistent that the nega- Neelsen 2020; Watkins et al. 2017). Even where tive financial impacts on households from NCDs are people are in principle covered by SHI or entitled to substantial across the developing world, with major services under general revenues financed systems, deficits in financial protection. Recent work from financial protection is often shallow due to limited developing countries, which carefully distinguishes service packages, co-payment requirements and/or CHE from NCDs and communicable diseases, finds informal payments, or other factors such as costs of that CHE from NCDs tends to increase with the share accessing health services. Comparative analysis of a of NCDs in the overall burden of disease, though UHC index combining service coverage and degree with major exceptions such as China and South Af- of financial protection finds a mixed picture across rica, mediated by the health system. NCD CHE rates developing countries with respect to financial pro- were also consistently higher in lower wealth quin- tection. Some countries (e.g., Ghana, Indonesia, and tiles within countries (Haakenstad et al. 2022). There Viet Nam) have increased both financial protection 30   Catastrophic health spending is defined as a proportion of total household expenditure - usually between 10 and 40 percent of total consumption - or 40 percent of non-food consumption. Impoverishing health spending is defined as taking a household below a defined poverty threshold. Jan et al. (2018) provide a useful comparison of studies on CHE shares using a range of measures. 31   For sub-Saharan Africa, Eze at al. (2022) find that the incidence of CHE continued to rise from 2010-20. These measures are not for NCD-related health expenditures specifically. Haakenstad et al. 2020 find that across 39 countries around 45 percent of CHE cases could not be linked to a specific cause. 32   The result for LICs appears counter-intuitive, given the greater likelihood of households being poor in the first place. While the authors do not speculate, it may be that those in the poorest countries are more likely to forego health treatment and spending altogether due to income and supply side constraints. Somewhat distinct from these results, Wagstaff and Neelsen found that GDP per capita was positively and significantly associated with CHE, and higher inequality had a positive association with CHE at all country income levels (Wagstaff and Neelson 2020, noting that low service coverage in LICs may drive the former finding). CHAPTER 2 53 and service coverage, while financial protection in Brazil on informal worker household which finds a others has declined (Wagstaff and Neelsen 2020). significant added worker effect from health-induced The impacts of health spending by households work absence (Reis 2007). Work from Chile has on human capital are both direct and indirect (Jan et found more nuanced impacts, with the added work- al. 2018). The most direct negative effect will be on er effect observed for some chronic conditions (e.g., household non-health consumption. This may impact arthritis) but not others (hypertension and asthma), human capital in various ways, including reduced nu- and variable according to the age of the ill and well trition (impacting human capital formation early in spouses (Acuna et al. 2019). The studies highlight the life and its preservation later in life); negative coping importance of considering the specific health condi- strategies such as withdrawal of children from school tion and its implications for whether spouses are able due to household budget constraints and increased to work or need to focus on providing care. participation in child labor; and selling assets and/ Apart from social health insurance, unemploy- or running down savings (both likely to have impacts ment, disability, and work injury insurance pro- on productivity, and hence reduce returns to human grams can play important roles in providing finan- capital). When an adult female in the household has cial protection and protecting human capital. Like an NCD which compromises her capacity for unpaid SHI, to the extent that these other forms of social work/care responsibilities, this is also more likely to insurance protect incomes in the face of shocks, they be redistributed to other females in the household, should mitigate both direct consumption impacts often girls. Reduced household income from health on human capital and negative coping mechanisms spending may also result in failure to seek care when which compromise it. However, unlike SHI where needed subsequently, directly impacting the health subsidized approaches have become increasingly dimension of human capital. At the same time, there popular in developing countries, coverage of these is evidence from a number of developing countries programs tends to be confined to formal sector that illness of one family member often induces in- workers who are often a minority of the labor force. creased work effort from others to mitigate the in- These major coverage gaps were highlighted during come shock, including migration (Nikoloski, Chris- the COVID-19 pandemic. Coverage of non-health tiaensen, and Hill (2018) for Africa, for example). forms of social insurance is also much more likely to Estimates of the scale of the so-called added worker exhibit income and gender bias due to coverage be- effect tend to focus on participation of non-ill adults ing focused on the individual rather than the house- rather than net income impacts. A study from Korea hold. The exception to this coverage deficit in many for example found onset of chronic illness of a hus- developing countries is disability transfers under so- band reduced the probability of the wife exiting the la- cial assistance programs which are targeted on the bor force by 9.2 percentage points, and the effect was basis of severity of disability and in some countries more pronounced for chronic than acute conditions also according to poverty/vulnerability (World Bank (Kim et al. 2018). This is consistent with work from 2018), though adequacy is often modest. INEQUALITY AND GENDER DIMENSIONS OF NCD AND HUMAN CAPITAL PATHWAYS In thinking through the pathways discussed in previous sections, it is vital to consider distributional aspects. Inequalities may have several layers, from differentials in NCD incidence by different characteristics, to dif- ferences in human capital deployment/returns in the labor market, to differences in access to and provision of care, to financial protection. There are also likely to be important intersections between different dimensions of inequality, e.g., between gender and poverty. This section focuses on two important sources of inequality and their relationship to NCDs and human capital: socio-economic status (SES) and gender. With respect to SES, the first clear inequality is be- countries, there is a strong relationship between GDP tween income and life expectancy, both across coun- per capita (PPP) and life expectancy.33 While there tries and across households within countries. Across are outliers on the up- and downsides, and consid- 33  Clio Infra Project. Life Expectancy at Birth (total). Available at: https://clio-infra.eu/Indicators/LifeExpectancyatBirthTotal.html 54 CHAPTER 2 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E erable improvements in developing country life ex- HICs also found on balance an association between pectancy over time, the overall pattern of rising life SES and NCD incidence (Sommer et al. 2015). expectancy with country income is strong. Similarly, Looking at NCD risk factors in developing the likelihood of death from specific NCDs remains countries, a systematic review of low and low- substantially higher in lower income countries ver- er-middle income countries found that in most sus richer ones (e.g., a woman in a LMIC has around country studies low SES groups had significantly twice the annual likelihood of dying from CVD than higher prevalence of tobacco and alcohol use, and her peer in a HIC, Yusuf et al. 2014). consumed less fruit, vegetables, fish, and fibre than Within countries, the income gradient of life high SES groups (Allen et al. 2017). At the same expectancy across individuals has been shown in a time, high SES groups were less physically active number of developed countries (e.g., Chetty 2016; and consumed more fats, salt, and processed foods. Galama et al. 2018; Lleras-Murney 2022; Mortensen The Sommer et al. review also found an association et al. 2016). The evidence from developing coun- between SES and NCD risk factors, with low SES tries is thinner, but available multi-country studies increasing the risk of developing CVD, lung and show a clear relationship between lower income, gastric cancer, type 2 diabetes, and COPD. especially very low income, and higher mortality, Apart from incidence and risk factors, other though the direction of causation between health aspects which may impact human capital often ex- status and income is less clear (Banerjee and Duflo hibit SES inequality. One as noted above is coverage 2010). Interestingly, one study of China, Costa Rica, of social insurance. Those in the informal sector Indonesia, Mexico, South Africa, and South Korea and with lower SES status are more likely either to found more nuanced relationships between educa- lack SHI coverage, or – where SHI participation is tion levels and adult mortality than in richer coun- subsidized for poor households or those outside the tries, with the exception of a consistent advantage formal sector – to have SHI packages which may for the tertiary educated (Sudharsanan et al. 2020). be more limited than those in formal sector or civil As in richer countries, the available evidence service schemes or more rationed in terms of ser- from developing countries shows clear socio-eco- vice access. In almost all developing countries, low nomic inequalities in incidence of both NCDs and SES informal workers (and those outside the labor risk factors, though evidence remains incomplete force altogether) are highly likely to be outside un- and the quality of studies on the issue mixed (Hos- employment, disability, work injury and other so- seinpoor et al. 2012; Williams et al. 2018; Allen et cial insurance programs, except to the extent that al. 2017; Sommer et al. 2015). The simplistic notion those risks have some protection under social assis- that NCD are rich people’s diseases in developing tance systems (though even then, the adequacy of countries thus appears not to be borne out by the cover will be less than in social insurance systems). evidence and suggests that inequalities in preva- The risks of uninsured income shocks from these lence and risk factors should be take into account sources and negative coping strategies that may in policy. For example, a majority of studies in one compromise human capital are accordingly higher. systematic review found higher incidence of cancer, With respect to gender, there are multiple lay- CVD and multiple NCDs for low SES households, ers of gendered differences with respect to NCDs while obesity in contrast was higher among the bet- and human capital which are important to consid- ter off (Williams et al. 2018). However, for all condi- er (Bonita and Beaglehole 2014; Knaul et al. 2021; tions examined, the results were not uniform across Langer et al. 2015). There are several underlying countries (and in some cases also within larger differences in NCD incidence and risk factors, but countries like India) and studies in terms of SES also a range of other inequalities which may also and NCD incidence, suggesting a need for further impact human capital accumulation, deployment research. In an earlier review, angina, arthritis, asth- and protection. With respect to NCD incidence ma, depression and NCD comorbidity were associ- and risk factors, NCDs dominate the disease bur- ated with lower SES and education within countries, den globally for both genders (and for women ac- while for diabetes the association was in the oppo- count for around two thirds of deaths annually), but site direction (Hosseinpoor et al. 2012). Looking there are some differences in the gender-specific across developing countries, all these inequalities patterns of NCD incidence and risk factors. For ex- were more pronounced in LICs relative to MICs. A ample, male smoking rates (and lung cancer deaths) review of systematic reviews across LICs, MICs and are substantially higher than women globally, and CHAPTER 2 55 harmful alcohol consumption tends to be lower LICs and LMICS around 50 percent higher than in among women, while obesity is higher for women UMICs and HICs. In addition to unpaid care work, than men in most countries, and women dispropor- there is a strong gender dimension to paid health- tionately suffer and die from Alzheimer’s and other care work also, with women over-represented in dementias (WHO 2022). Of particular note is the health sector workforces across the world, though widespread prevalence globally of gender-based vi- often in lower skilled and lower paid jobs (or not olence as a risk factor which drives higher incidence paid at all, as with community health volunteers in of mental health disorders, other conditions such many developing countries). Whether paid or un- as head trauma and consequent cognitive impacts, paid, the dual role of women as both consumers of and higher rates of behavioral risk factors for NCDs healthcare and aged care and providers is notably (Devries et al. (2013) estimate around one third of more pronounced than for men (Langer et al. 2015). women globally experience intimate partner vio- A related but less obvious gender dimension lence). The combination of lower age-specific death in terms of health outcomes is the disproportionate rates for women from NCDs and their greater lon- burden on women of household tasks which are not gevity means that around 88 percent as many wom- directly care-related. A number of the most time-in- en as men die annually from NCDs (WHO 2022). tensive tasks of women in households such as cook- A second source of inequality relates to the gen- ing, cleaning, and fetching water (often called el- der patterns of unpaid care provision. While NCDs ements of “reproductive labor”) have very direct represent a major health burden for all genders, the health and nutritional benefits for households, low- burden of informal care falls disproportionately on ering their likelihood of both communicable and women. Globally, around 76 percent of all unpaid non-communicable diseases. The wellbeing and care work (including household work, direct per- economic benefits of this labor are almost never sonal care, and volunteer work) is undertaken by measured but are likely to be substantial. One such women, though is not typically valued in national analysis for Mexico is instructive, suggesting that accounts nor in most estimates of returns to human proper accounting for the health benefits of mixed capital.34 That share can run to over 90 percent in household work which also promotes better health some countries (e.g., India, Pakistan, Cambodia, in the household would significantly multiply the and Mali). The gender differentials in unpaid care economic contribution of women’s direct unpaid work are in all regions and at all levels of income, care work. For Mexico, the total value of women’s though they are higher in LICs and MICs than HICs unpaid contribution to health increased from 0.5 on average. Almost half of women outside the labor percent of GDP to 2 percent once the health-pro- market in MICs and over a third in LICs cite unpaid moting value of mixed work was accounted for, care duties - broadly defined - as the main reason valuing time at minimum wage (Langer et al. 2015). (ILO 2018; Charmes 2019; Folbre 2015). Thus, there Broader but related to the points above, a major is a massive underutilization of women’s human cap- source of gender difference relates to measurement of ital due to caring duties using standard valuations deployment and returns to human capital. The lower to returns to human capital. Based on data from 53 labor force participation rates of women in market countries across income levels, ILO estimates that work and the common failure to value their non-mar- women’s unpaid care work would amount to 6.6 ket work results in a major downward bias in returns percent of global GDP. Using a more focused defini- to women’s human capital. To the extent that NCDs tion of the value of women’s contribution to health compromise labor force outcomes, these factors will work globally, Langer et al. (for the Lancet Com- also result in major under-estimation of the negative mission on Women and Health) estimate unpaid impacts of NCDs on the human capital of women health-related care work of women at between 1.1 and societies and on growth. Even where women are and 3.1 percent of GDP globally, depending on the working in the labor market, the gender wage gap will valuation method used, and with the GDP share in also contribute to lower estimated impacts of a given 34   ILO estimates that the bulk of unpaid care work is household duties other than direct personal care, though acknowledged that multi-tasking across regular household duties and care specifically makes precise estimates of shares of type of unpaid care difficult to unpick even with time use surveys. As noted by Langer et al. (2015) “...four methods are typically used [to value unpaid care], each with advantages and disadvantages: opportunity cost; proxy good; contingent valuation; and conjoint analysis”. 56 CHAPTER 2 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E burden of NCDs for women relative to men. ical differences, provider bias or other factors is not A further source of gender inequality with re- easy and evidence from developing countries is very spect to NCDs relates to access to healthcare and thin (Chang et al. 2007). To the extent such differen- patterns of care provided. This is an area with a tial treatment exists, it may, as noted, reflect in part thinner evidence based in developing countries. an ongoing prioritization of sexual and reproductive Overall, there appear to be gendered patterns of health as the dominant concern in women’s health access to care, and the nature and quality of care (Langer et al. 2015; Bonita and Beaglehole, 2014). provided if accessed, shaped by social determinants As in many areas of social policy, the gender including gender inequality, poverty and cultural inequalities in NCD incidence, access to care and factors (Langer et al. 2015). The first source of gen- NCD management, financial protection and other der difference relates to women’s lack of control of factors often co-exist with other sources of depriva- household resources in many developing countries tion, and the various factors are likely to be reinforc- and the likelihood of this impacting their access to ing. This may include race or ethnicity, language, care relative to men. This may reflect either higher illiteracy, disability status, poverty, sexual orien- poverty rates of female-headed households, or with- tation, and other factors. This inter-sectionality of in male-headed households lack of prioritization of deficits or risks applies to all population groups but women in whatever is spent on healthcare. While are likely to be more pronounced for women. extending and deepening financial protection for As with NCD and health-related issues noted health will benefit all people, lower financial barri- above, there are clear gender differentials in both edu- ers are thus likely to benefit women disproportion- cation and labor market participation and outcomes, ately (and those benefits to be outsized for poorer though the picture on education has become some- women). Another factor with respect to access is so- what more nuanced in recent decades (Evans, Akmal, cial and cultural factors which limit women’s access and Jakiela 2020; Saavedra et al. 2021). Historically, to care where there are strong norms around wom- gender gaps in girls’ school enrollment were ob- en being treated by men and there are insufficient served in nearly all developing countries and remain female providers as evidence from Afghanistan has in many. The positive news is that the median gender shown for example (Save the Children 2008). gap in enrollment reduced globally between 1960 and Apart from failure to access care when avail- 2010 from 1.2 years to 0.8 years, and that around two able, there may be gender biases or gaps in the defi- thirds of all countries have achieved gender parity nition of publicly financed basic packages of care. in primary enrollment. At the same time, the Bar- For example, it has been common in basic packages ro-Lee data set for 146 countries finds that while 104 to prioritize sexual and reproductive health inter- countries reduced their gender gap in that period, it ventions for women, which may come at the cost of increased in 42 countries (Evans, Akmal, and Jakiela NCD-focused interventions for prevention or man- 2020). While a significant number of those are fragile agement (Bonita and Beaglehole 2014). An example states, India is one which saw an increase. would be failure to include mammography in ba- Looking below high-level averages, the picture sic packages in some countries. While such neglect remains concerning, however. In countries affected may not be intentional, it may be reflective of lack by fragility, conflict and violence girls remain 2.5 of prioritization of research on women’s NCD in- times more likely to be out of school than boys. Also, cidence and risk factors to strengthen the evidence in low-income countries the school completion rates base upon which disease control priorities are for- of girls at both primary and secondary levels remain mulated at global and national levels. below that of boys (63 percent for girls against 67 Similar issues arise relating to care received when percent for boys at primary level, and 36 percent care is accessed. First, even where women do seek versus 44 percent at secondary level – Saavedra et care, this may be more likely to be delayed and time- al. 2021). These deficits are more pronounced in liness of care compromised (see for example, Liakos economic terms, given that the average rate of re- and Parikh (2018) for the US). While the evidence is turn to an additional year of schooling is higher for largely from developed countries, there appear to be girls, at around 12 percent against 10 percent for gender differences in diagnosis and management by boys. Across developing countries girls who have doctors, with women being asked fewer diagnostic multiple disadvantages such as coming from poor questions, and given fewer diagnostic tests, though households, living with disabilities, or coming from unpicking how differences may be driven by biolog- minorities or rural areas are furthest behind. Inter- CHAPTER 2 57 estingly though, the learning poverty rates of girls many countries are exacerbated during the deploy- in low- and middle-income countries are on average ment and further accumulation stages. In terms of actually lower than for boys (55 versus 59 percent), returns to human capital, the participation gap is and in many developing countries tertiary enroll- exacerbated by the gender pay gap in all countries, ment rates are slightly higher for young women than and the gap is sharply higher for women with chil- young men.35 At the same time, there remain strong dren (Terada-Hagiwara et al. 2018; Si et al. 2021).36 gendered patterns in subject mix and more general Gender differentials in terms of labor force biases in schools and classrooms which often rein- participation and wages are in turn reflected in force more limited expectations for girls (e.g., the differential gender coverage and adequacy of con- substantial under-representation of girls in STEM tributory pension, unemployment, disability, and subjects). While the situation has therefore im- work injury social insurances. While SHI tends to proved in many developing countries over time and be provided on a household basis, the other social become more nuanced, the averages also conceal insurances attach to the individual worker and significant regional and other variations, and gender overwhelmingly in developing countries to workers equality in education remains an unrealized goal for in the formal sector. To the extent that women have significant numbers of developing countries and for lower overall labor force participation and higher significant shares of girls within countries. rates of informal sector employment when they are When considering deployment of human cap- doing market work, they have lesser financial pro- ital, there remain often substantial gender gaps in tection against these non-health risks and in old age. labor force participation in much of the developing Even where they are working in employment where world, and the gaps are particularly pronounced in non-health insurances are provided, they are like- South Asia (SAR) and Middle East and North Afri- ly to have shallower coverage of protection during ca (MENA). Globally, female labor force participa- working life (due to vesting periods and breaks in tion is around 20 percentage points lower than for work history due to child rearing for many wom- men overall, and over 40 percentage points lower in en), lower contribution densities across their work- MENA and SAR regions. Even in the developing re- ing lives for pensions, resulting in often significantly gions with the highest labor force participation such lower pensions in later life, and lower adequacy of as EAP, there remains a participation gap between non-health insurances in general reflecting the gen- women and men in terms of market work, which der wage gap. Where women do not work but their increases for higher skilled and upper level and spouses do, the prevalence and adequacy of survi- leadership positions, including in the health sector vor pensions for female spouses is also mixed across (Knaul et al. 2021). As a result, the gender gaps at the developing world (Chłoń-Domińczak 2015). the intensive human capital accumulation phase in 35   Learning poverty is a concept popularized by the World Bank which focuses on the ability of children aged ten to read and comprehend a simple paragraph, as well as shares of children out of school altogether. 36   In response, Target 8.5 of the 2030 Sustainable Development Goals calls for “equal pay for work of equal value.” 58 CHAPTER 2 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E WELLBEING EFFECTS OF ADDRESSING NCDS While the bulk of literature discussed above is on the links between economic growth or productivity and health, longevity and education/skills, it is important also to take account of the direct impacts of health, and NCDs more specifically, on non-economic measures of wellbeing at the individual and societal levels. This is a rapidly growing area of research, with different measures of life satisfaction used, including life satisfaction, quality of life, health-related quality of life, and happiness.37 In economist speak, good health has consumption value in terms of improved quality of life, over and above its contribution to individual and societal welfare through its investment value and impacts on productivity, growth, and individual earnings. Alternatively, one can look at health (and education) as functions which contribute directly to capabilities which are the basis for the well- being, freedom and dignity of people, over and above their narrow human capital contribution to production (Sen 1997; Nussbaum 2011). They are both the result of capabilities and prerequisites for capabilities.38 In recent years, the literature on the relationship between health status and different measures of wellbeing has been grow- ing (Deaton 2008; Steptoe et al. 2015; Steptoe 2019).39 As noted earlier, wellbeing has different dimensions: eval- uative wellbeing (or life satisfaction), hedonic - or affective - wellbeing (feelings such as anger, happiness, sadness etc.), and eudemonic wellbeing (sense of meaning or purpose). Interestingly, measures of wellbeing across the life course vary across regions, with richer English-speaking countries exhibiting a U-shaped pattern of lowest life satisfaction around mid-40s, while former Soviet and LAC countries see a steady increase in dissatisfaction with age, and Africa shows no strong age pattern (Steptoe et al. 2015, using Gallup data from 160 countries). Evidence is growing that subjective measures of assessments of eudemonic wellbeing, and momen- wellbeing matter for health, and that health – and tary hedonic measures is in Steptoe et al. 2015.40 A specifically NCD status - impacts subjective mea- growing body of literature suggests that being happi- sures of wellbeing in a bidirectional relationship. er (often measured by absence of depression or dis- Poor health has been shown to reduce subjective tress) is good for health. A major meta-analysis finds wellbeing (and conversely, positive health behaviors that it reduces all-cause mortality (Martin Maria et such as physical activity positively affect wellbeing), al. 2017). And a range of studies find similar posi- and lower subjective wellbeing increases the risks of tive impacts on specific NCD conditions (Davidson, premature mortality and a range of NCDs, including Mostofsky, and Whang 2010; Boehm and Kubzan- coronary heart disease, diabetes, and other chron- sky 2012), including coronary heart disease (Carney ic conditions (Chida and Steptoe 2008; Windle et and Freedland 2017), stroke (Li et al. 2015) and type al. 2010; Steptoe et al. 2015, Steptoe 2006). Just as 2 diabetes (Hackett and Steptoe 2017). Depression interestingly, newer research suggests that positive also predicts increased mortality among people with subjective wellbeing may be a protective factor for COPD, diabetes, CHD, stroke, and some cancers health and is associated with longer lives and lower (studies summarized in Steptoe 2019). For NCDs morbidity. A brief summary of the literature from specifically, studies also show increases in depres- prospective epidemiological studies, retrospective sion after diagnoses of various conditions, including 37   While the focus here is primarily on the WHO Quality of Life and happiness measures, there has been a proliferation of initiatives which aim to go beyond GDP as the dominant measure of human wellbeing, including the OECD’s Better Life Initiative, and the work of the Stiglitz/Sen/Fitoussi Commission which is informing quality of life measures by the EU. The UN Human Development Index inspired by Sen has a similar motivation. 38   Nussbaum also contrasts internal capabilities (which are personal), and combined capabilities that are “defined as internal capabilities together with the social/political/economic conditions in which functioning can actually be chosen”. Combined capability “combines in- ternal preparedness with external opportunity in a complicated way, so that measurement is likely to be no easy task.” (Nussbaum 2011). 39   While not the focus of this chapter, this is also true for the studies on the wellbeing effects of education, which look at both direct impacts on wellbeing (for which findings are inconsistent across studies, and weaker than for health when positive) and impacts on non-pecuniary factors which are associated with better quality of life, within the labor market (e.g., job satisfaction, positive mental health impacts of lower unemployment probability, or social status), and outside it (e.g., better health behaviours, marriage market outcomes, childrens’ education- al performance, fertility behaviour for women, parenting behaviour, and variables such as levels of trust and social capital. See Oreopoulos and Slavanes 2011; Grossman 2006; Vila 2000; Helliwell and Putnam 1999; Powdthavee et al. 2015; Blackstone and Van Rensselaer 2019. 40   While more evidence is available from prospective studies to date, with their potential shortcomings including confounding factors and reverse causality, retrospective work points to similar conclusions. CHAPTER 2 59 diabetes, coronary heart disease, some cancers, and sion in combination with other NCDs had particu- chronic kidney disease (Satin et al. 2009; Hedayati et larly strong negative effects). A study from Brazil also al. 2010; Meijer et al. 2011). Even more interesting- found that the negative impacts of NCDs on QOL ly, studies have found that the association between scores were more pronounced for older people, and happiness, mortality and specific conditions seems those with low income levels, low education, and to hold even after controlling for negative emotions Black people, suggesting that the socioeconomic dif- (Tilvis et al. 2012; Steptoe and Wardle 2012). ferentials which are apparent for NCDs themselves Broader measures of quality of life and its may be exacerbated when considering wider mea- linkage to health have also been a growing focus sures of life quality (Höfelmann et al. 2017). of WHO. The WHO Quality of Life (WHOQOL) Using a simpler measure of wellbeing, several method aims to measure an individual’s “percep- rounds of the World Happiness Report find that bet- tions of their position in life in the context of the ter health is strongly associated with happiness. A culture and value systems in which they live and in range of factors are positively associated with higher relation to their goals, expectations, standards and life satisfaction, including income (notably relative concerns”, aiming for a composite (and decompos- income) and employment, education and family life, able) measure which provides a measure of quality and mental and physical health (Helliwell, Layard, of life that goes well beyond traditional measures and Sachs 2017; Clark et al. 2017. However, mental of health and self-perceived health (WHO 2012). health has a particularly strong impact, in Western The WHOQOL instrument was developed across countries studied more than income, employment 15 centers globally (including in some developing or physical illness, and second only to income in the countries) in an effort to reflect culturally diverse developing country analyzed (Indonesia). There is perceptions of quality of life and asks 100 ques- a similar ranking when one focuses on the drivers tions across a range of domains of wellbeing (phys- of misery rather than satisfaction, with elimination ical, psychological, level of independence, social of depressive and anxiety disorders the single most relations, environment, spirituality/religion, and impactful intervention in simulations (and the most personal beliefs), as well as an overall measure of cost-effective). Childhood factors such as emotion- wellbeing and health. The instrument has also been al health and conduct are also strong predictors of adapted to the WHOQOL-BREF, which is a short- adult happiness (Layard et al. 2014), and childrens’ ened form of the instrument better suited to inclu- emotional health is in turn impacted by parental sion in surveys which may be measuring multiple health, particularly the mental health of the moth- things beyond quality of life. er (Helliwell, Layard, and Sachs 2017). The findings The research from developing countries which are consistent with country-specific studies, which relates NCD incidence to WHOQOL measures is rel- also point to the importance of the health measures atively nascent but growing. Studies from middle-in- used, with conditions which affect daily functioning come and lower-income countries have found neg- more closely associated with impacts on happiness ative associations between a range of specific NCDs (e.g., Angner et al. 2013). Looked at from a health and lower QOL using the QOL-BREF measure (Lee capital perspective, Arrow et al. (2014) estimate that et al. 2015; Arokiasamy et al. 2015), and cross-coun- the direct well-being impact of better health (that is try work on middle-income countries has also ana- not mediated through production of good and ser- lyzed how multi-morbidity NCD dyads negatively vices as measured in GDP) is likely larger than their impact QOL (Sum et al. 2019, for example finding indirect effects through production. that diabetes, stroke and depression had the largest negative impacts on mean QOL scores, while depres- 60 CHAPTER 2 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E POLICIES AND INTERVENTIONS FOR NCDS AND HEALTHY LONGEVITY A wide range of policies can impact human capital across the life course, as well as the likelihood of developing NCDs which negatively impact healthy longevity. This section looks at policies that have the greatest potential to impact NCDs and healthy longevity across the life course. These can contribute to achieving the UN SDG Target 3.4 of reducing premature mortality from NCDs by a third between 2015 and 2030. Across the life course we focus on policies and interventions to improve NCD outcomes and support healthy longevity, including for nutrition, NCD prevention and management, as well as interventions for long-term care, the built environment and tax, air pollution and traffic accident reduction. The policies and interventions considered here are those with the most di- rect impacts on NCDs: the full spectrum of policy areas includes education, labor, and social protection, to name a few (which are summarized in Annex Tables 2A.2 and 2A.3). This is a menu, and countries would need to choose judiciously what combination of policies and interventions were cost-effective and consistent with financial and capacity constraints. Evidence to guide these choices is provided in several of the chapters on HLI research present- ed in this compendium, in particular on the economic burden of NCDs and cost-effectiveness of different interven- tions. Within a number of these areas, tax and fiscal policies can play a useful role. There is also a common need for better data and measurement tools, as well as research to inform policy development and evaluate outcomes. i. Nutrition Nutrition policies are the most foundational for hu- in early nutrition one of the best value-for-money man capital, NCDs and healthy longevity given their development actions” (Shekar et al. 2017). important short and long-term impacts. The policy There have also been considerable investments menu of nutritional interventions focuses primarily in complementary interventions, including agri- on women of child-bearing age and young children cultural programs to promote more diverse food and combines both direct nutritional interventions cultivation and adoption, food systems, WASH in- and multi-sectoral policies which address the un- terventions, women’s empowerment programs, and derlying determinants of poor nutrition (Bhutta increasingly conditional or “cash-plus” cash trans- et al. 2008 for a meta-review of studies on different fers which help address the demand side and in- interventions for maternal and child nutrition and come constraints to better nutrition.41 While more their effectiveness; Von Salmuth et al. 2021; IFPRI research is needed to understand the cost-effective- 2016; Shekar et al. 2017). The direct nutritional in- ness and interactions of these interventions, there is terventions include strengthening supply-side inter- emerging consensus on a core set of maternal and ventions, including micro-nutrient supplementation child nutrition interventions which will impact (e.g., Vitamin A, zinc and iron for children and folate childhood and particularly adult NCDs and provide for pregnant women); pre-school and school feeding the cognitive foundations for better learning and programs (Verguet et al. 2020); social and behavioral human capital acquisition and deployment. There is change and parenting programs, including breast- also growing attention to obesity in childhood, both feeding and complementary feeding education; and as an NCD itself and as a pre-disposing condition for management of severe acute malnutrition following many NCDs later in life. WHO guidelines. World Bank costing of a core pack- While most attention in development literature age of nutrition-specific interventions in 2017 sug- is understandably focused on young people, the im- gested a cost of around US$ 70 billion over a decade portance of nutrition for good physical and mental to address childhood stunting and wasting, maternal functioning of older people has received increasing anemia and promotion of exclusive breastfeeding. attention in recent years. Poor diet and associated Apart from averting an estimated 3.7 million child micronutrient deficiencies have been linked to in- deaths annually, the economic returns per dollar in- creased morbidity, hospitalization, and mortality in vested were between $4 and $35, “making investing older people, including to a range of NCDs, includ- 41   For example, Manley at el. (2020) in a systematic review and meta-analysis of the impact of cash transfers on child nutrition find transfers programs had significant positive impacts on dietary diversity, animal-sourced food consumption, linear growth and stunting reduction, though program design matters for the degree of impact. CHAPTER 2 61 ing heart disease, cancer, osteoporosis and weak im- above. For the explosion of obesity at all ages, but mune systems in general (Kaur et al. 2015). There is a particularly childhood, the main areas for attention host of reasons for inadequate nutrition in older peo- relate to diet and lack of physical activity, both at the ple, from affordability to loss of enjoyment in eating, individual level - including limiting consumption of physical difficulties in eating for physical or cognitive fats and sugars, increasing intake of fruits, vegeta- reasons, mobility constraints in acquiring food, and bles, legumes, whole grains and nuts, and increasing motivation and depression.42 Studies in developing physical activity – and at industry and societal levels countries have found rates of malnutrition among reducing the fat, sugar and salt content of processed older people which often rival those of children, with foods, increasing availability and affordability of one cross-country synthesis for Africa for example health foods, better regulating marketing of obe- finding malnutrition rates between 10 and 36 per- sity-inducing food and taxing them more (see be- cent and 13.1 and 27 percent for men and women low), and promoting regular physical activity at the respectively (Charlton and Rose 2001). In terms of societal and workplace level. interventions to address poor nutrition at older ages, there are direct measures such as supplementation to ii. Preventing, delaying, and managing address common deficiencies (e.g., calcium, amino NCDs across the life course acids, iron, zinc and vitamin B), but also addressing the access, mobility and other social causes of poor There has been intensive work on the package of nutrition is at least as important (Roberts et al. 2019). policies and interventions that can best address the In addition to nutritional interventions for un- NCD epidemic, including how to do so as cost ef- der-nutrition at different ages, a growing concern in fectively as possible. These are thoroughly summa- the developing world is the double burden of mal- rized in Disease Control Priorities (DCP), including nutrition with the rapid increase in obesity (Popkin mental and neurological disorders (DCP 3rd edition, et al. 2020). This has two dimensions. First is the 2017; Lancet 2017) and WHO “best buys” for ad- direct prevalence of obesity itself and associated dressing the risk factors for NCDs (tobacco and al- NCD risks, including increasingly in childhood. cohol use, unhealthy diets, and physical inactivity) According to WHO, in 2016, more than 1.9 billion and managing the “big four” NCD categories (CVD, adults were overweight, and of these, over 650 mil- diabetes, cancers, and chronic respiratory diseases lion were obese, or 13 percent of the global adult (WHO 2017). While already crucial, there has also population. Obesity is associated with a range of been increased attention to NCDs as pre-disposing NCDs, including cardiovascular diseases, diabetes, for severe impacts from COVID-19 (Nikoloski et al. some cancers (endometrial, breast, ovarian, pros- 2021), as was also seen in the SARS and MERA out- tate, liver, gallbladder, kidney, and colon) and mus- breaks (Chan et al. 2003; Arabi et al. 2017). culoskeletal disorders. In addition, over 340 million Broadly, effective policies to reduce exposure children and adolescents were overweight or obese and incidence of and manage NCDs stretch across in 2020. Most of the world’s population now lives in a range of areas, many within the health sector, but countries where overweight and obesity kills more also quite a number beyond it. They include: people than under-nutrition (WHO 2021). The levels and rates of increase in childhood obesity in • Strategies to reduce exposure to NCD risk factors, developing countries are particularly alarming, and including taxes and other interventions to control associated with a range of NCDs, including type 2 tobacco, alcohol, sugar, trans-fat and salt intake diabetes, the early-onset metabolic syndrome, sub- use (see below); nutritional and dietary policies clinical inflammation, dyslipidemia, coronary ar- (see above); vaccinations to reduce risks of hep- tery diseases, and adulthood obesity (Gupta et al. atitis B and liver cancer; and measures to reduce 2012). Second is growing evidence on the increased indoor and outdoor air pollution. likelihood of adult obesity as a result of under-nu- trition in childhood (Heidari-Beni 2019). For the • Strategies to manage key NCD risks such as second dimension, the interventions for pregnant screening and management for cholesterol and women and children are common to those noted BP, diabetes, and cancers; promotion of self-man- 42   www.mayoclinic.org/healthy-lifestyle/caregivers/in-depth/senior-health/art-20044699. 62 CHAPTER 2 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E agement through education and support services, provide good value for money; (ii) be feasible to im- including promoting medication adherence; and plement in low- and middle-income countries; and widespread availability of palliative care. (iii) address a significant disease burden, drawing on systematic reviews of economic evaluations, ep- • Development or refinement of layered strat- idemiological data, clinical effectiveness studies, and egies for mental, neurological, and substance expert judgment. The key priorities identified were: use including community-level interventions in life-skills and building social and emotion- • For cancer control and management, a range of al competencies; promoting self-management; interventions for prevention, detection, diagno- strengthening primary care and outreach ser- sis, surgery, other treatments, and palliative care vices; and strengthening general and specialist (Gelband et al. 2015). These include “prevention care for severe and emergency cases and integra- of tobacco-related cancer and virus-related liver tion with other health care channels. and cervical cancers (including vaccinations), diagnosis and treatment of early breast cancer, • Building an effective coordinate care strategy cervical cancer, and selected childhood cancers; which is built on a central role for primary care and widespread availability of palliative care, in- with affordable access to care and availability of cluding opioids”. essential medications. • For cardiovascular, respiratory, and related dis- • Strategies to get better value from health systems, orders, Prabhakaran et al. 2017 provide a costed including moving from line item and fee-for- essential package of interventions to reduce risk service payments to combination of capitation of and manage CVRDs, including tobacco taxa- for primary care and global budgets with case tion, salt reduction interventions, bans on trans- mix-adjusted for hospitals; strategic pharma pur- fats, availability of essential medications, and chasing with reference pricing; transparent and strengthening of primary care. evidence-based processes for prioritizing medical interventions and technology; and building stra- • For mental, neurological and substance abuse tegic purchasing capacity. (MNS) disorders, “a variety of interventions, in- cluding drugs, psychological treatments, and • Strategies within the health system to slow func- social interventions, can prevent and treat MNS tional decline among older people, including for disorders. At the population-level, best practic- weakening eyesight, hearing and bones/joints. es include legislative measures to restrict access to means of self-harm or suicide and to reduce • Strategies outside the health system to prepare for the availability of and demand for alcohol. At the and mitigate functional decline in old age, includ- community-level, best practices include life-skills ing access to affordable assistive devices, access to training in schools to build social and emotional social participation, recreation and public spac- competencies. For healthcare, three channels are es, workplace health promotion programs, and identified: self-management, primary care, and improvement/adaptation in workplaces environ- community outreach; hospital care, including ments (see below). both specialist services and first-level hospitals for those with severe, refractory, or emergency The Lancet 2017 summaries of the DCP (3rd edition) presentations; and promoting the integration of spell out in more detail the priority interventions for mental healthcare in other channels of healthcare different groups of NCDs, with a model list of 21 in- (Patel and Chatterji 2015). terventions (the Lancet papers on DCP3; Bundy et al. 2017 for childhood and adolescent conditions; Black In addition to the looking at the cost-effectiveness et al. 2013 for reproductive, maternal, newborn and of policies and interventions to address NCDs, childhood interventions; and Debas et al. 2015 for chapter 15 by Watkins et al. in this compendium essential surgeries).43 The interventions all aim to: (i) incorporates additional factors to take into account 43   https://www.thelancet.com/disease-control-priorities-3 CHAPTER 2 63 as countries at different levels of development pri- ly in developing countries – frontline care will fall oritize their essential health packages. These include to non-specialists and non-medical people, making the degree of financial protection and equity con- the integration of basic preventive and management siderations, features of health system performance, geriatric care into GP, nursing, social work and relat- and fiscal space. They also provide a costed set of ed disciplines particularly important. priority interventions which allows for variation in There have also been rapid advances in dementia the Incremental Cost-Effectiveness Ratio (ICER) risk reduction and care, reflected in the 2019 WHO of specific interventions across regions.44 Overall, guidelines for risk reduction of cognitive decline and the authors note that UHC approaches to date have dementia and evidence that “multidomain approach- often emphasized high coverage with low OOP es targeting simultaneously multiple risk factors and spending and low cost, an approach which may not tailored at both individual and population level, are accommodate well within essential health packages likely to be most effective and feasible in dementia the broad nature of NCDs and may fail to exploit risk reduction” (Stephen et al. 2021). Apart from de- the synergies of shared platforms. At the same time, mentia specifically, there has been a proliferation in they find that a package if high priority NCD in- recent years of research on delaying, and even revers- terventions can be implemented across developing ing, normal cognitive decline in healthy older adults regions at a reasonable incremental cost with major - so-called cognitive aging – through combinations gains in deaths and DALYS averted. of exercise, diet, sleep, mental stimulation, and social Apart from condition-specific interventions, interaction (Blazer et al. 2015; Dumas 2017). Sim- global and national attention is needed for global ilarly, palliative care in both residential and outpa- public goods which are necessary for preventing tient settings, and using digital technologies, has led and managing NCDs. In particular, promoting an to improvements in patients’ and caregivers’ wellbe- R&D structure for relevant drugs, diagnostics, vac- ing at end-of-life and often cost-effective (Naoum et cines, and related products which balances legal al. 2021; WHO 2020). Despite that, WHO estimates protections and financial incentives for research that only around 14 percent of those needing pallia- and production with health needs and global equity tive care globally each year (most of them in develop- is an ongoing challenge. ing countries and almost 90 percent of them suffer- In addition to the general interventions for ing from one or more of the “big four” NCDs) have NCDs across the life course (and where the higher access to such care (WHO 2020). rates of NCDs and co-morbidities of older people make interventions especially pertinent for older iii. Long-term care people), there is a range of interventions which tar- get older people and have the potential to prolong life Closely related to aspects of healthcare, long-term and compress morbidity at older ages. Some of these care is an area which has gained growing attention will be focused on individuals and some ideally take from policymakers in recent years and where there a population-focused approach. These have been are direct and indirect impacts on health, human framed around the “4Ps”: predictive, preventive, capital, wellbeing, and productivity. In developing personalized and participatory. This would include (and, to a significant extent, developed) countries, population-based public health interventions such as this continues to be an area dominated by informal education on self-management of health, including provision of care to frail elderly people by families maintaining proper nutrition and exercise at older and other private arrangements. However, the eco- ages (Izquierdo et al. 2021). It also includes enhanced nomic and social cases for expanded public sector screening of the evolving capacity of older people involvement are compelling. In direct terms, ex- based on a biopsychosocial assessment, which would tending the period in which older people can live include not only general NCDs, but frailty, social safely in their own homes through access to decent isolation, dementia, and other factors that are most quality home- and community-based care improves pronounced at older ages (Stuck et al. 1993). While wellbeing of care recipients (who consistently prefer there will be a growing need for specialized geriatric those forms of care to residential care), increases the care skills as societies age, inevitably – and especial- return on earlier investments in the health capital 44   The geographic disaggregation is China; Eurasia and Mediterranean; India; LAC; and Sub-Saharan Africa. 64 CHAPTER 2 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E of both cared and carers, and is likely to be more ly-financed LTC and who is eligible for them, financ- cost-effective through delaying costly residential ing, contracting and provider payment mechanisms, care (Low et al. 2011; You et al. 2012). Long-term and human resources (Yiengprugsawan and Piggott care has also been seen to be a significant and grow- (2022) and O’Keefe and Yiengprugsawan (2023) on ing source of employment across the world given the experience of emerging Asia for example). the major shortages in care workers in both devel- oping and most developed countries. Realizing the iv. The built and community environment employment effects also often requires attention to migration policies, with immigration of care work- While much of the literature on NCDs tends to ers from younger and poorer countries a win-win focus on the underlying diseases, it is vital also to for sending and receiving countries (OECD 2020).45 think about how NCDs play out in terms of func- In indirect terms, there are demonstrated positive tional capacity. As WHO has emphasized in its impacts on the wellbeing of informal carers from World Report on Ageing and Health in 2015 (sum- formal care interventions which supplement infor- marized in Beard et al. 2016), functional capacity mal care, and potential for extending the working is an outcome of the interaction of intrinsic capac- lives and/or intensity of market work of adult chil- ity (“the composite of all the physical and mental dren and in-laws (particularly women). capacities of an individual”) and the external en- While these benefits for health, human capital, vironment within which people live. That external budgets, and output are clear, long-term care is a environment has many dimensions, but a crucial sector where there remains enormous diversity in one is the built environment within which people terms of institutional boundaries between health live.46 As noted, a common underlying disease or and social care services, in the mix of services, lev- condition may have very different impacts on a per- els of public spending, and public/private roles even son’s functioning depending on whether their built in developed countries. In developing countries, environment is accommodative or hindering. This formal care systems remain largely in their infancy, social model of disability is highly relevant to the though a growing number are in the initial stages impact of NCDs on people’s ability to live a full and of developing them (see Glinskaya et al. 2018 for satisfying life. For the built environment, promot- discussion of China), with particular challenges ing universal design for buildings and public spac- in working through the appropriate institutional es, appropriate housing, multi-generational spaces, arrangements and financing models. For most de- more accessible and inclusive public transport sys- veloping countries, the starting point is likely to be tems, effective spatial planning to facilitate access, developing a national long term care strategy which and other features of the physical environment is delineates the roles of state, market, communities, crucial to maximizing the potential of people with and households and is built around an “ageing in NCDs as producers, satisfied citizens, and contrib- place” approach which also promotes a continuum uting family members (Das et al. 2022).47 While this of care across home, community-based and health will likely involve retrofitting on the stock of such facility settings. For countries which are further infrastructure (which may benefit from financial along in development of their long-term care sec- incentives beyond the public sector), it will be far tors, there is a range of issues which need to be ad- more cost-effective to build in universal design at dressed as the formal LTC sector matures, including the design stage, requiring incorporation in plan- institutional mandates, defining packages of public- ning and approval processes. 45   At the same time, the relatively low productivity of the sector makes it important to assess the net increase in output and returns to care worker human capital, which will also have a strong gender dimension given the preponderance of female workers. 46   Built environment has varying definitions but at a basic level covers the man-made or modified structures and spaces that provide people with living, working, services and recreational spaces, including man-made parks and green spaces and utilities and transport systems which provide the supporting infrastructure to utilize it. See for example US EPA: https://www.epa.gov/smm/basic-informa- tion-about-built-environment. 47   Universal design is built on seven guiding principles: Equitable Use; Flexibility in Use; Simple and Intuitive Use; Perceptible Information; Tolerance for Error; Low Physical Effort; Size and Space for Approach and Use. The original principles have been supplemented by 8 goals of universal design, namely: Body Fit, Comfort, Awareness, Understanding, Wellness, Social Integration, Personalization, Cultural Appro- priateness. https://universaldesign.ie/What-is-Universal-Design. CHAPTER 2 65 Beyond the level of individual infrastruc- Mexico, South Africa, and Chile shows the positive ture or specific systems such as public transport, health impacts of raising taxes on sugar-sweetened the focus in more advanced policy settings is on beverages (SSBs) (Basto-Abro et al. 2019; Stacey et whole-of-community approaches to creating envi- al. 2021). The case for raising tobacco, alcohol and ronments which are accommodative of people with SSB taxes in most LICs and MICs is therefore clear, functional limitations. The focus is often elderly despite the arguments of industry that illicit trade people, but such initiatives are beneficial for any- increases. This has been happening in developing one with functional limitations. This can be seen countries, most commonly through increases in in planning and areas-based approach in countries excise taxes. A CGD study (Lane, Glassman, and such as Singapore and Japan, usually in the context Smitham 2021) found 124 policy commitments of building age-friendly communities. in 43 countries in IMF and World Bank programs on the taxation of alcohol, tobacco, and SSBs, with v. Cross-cutting policies directly impacting more than half relating to tobacco products and one NCDs and their management third to alcoholic beverages. Of course, taxes are not the only instrument While many cross-sectoral policies influence NCDs in seeking to modify these risk factors, albeit they and human capital more broadly, four are worth are the most effective in the case of tobacco. For highlighting as especially salient. The first is tax- example, the WHO Framework Convention on To- ation policy and tax treatment of tobacco, alcohol, bacco Control recommends a multi-prong approach and sweetened beverages (SSBs). While the politi- labelled “MPOWER” i.e., Monitor tobacco use and cal economy of increases taxes on all three is always prevention policies; Protect people from tobacco use challenging (Elliott et al. 2020, the benefits in reduc- (particularly through smoke-free indoor public plac- ing risk behaviors and ultimately NCDs are very well es); Offer help to quit tobacco use; Warn about the documented (Jha et al. 2014). The impact in terms dangers of tobacco; Enforce bans on tobacco adver- of mortality and morbidity, costs to households and tising, promotion, and sponsorship; and Raise taxes reduction in other beneficial forms of consump- on tobacco. Ekpu and Brown (2015) provide an ex- tion, costs to health systems in treating tobacco-re- cellent summary of interventions to reduce smoking lated disease are enormous. WHO estimates over and their economic cost and benefits. Similar mixes 8 million deaths globally per year from tobacco, of policies are relevant for alcohol and SSBs. of whom around 1.2 million are non-smokers, and A second area of direct relevance to NCDs is that around 80 percent of the world’s 1.3 million policies to reduce air pollution. According to WHO’s tobacco users are in LICs and MICs.48 The Asian World Global Ambient Air Quality Database nine in Development Bank (ADB) estimates for Asia also 10 people globally breathe highly polluted air, and suggest that productivity losses from tobacco and in 2019, 99 percent of the world’s population was alcohol use combined run to around 2 percent of living in places where the WHO air quality guide- GDP annually (Lane (2022) for ADB). At the same lines levels were not met.49 The health impacts of time, taxes on tobacco in most developing countries outdoor air pollution are exacerbated by household average less than half of the price of products, as air pollution from burning fuel inside houses, nota- against around two thirds or higher in many HICs bly for women. The combined impacts of outdoor (Verguet et al. 2015; Lane 2022). WHO estimates and household air pollution are about 7 million that an increase in tobacco prices by 10 percent de- premature deaths each year from NCDs (almost 90 creases consumption by about 5 percent in low- and percent of them in developing countries) including middle-income countries. In addition, despite as- from heart disease, chronic obstructive pulmonary sumptions to the contrary, raising taxes appears to disease, lung cancer, acute respiratory infections, benefit the poor more due to their price elasticities, and stroke.50 Exposure also has an equity dimension, so is equitable as well (see Verguet et al. 2015 for with people in developing countries more exposed China). Similarly, evidence from countries such as to both types of pollution, and poor and marginal- 48   https://www.who.int/news-room/fact-sheets/detail/tobacco. 49   https://www.who.int/publications/m/item/who-air-quality-database-2022. 50  https://www.who.int/teams/environment-climate-change-and-health/air-quality-and-health/health-impacts/exposure-air-pollution. 66 CHAPTER 2 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E ized communities within those countries more ex- ital, reducing further deterioration of health condi- posed that their better-off fellow citizens. WHO also tions by allowing recovery time, but also protection notes the specific impacts on women and children, of fellow workers from infectious conditions. In ad- with air pollution increasing low birth weight, pre- dition, unemployment benefits may help to mitigate term birth and small for gestational age births, and the negative mental health effects that have been evidence suggesting that air pollution may affect shown to accompany job loss and unemployment diabetes and neurological development in children. (McKee-Ryan et al. 2005; Paul et al. 2009). See An- The policy responses to limit air pollution are multi- nex Table 2A.3 for examples of social insurance and ple, and include efforts on clean household fuel and other social protection policies which are relevant energy, clean energy technologies, clean use of tech- to NCDs and HC across the life cycle. nology in industry, urban planning and transport The fourth is interventions to improve road practices that reduce pollution, taxes, and market safety and reduce the enormous toll of mortality mechanisms to incentivize green technologies, and and injury from traffic accidents. While not NCDs reforming waste management practices.51 in the main sense of the HLI, road accidents claim The third is expanding coverage of social in- 1.35 million lives per year (over 90 percent of them surance. This applies first and foremost to social in developing countries), and up to 50 million peo- health insurance, where more rapid progression to- ple sustain non-fatal injuries, accounting for out- ward UHC, and inclusion of cost-effective NCD in- put losses of 7 to 22 percent of GDP over a 24-year terventions in essential health packages, has the po- period.52 Addressing the causes of traffic accidents tential to impact NCD prevalence and management includes policies on road and roadside design and in several ways. But is also applies to other forms of operation, safety standards for vehicles, behavioral social insurance such as sickness, workplace injury, issues such as drunk driving and seat belt and hel- disability, and unemployment insurance. For exam- met use, and improving post-crash emergency and ple, sickness cover can be protective of human cap- healthcare services (WHO 2018). CONCLUSION NCDs are one of the major health and development challenges of our time. The preconception that NCDs are mainly conditions of rich countries is increasingly untenable, and the HLI is one of a growing set of efforts to increase the urgency of policy response across the world. An operational framework is imperative to address the impact of the rising toll of NCDs on the lives and wellbeing of people in low and middle-income countries. NCDs reduce the possibility of healthy longevity and their negative impact on human capital and productivity is com- plemented by their effect on wellbeing, itself an intrinsic outcome of development. This chapter has attempted to unbundle the intricate web of relationships between NCDs, human capital, and the end outcomes of healthy longevity, inclusive growth, and wellbeing. The three main channels through which NCDs NCD prevalence and economic growth, though the may impact end outcomes are the macroeconom- causal nature of the relationship remains a subject ic/savings channel, the fiscal channel, and the hu- of debate. NCDs also have clear fiscal impacts on man capital channel. On the first channel, there the expenditure and revenue sides, increasing pub- has been much debate on the impact of popula- lic spending on healthcare and aged care, and re- tion health on economic growth, and an inherent ducing revenues from both workers and firms. At challenge is how to account for non-health factors the same time, taxes on tobacco, alcohol and sug- such as institutions and geography. But a growing ar-sweetened beverages are crucial instruments for body of empirical work, including from developing prevention of NCDs, increasing longevity, and pro- countries, supports a negative relationship between moting future productivity. 51   https://www.who.int/news-room/fact-sheets/detail/ambient-(outdoor)-air-quality-and-health. 52   https://www.worldbank.org/en/news/feature/2019/05/09/how-can-you-help-save-lives-on-the-road. CHAPTER 2 67 In terms of the human capital channel, a life consistently demonstrates that NCDs compromise course perspective is vital, and the discussion has quality of life across various wellbeing measures, and attempted to unbundle how NCDs impact HC from that the relationship is also bidirectional: subjective conception through to old age, and also across gen- measures of wellbeing matter for health (including erations. NCDs compromise the entire trajectory of appearing to be a protective factor), and NCDs im- HC, from its formation to deployment to protection pact subjective measures of wellbeing. and preservation. The immediate and long-term Just as NCDs impact HC and end outcomes impacts of NCDs and risk behaviors play out in the across the life course, policy responses to prevent and labor market in multiple ways, compromising indi- control them are needed from before birth to the end vidual welfare, and productivity and growth at the of life. The chapter has outlined some of the major societal level. The relationships between NCDs and policy areas, with detailed consideration of specific HC, and between them and end outcomes, are also policy domains in various areas of HLI research. The multi-directional. NCDs and HC have a two-way most direct are the areas of nutrition, public health, impact on each other, and NCDs and HC are them- and healthcare services, where the economic and selves affected by levels of economic development social returns to investments are high, and afford- and wellbeing. While the tangled nature of the re- able packages of essential NCD interventions are lationships makes attribution of causality challeng- available. While cost-effectiveness is a key consider- ing, the expanding literature from rich and devel- ation for developing countries in prioritizing limited oping countries increasingly suggests a causal link budgets, it is also important to integrate equity and between NCDs, human capital and end outcomes. financial protection considerations, as well as feasi- Apart from the aggregate impacts of NCDs, the bility of implementation. Beyond the health sector, chapter has highlighted the importance of distribu- there are improvements needed in taxation policies tional implications. In terms of socio-economic sta- to shift risk behaviors (and raise revenue), and pol- tus, most NCDs and risk factors disproportionately icies and interventions to reduce air pollution, im- affect poorer and more vulnerable people within prove road safety, adapt the built environment, and countries, both in terms of their incidence and in expand coverage of social security systems. terms of lower financial protection offered by public Finally, while some of the channels through programs. These disparities appear to be even more which NCDs influence the end outcomes of inter- pronounced within LICs. There are also important est are becoming better understood, there remains gender inequality considerations, both in terms of a huge outstanding research agenda, especially for the NCDs of women themselves and how they are developing countries. The knowledge base on NCDs managed by health systems, and in terms of the and their socioeconomic and wellbeing impacts in disproportionate burden on women of caring for developing countries is growing rapidly, but that evi- NCD-affected family members. dence also points to the diversity of situations across NCDs also have crucial effects on human well- countries and within them. In particular, making the being itself, over and above their impact on human economic case for why NCDs matter for develop- capital, productivity, and growth. That intrinsic ben- ment will be crucial to persuading policymakers of efit is a fundamental motivation for addressing the the urgency of and high returns to enhancing invest- rising toll of NCDs in developing countries. Research ments in NCD prevention and management. on the relationships between NCDs and wellbeing REFERENCES 1. Abdullah, Asnawi. 2015. “The Double Burden of Undernutrition 3. Acemoglu, Daron, and Simon Johnson. 2007. “Disease and and Overnutrition in Developing Countries: An Update.” Cur- Development: The Effect of Life Expectancy on Econom- rent Obesity Reports 4 (3): 337–49. https://doi.org/10.1007/ ic Growth.” Journal of Political Economy 115 (6): 925–85. s13679-015-0170-y. https://doi.org/10.1086/529000. 2. Acemoglu, Daron, Francisco Gallego, and James A. Robinson. 4. Acuña, Carlos, Héctor Acuña, and Diego Carrasco. 2019. “Health 2014. “Institutions, Human Capital and Development.” National Shocks and the Added Worker Effect: A Life Cycle Approach.” Bureau of Economic Research. February 1, 2014. https://doi. Journal of Applied Economics 22 (1): 273–86. https://ideas. org/10.3386/w19933. repec.org/a/taf/recsxx/v22y2019i1p273-286.html. 68 CHAPTER 2 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E 5. Adda, Jérôme, James Banks, and Hans-Martin von Gaudecker. https://ideas.repec.org/p/zbw/hohdps/012017.html. 2009. “The Impact of Income Shocks on Health: Evidence from 16. Banerjee, Abhijit, and Esther Duflo. 2010. Aging and Death un- Cohort Data.” Journal of the European Economic Association der a Dollar a Day. In Research Findings in the Economicsof 7 (6): 1361–99. https://doi.org/10.1162/jeea.2009.7.6.1361. Aging. Edited by David Wise. University of Chicago Press. 6. Aghion, Philippe, and Peter Howitt. 1998. Endogenous Growth 17. Bastagli, Francesca, Jessica Hagen-Zanker, and Georgina Theory. Cambridge, MA: The MIT Press. Sturge. 2016. “Cash Transfers: What Does the Evidence Say? A 7. Allen, Luke, Julianne Williams, Nick Townsend, Bente Mik- Rigorous Review of Impacts and the Role of Design and Imple- kelsen, Nia Roberts, Charlie Foster, and Kremlin Wickramasing- mentation Features.” United Kingdom: Overseas Development he. 2017. “Socioeconomic Status and Non-Communicable Institute (ODI). https://odi.org/en/publications/cash-transfers- Disease Behavioural Risk Factors in Low-Income and Low- what-does-the-evidence-say-a-rigorous-review-of-impacts- er-Middle-Income Countries: A Systematic Review.” The Lancet and-the-role-of-design-and-implementation-features/. Global Health 5 (3): e277–89. https://doi.org/10.1016/s2214- 18. Bastagli, Francesca, Jessica Hagen-Zanker, Luke Harman, Valen- 109x(17)30058-x. tina Barca, Georgina Sturge, Tanja Schmidt, and Luca Pellerano. 8. Angner, Erik, Jennifer Ghandhi, Kristen Williams Purvis, Daniel 2016. “Cash Transfers: What Does the Evidence Say? A Rigorous Amante, and Jeroan Allison. 2012. “Daily Functioning, Health Sta- Review of Programme Impact and of the Role of Design and Im- tus, and Happiness in Older Adults.” Journal of Happiness Stud- plementation Features.” London: Overseas Development Institute. ies 14 (5): 1563–74. https://doi.org/10.1007/s10902-012-9395-6. 19. Basto-Abreu, Ana, Tonatiuh Barrientos-Gutiérrez, Dèsirée 9. Arabi, Yaseen M., Hanan H. Balkhy, Frederick G. Hayden, Ab- Vidaña-Pérez, M. Arantxa Colchero, Mauricio Hernández-F., derrezak Bouchama, Thomas Luke, J. Kenneth Baillie, Awad Mauricio Hernández-Ávila, Zachary J. Ward, Michael W. Long, Al-Omari, et al. 2017. “Middle East Respiratory Syndrome.” New and Steven L. Gortmaker. 2019. “Cost-Effectiveness of the Sug- England Journal of Medicine 376 (6): 584–94. https://doi. ar-Sweetened Beverage Excise Tax in Mexico.” Health Affairs 38 org/10.1056/nejmsr1408795. (11): 1824–31. https://doi.org/10.1377/hlthaff.2018.05469. 10. Armengaud, J. B., C. Yzydorczyk, B. Siddeek, A. C. Peyter, and U. 20. Beaglehole, Robert, Ruth Bonita, and Roger S. Magnusson. Simeoni. 2021. “Intrauterine Growth Restriction: Clinical Conse- 2011. “Global Cancer Prevention: An Important Pathway to quences on Health and Disease at Adulthood.” Reproductive Global Health and Development.” Public Health 125 (12): 821– Toxicology 99 (January): 168–76. https://doi.org/10.1016/j. 31. https://doi.org/10.1016/j.puhe.2011.09.029. reprotox.2020.10.005. 21. Ben-Porath, Yoram. 1967. “The Production of Human Capital 11. Arokiasamy, Perianayagam, Uttamacharya Uttamacharya, and the Life Cycle of Earnings.” Journal of Political Economy Kshipra Jain, Richard Berko Biritwum, Alfred Edwin Yawson, 75 (4, Part 1): 352–65. https://doi.org/10.1086/259291. Fan Wu, Yanfei Guo, et al. 2015. “The Impact of Multimorbidity 22. Bennett, James E, Gretchen A Stevens, Colin D Mathers, Ruth on Adult Physical and Mental Health in Low- and Middle-In- Bonita, Jürgen Rehm, Margaret E Kruk, Leanne M Riley, et al. come Countries: What Does the Study on Global Ageing and 2018. “NCD Countdown 2030: Worldwide Trends in Non-Com- Adult Health (SAGE) Reveal?” BMC Medicine 13 (1). https://doi. municable Disease Mortality and Progress towards Sustainable org/10.1186/s12916-015-0402-8. Development Goal Target 3.4.” The Lancet 392 (10152): 1072– 12. Arrow, Kenneth J., Partha Dasgupta, Lawrence H. Goulder, Kev- 88. https://doi.org/10.1016/s0140-6736(18)31992-5. in J. Mumford, and Kirsten Oleson. 2013. “Sustainability and the 23. Bertuccio, Paola, Gianfranco Alicandro, Matteo Malvezzi, Gre- Measurement of Wealth: Further Reflections.” Environment ta Carioli, Paolo Boffetta, Fabio Levi, Carlo La Vecchia, and Eva and Development Economics 18 (4): 504–16. https://doi. Negri. 2020. “Childhood Cancer Mortality Trends in Europe, org/10.1017/s1355770x13000193. 1990-2017, with Focus on Geographic Differences.” Cancer 13. Ashraf, Quamrul H., Ashley Lester, and David N. Weil. Epidemiology 67 (August): 101768. https://doi.org/10.1016/j. 2008. “When Does Improving Health Raise GDP?” NBER canep.2020.101768. Macroeconomics Annual 23 (1): 157–204. https://doi. 24. Bhalotra, Sonia, and Samantha B. Rawlings. 2011. “Intergen- org/10.1086/593084. erational Persistence in Health in Developing Countries: The 14. Atella, Vincenzo, and Joanna Kopinska. 2014. “Body Weight, Penalty of Gender Inequality?” Journal of Public Economics 95 Eating Patterns, and Physical Activity: The Role of Education.” (3-4): 286–99. https://doi.org/10.1016/j.jpubeco.2010.10.016. Demography 51 (4): 1225–49. https://doi.org/10.1007/ 25. Bhutta, Zulfiqar A, Tahmeed Ahmed, Robert E Black, Simon s13524-014-0311-z. Cousens, Kathryn Dewey, Elsa Giugliani, Batool A Haider, et 15. Baldanzi, Annarita, Alberto Bucci, and Klaus Prettner. 2017. al. 2008. “What Works? Interventions for Maternal and Child “Children’s Health, Human Capital Accumulation, and R&D- Undernutrition and Survival.” The Lancet 371 (9610): 417–40. Based Economic Growth.” Hohenheim Discussion Papers in https://doi.org/10.1016/s0140-6736(07)61693-6. Business, Economics and Social Sciences. University of Ho- 26. Black, Robert E, Cesar G Victora, Susan P Walker, Zulfiqar A Bhutta, henheim, Faculty of Business, Economics and Social Sciences. Parul Christian, Mercedes de Onis, Majid Ezzati, et al. 2013. “Ma- CHAPTER 2 69 ternal and Child Undernutrition and Overweight in Low-Income Adults Aged 20-79 Years: A Cost-of-Illness Study.” The Lan- and Middle-Income Countries.” The Lancet 382 (9890): 427–51. cet. Diabetes & Endocrinology 5 (6): 423–30. https://doi. 27. Black, Sandra E., Paul J. Devereux , and Kjell G. Salvanes. 2012. org/10.1016/S2213-8587(17)30097-9. “Losing Heart? The Effect of Job Displacement on Health.” NBER 39. Bonita, Ruth, and Robert Beaglehole. 2014. “Women and NCDs: Working Paper W18660. Cambridge, MA: National Bureau of Overcoming the Neglect.” Global Health Action 7 (1): 23742. Economic Research. https://papers.ssrn.com/sol3/papers. https://doi.org/10.3402/gha.v7.23742. cfm?abstract_id=2194785. 40. Boucekkine, Raouf, David de la Croix, and Omar Licandro. 2002. 28. Blackstone, Tanja, and Kristen Rensselaer. 2021. “Does Edu- “Vintage Human Capital, Demographic Trends, and Endogenous cation Contribute to Happiness? A Cross-Country Study.” Re- Growth.” Journal of Economic Theory 104 (2): 340–75. https:// search in Business and Economics Journal 14. https://www. ideas.repec.org/a/eee/jetheo/v104y2002i2p340-375.html. aabri.com/manuscripts/203179.pdf. 41. Braakmann, Nils. 2011. “The Causal Relationship between Edu- 29. Blazer, Dan G., Kristine Yaffe, and Jason Karlawish. 2015. “Cog- cation, Health and Health Related Behaviour: Evidence from a nitive Aging.” JAMA 313 (21): 2121. https://doi.org/10.1001/ Natural Experiment in England.” Journal of Health Economics jama.2015.4380. 30 (4): 753–63. https://doi.org/10.1016/j.jhealeco.2011.05.015. 30. Bleakley, Hoyt. 2010. “Health, Human Capital, and Develop- 42. Brouwers, Evelien P. M. 2020. “Social Stigma Is an Underestimat- ment.” Annual Review of Economics 2 (1): 283–310. https:// ed Contributing Factor to Unemployment in People with Men- doi.org/10.1146/annurev.economics.102308.124436. tal Illness or Mental Health Issues: Position Paper and Future 31. Bloom, David E., Elizabeth T. Cafiero-Fonseca, Mark E. McGov- Directions.” BMC Psychology 8 (1). https://doi.org/10.1186/ ern, Klaus Prettner, Anderson Stanciole, Jonathan Weiss, Sam- s40359-020-00399-0. uel Bakkila, and Larry Rosenberg. 2014. “The Macroeconomic 43. Buckles, Kasey, Andreas Hagemann, Ofer Malamud, Melinda Impact of Non-Communicable Diseases in China and India: Morrill, and Abigail Wozniak. 2013. “The Effect of College Ed- Estimates, Projections, and Comparisons.” The Journal of the ucation on Health.” NBER Working Paper 19222. Cambridge, Economics of Ageing 4 (December): 100–111. https://doi. MA: National Bureau of Economic Research. https://econpa- org/10.1016/j.jeoa.2014.08.003. pers.repec.org/paper/nbrnberwo/19222.htm. 32. Bloom, David E., David Canning, and Bryan Graham. 2003. 44. Bundy, Donald A.P., Nilanthi de Silva, Susan Horton, George C. “Longevity and Life-Cycle Savings.” The Scandinavian Jour- Patton, Linda Schultz, and Dean T. Jamison. 2017. “Child and nal of Economics 105 (3): 319–38. https://www.jstor.org/sta- Adolescent Health and Development: Realizing Neglected ble/3440944. Potential.” In Child and Adolescent Health and Development. 33. Bloom, David E., David Canning, and Jaypee Sevilla. 2004. “The 3rd Edition. Washington, D.C.: World Bank. doi: 10.1596/978-1- Effect of Health on Economic Growth: A Production Function 4648-0423-6_ch1. Approach.” World Development 32 (1): 1–13. https://doi. 45. Cancelliere, Carol, J David Cassidy, Carlo Ammendolia, and Pierre org/10.1016/j.worlddev.2003.07.002. Côté. 2011. “Are Workplace Health Promotion Programs Effective 34. Bloom, David E., Michael Kuhn, and Klaus Prettner. 2018. at Improving Presenteeism in Workers? A Systematic Review and “Health and Economic Growth.” IZA Discussion Paper 11939. Best Evidence Synthesis of the Literature.” BMC Public Health 11 Bonn, Germany: IZA Institute of Labor Economics. https://pa- (1). https://doi.org/10.1186/1471-2458-11-395. pers.ssrn.com/sol3/papers.cfm?abstract_id=3301688. 46. Capatina, Elena, Michael Keane, and Shiko Maruyama. 2020. 35. Bloom, David, Michael Kuhn, and Klaus Prettner. 2015. “The “Health Shocks and the Evolution of Earnings over the Life-Cy- Contribution of Female Health to Economic Development.” cle.” UNSW Economics Working Paper 2018-14a. Sydney, Aus- NBER Working Paper 21411. Cambridge, MA: National Bureau tralia: School of Economics, The University of New South Wales. of Economic Research. https://www.nber.org/system/files/ https://ideas.repec.org/p/swe/wpaper/2018-14b.html. working_papers/w21411/w21411.pdf. 47. Carney, Robert M., and Kenneth E. Freedland. 2016. “Depres- 36. BMJ. 2018. “The Health, Poverty, and Financial Consequences sion and Coronary Heart Disease.” Nature Reviews Cardiology of a Cigarette Price Increase among 500 Million Male Smokers 14 (3): 145–55. https://doi.org/10.1038/nrcardio.2016.181. in 13 Middle Income Countries: Compartmental Model Study.” 48. Cervellati, Matteo, and Uwe Sunde. 2009. “Life Expectancy and BMJ, April, k1162. https://doi.org/10.1136/bmj.k1162. Economic Growth: The Role of the Demographic Transition.” 37. Boehm, Julia K., and Laura D. Kubzansky. 2012. “The Heart’s IZA Discussion Paper No. 4160. Bonn, Germany: Institute for Content: The Association between Positive Psychological the Study of Labor (IZA). https://doi.org/10.2139/ssrn.1405928. Well-Being and Cardiovascular Health.” Psychological Bulletin 49. Chakraborty, Debapriya, Daphne Wu, and Prabhat Jha. 2024. 138 (4): 655–91. https://doi.org/10.1037/a0027448. “Exploring the Labour Market Outcomes of the Risk Factors for 38. Bommer, Christian, Esther Heesemann, Vera Sagalova, Jennifer Non-communicable Diseases: A Systematic Review.” In Unlocking Manne-Goehler, Rifat Atun, Till Bärnighausen, and Sebastian the Power of Healthy Longevity: Compendium of Research for Vollmer. 2017. “The Global Economic Burden of Diabetes in the Healthy Longevity Initiative. Washington, D.C.: World Bank. 70 CHAPTER 2 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E 50. Elliott, Lana M, Sarah L Dalglish and Stephanie M. Topp. 2020. spective Observational Studies.” Psychosomatic Medicine 70 “Health Taxes on Tobacco, Alcohol, Food and Drinks in Low- (7): 741–56. https://doi.org/10.1097/psy.0b013e31818105ba. and Middle-Income Countries: A Scoping Review of Policy 62. Chłoń-Domińczak , Agnieszka, and Dariusz Stańko. 2015. “Are Content, Actors, Process and Context.” International Journal Survivor Pensions Still Needed?” In OECD Pensions Outlook of Health Policy and Management 11: 414 - 428. 2018. Paris: OECD Publishing. https://www.oecd-ilibrary.org/ 51. International Food Policy Research Institute (IFPRI). 2016. Glob- sites/pens_outlook-2018-10-en/index.html?itemId=/content/ al Nutrition Report 2016: From Promise to Impact: Ending Mal- component/pens_outlook-2018-10-en. nutrition by 2030. Washington, DC 63. Clark, Andrew E., Sarah Flèche, Richard Layard, Nattavudh 52. International LabourOrganization (ILO). “Care work and care Powdthavee, and George Ward. 2017. “The Key Determinants jobs for the future of decent work.” Geneva, Switzerland: Inter- of Happiness and Misery.” CEP Discussion Papers. UK: Centre national Labour Office. for Economic Performance, LSE. https://ideas.repec.org/p/cep/ 53. Save the Children. 2008. “Saving the lives of mothers and new- cepdps/dp1485.html. borns inAfghanistan: Afghanistan newborn health situation 64. Clark, Damon, and Heather Royer. 2013. “The Effect of Edu- analysis.” Kabul: Save the Children. cation on Adult Mortality and Health: Evidence from Britain.” 54. ———. 2015. “The Economic and Demographic Transition, American Economic Review 103 (6): 2087–2120. https://doi. Mortality, and Comparative Development.” American Eco- org/10.1257/aer.103.6.2087. nomic Journal: Macroeconomics 7 (3): 189–225. https://doi. 65. Conti, Gabriella, and Christopher Hansman. 2013. “Personality org/10.1257/mac.20130170. and the Education–Health Gradient: A Note on ‘Understand- 55. Chaker, Layal, Abby Falla, Sven J. van der Lee, Taulant Muka, Da- ing Differences in Health Behaviors by Education.’” Journal of vid Imo, Loes Jaspers, Veronica Colpani, et al. 2015. “The Global Health Economics 32 (2): 480–85. https://doi.org/10.1016/j. Impact of Non-Communicable Diseases on Macro-Economic jhealeco.2012.07.005. Productivity: A Systematic Review.” European Journal of Ep- 66. Conti, Gabriella, and James J. Heckman. 2010. “Understanding idemiology 30 (5): 357–95. https://doi.org/10.1007/s10654- the Early Origins of the Education–Health Gradient.” Perspec- 015-0026-5. tives on Psychological Science 5 (5): 585–605. https://doi. 56. Chan, J.W.M., C.K. Ng, Y.H. Chan, T.Y.W. Mok, W.H.O.S. Lee, S.Y.Y. org/10.1177/1745691610383502. Chu, W.L. Law, M.P. Lee, and P.C.K. Li. 2003. “Short Term Out- 67. Conti, Gabriella, James Heckman, and Sergio Urzua. 2010. “The come and Risk Factors for Adverse Clinical Outcomes in Adults Education-Health Gradient.” American Economic Review 100 with Severe Acute Respiratory Syndrome (SARS).” Thorax 58 (2): 234–38. https://doi.org/10.1257/aer.100.2.234. (8): 686–89. https://doi.org/10.1136/thorax.58.8.686. 68. Currie, Janet, and Mark Stabile. 2006. “Child Mental Health 57. Chang, Anna Marie, Bryn Mumma, Keara L. Sease, Jennifer L. and Human Capital Accumulation: The Case of ADHD.” Jour- Robey, Frances S. Shofer, and Judd E. Hollander. 2007. “Gen- nal of Health Economics 25 (6): 1094–1118. https://doi. der Bias in Cardiovascular Testing Persists after Adjustment org/10.1016/j.jhealeco.2006.03.001. for Presenting Characteristics and Cardiac Risk.” Academic 69. Cutler, David, and Adriana Lleras-Muney. 2006. “Education and Emergency Medicine: Official Journal of the Society for Ac- Health: Evaluating Theories and Evidence.” NBER Working Pa- ademic Emergency Medicine 14 (7): 599–605. https://doi. per 12352. Cambridge, MA: National Bureau of Economic Re- org/10.1197/j.aem.2007.03.1355. search. https://www.nber.org/system/files/working_papers/ 58. Charlton, Karen E., and Donald Rose. 2001. “Nutrition among w12352/w12352.pdf. Older Adults in Africa: The Situation at the Beginning of the 70. Das, Maitreyi B., Arai Yuko, Terri B. Chapman, and Vibhu Jain. Millenium.” The Journal of Nutrition 131 (9): 2424S2428S. 2022. “Silver Hues: Building Age-Ready Cities.” Washington, https://doi.org/10.1093/jn/131.9.2424S. D.C.: World Bank. http://hdl.handle.net/10986/37259. 59. Charmes, Jacques. 2019. “The Unpaid Care Work and the La- 71. Davidson, Karina W., Elizabeth Mostofsky, and William Whang. bour Market. An Analysis of Time Use Data Based on the Latest 2010. “Don’t Worry, Be Happy: Positive Affect and Reduced 10- World Compilation of Time-Use Surveys.” Geneva, Switzerland: Year Incident Coronary Heart Disease: The Canadian Nova Sco- International Labour Office. https://www.ilo.org/wcmsp5/ tia Health Survey.” European Heart Journal 31 (9): 1065–70. groups/public/---dgreports/---gender/documents/publica- https://doi.org/10.1093/eurheartj/ehp603. tion/wcms_732791.pdf. 72. Deaton, Angus. 2008. “Income, Health, and Well-Being around 60. Chatenoud, Liliane, Paola Bertuccio, Cristina Bosetti, Fabio Levi, the World: Evidence from the Gallup World Poll.” Journal of Eva Negri, and Carlo La Vecchia. 2010. “Childhood Cancer Mor- Economic Perspectives 22 (2): 53–72. https://doi.org/10.1257/ tality in America, Asia, and Oceania, 1970 through 2007.” Can- jep.22.2.53. cer 116 (21): 5063–74. https://doi.org/10.1002/cncr.25406. 73. Deaton, Angus S., and Christina H. Paxson. 1998. “Aging and 61. Chida, Yoichi, and Andrew Steptoe. 2008. “Positive Psycholog- Inequality in Income and Health.” The American Economic ical Well-Being and Mortality: A Quantitative Review of Pro- Review 88 (2): 248–53. https://www.jstor.org/stable/116928. CHAPTER 2 71 74. Debas, Haile T., Peter Donkor, Atul Gawande, Dean T. Jamison, Dean T. Jamison, Hellen Gelband, Susan Horton, Prabhat Jha, Margaret E. Kruk, and Charles N. Mock, eds. 2015. Essential Ramanan Laxminarayan, Charles N. Mock, and Rachel Nugent. Surgery. Disease Control Priorities, Third Edition (Volume 1). PubMed. Washington (DC): The International Bank for Recon- Washington, D.C.: World Bank. struction and Development / The World Bank. 2017. https:// 75. Desbordes, Rodolphe. 2011. “The Non-Linear Effects of Life pubmed.ncbi.nlm.nih.gov/30212160/. Expectancy on Economic Growth.” Economics Letters 112 (1): 86. Etilé, Fabrice, and Andrew M. Jones. 2011. “Schooling and 116–18. https://doi.org/10.1016/j.econlet.2011.03.027. Smoking among the Baby Boomers – an Evaluation of the 76. Devercelli, Amanda E., and Frances Beaton-Day. 2020. “Bet- Impact of Educational Expansion in France.” Journal of Health ter Jobs and Brighter Futures: Investing in Childcare to Build Economics 30 (4): 811–31. https://doi.org/10.1016/j.jheale- Human Capital.” Washington, D.C.: World Bank. https://open- co.2011.05.002. knowledge.worldbank.org/server/api/core/bitstreams/a5e- 87. Evans, Brooks F., and Robert J. Palacios. 2015. “An Examina- 7a52e-115c-5dd1-97e6-c1b062c945c9/content. tion of Elderly Co-Residence in the Developing World.” So- 77. Devries, K. M., J. Y. T. Mak, C. García-Moreno, M. Petzold, J. C. cial Protection & Labor Policy Note No. 17. Washington, Child, G. Falder, S. Lim, et al. 2013. “The Global Prevalence of D.C.: World Bank. http://documents.worldbank.org/curated/ Intimate Partner Violence against Women.” Science 340 (6140): en/600291467992481736/An-examination-of-elderly-co-resi- 1527–28. https://doi.org/10.1126/science.1240937. dence-in-the-developing-world. 78. Dinerstein, Michael, Rigissa Megalokonomou, and Constan- 88. Evans, David, Maryam Akmal, and Pamela Jakiela. 2020. “Gen- tine Yannelis. 2022. “Human Capital Depreciation and Returns der Gaps in Education: The Long View.” Washington, D.C.: Cen- to Experience.” NBER Working Paper 27925. Vol. 112. Cam- ter for Global Development. https://www.cgdev.org/publica- bridge, MA: National Bureau of Economic Research. https:// tion/gender-gaps-education-long-view. doi.org/10.1257/aer.20201571. 89. Eze, Paul, Lucky Osaheni Lawani, Ujunwa Justina Agu, and 79. Do, Young Kyung, and Mary Ann Bautista. 2015. “Tobacco Use Yubraj Acharya. 2022. “Catastrophic Health Expenditure in and Household Expenditures on Food, Education, and Health- Sub-Saharan Africa: Systematic Review and Meta-Analysis.” care in Low- and Middle-Income Countries: A Multilevel Analy- Bulletin of the World Health Organization 100 (05): 337–51J. sis.” BMC Public Health 15 (1). https://doi.org/10.1186/s12889- https://doi.org/10.2471/blt.21.287673. 015-2423-9. 90. Field, John. 2013. “Learning through the Ages? Generational In- 80. Donfouet, Hermann P.P., Shukri F. Mohamed, and Eric Malin. equalities and Inter-Generational Dynamics of Lifelong Learn- 2021. “Socioeconomic Inequality in Tobacco Use in Kenya: A ing.” British Journal of Educational Studies 61 (1): 109–19. Concentration Analysis.” International Journal of Health Eco- https://doi.org/10.1080/00071005.2012.756172. nomics and Management, January. https://doi.org/10.1007/ 91. Finkelstein, Eric. A., Wan. C.K. Graham, and Rahul Malhotra. 2014. s10754-020-09292-0. “Lifetime Direct Medical Costs of Childhood Obesity.” PEDIAT- 81. Doshi, Kokila. 1994. “Determinants of the Savings Rate: An RICS 133 (5): 854–62. https://doi.org/10.1542/peds.2014-0063. International Comparison.” Contemporary Economic Poli- 92. Folbre, Nancy. 2015. “Valuing Non-Market Work.” Think Piece cy 12 (1): 37–45. https://doi.org/10.1111/j.1465-7287.1994. for 2015 UNDP Human Development Report . New York, NY: tb00410.x. United Nations Development Programme. https://hdr.undp. 82. Dumas, Julie A. 2017. “Strategies for Preventing Cog- org/system/files/documents/folbrehdr2015finalpdf.pdf. nitive Decline in Healthy Older Adults.” The Canadi- 93. Galama, Titus J., Adriana Lleras-Muney, and Hans van Kipper- an Journal of Psychiatry 62 (11): 754–60. https://doi. sluis. 2018. “The Effect of Education on Health and Mortality: org/10.1177/0706743717720691. A Review of Experimental and Quasi-Experimental Evidence.” 83. Ekpu, Victor U., and Abraham K. Brown. 2015. “The Econom- NBER Working Paper 24225. Cambridge, MA: National Bureau ic Impact of Smoking and of Reducing Smoking Prevalence: of Economic Research. https://ideas.repec.org/p/nbr/nber- Review of Evidence.” Tobacco Use Insights 8 (8): TUI.S15628. wo/24225.html. https://doi.org/10.4137/tui.s15628. 94. García, Sandra, and Juan E. Saavedra. 2017. “Educational Im- 84. Eliason, Marcus, and Donald Storrie. 2009. “Job Loss Is Bad pacts and Cost-Effectiveness of Conditional Cash Transfer for Your Health – Swedish Evidence on Cause-Specific Hos- Programs in Developing Countries: A Meta-Analysis.” Re- pitalization Following Involuntary Job Loss.” Social Science view of Educational Research 87 (5): 921–65. https://doi. & Medicine 68 (8): 1396–1406. https://doi.org/10.1016/j. org/10.3102/0034654317723008. socscimed.2009.01.021. 95. Garrow, Valerie. 2016. “Presenteeism: A Review of Current 85. Essue, Beverley M., Tracey-Lea Laba, Felicia Knaul, Annie Chu, Thinking.” Report 507. Brighton, UK: Institute for Employment Hoang Van Minh, Thi Kim Phuong Nguyen, and Stephen Jan. Studies. https://www.employment-studies.co.uk/system/files/ 2017. “Economic Burden of Chronic Ill Health and Injuries for resources/files/507_0.pdf. Households in Low- and Middle-Income Countries.” Edited by 96. Glinskaya, Elena E., and Zhanlian Feng. 2018. Building an Effi- 72 CHAPTER 2 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E cient and Sustainable Aged Care System : Option for Aged 107. Heckman, James J, and Ganesh Karapakula. 2019. “Intergen- Care in China. Washington, D.C.: World Bank Group. erational and Intragenerational Externalities of the Perry Pre- 97. Global Tobacco Economics Consortium. 2018. “The Health, school Project.” NBER Working Paper 25889. Cambridge, MA: Poverty, and Financial Consequences of a Cigarette Price In- National Bureau of Economic Research. https://www.nber. crease among 500 Million Male Smokers in 13 Middle Income org/papers/w25889. Countries: Compartmental Model Study.” BMJ, April, k1162. 108. Heckman, James J., John Eric Humphries, and Greg Veramendi. https://doi.org/10.1136/bmj.k1162. 2016. “Returns to Education: The Causal Effects of Education on 98. Gore, Fiona M, Paul J N Bloem, George C Patton, Jane Ferguson, Earnings, Health and Smoking.” IZA Discussion Paper No. 9957. Véronique Joseph, Carolyn Coffey, Susan M Sawyer, and Colin Bonn, Germany: Institute of Labor Economics (IZA). https://doi. D Mathers. 2011. “Global Burden of Disease in Young People org/10.2139/ssrn.2786041. Aged 10-24 Years: A Systematic Analysis.” Lancet 377 (9783): 109. Heckman, James J., and Thomas E. Macurdy. 1980. “A Life Cy- 2093–2102. https://doi.org/10.1016/S0140-6736(11)60512-6. cle Model of Female Labour Supply.” The Review of Economic 99. Greenhalgh, Elizabeth M., Michelle M. Scollo, and Merryn Studies 47 (1): 47. https://doi.org/10.2307/2297103. Pearce. 2016. “The Relationship between Tobacco Smoking 110. Hedayati, S. Susan. 2010. “Association between Major Depres- and Financial Stress - Tobacco in Australia.” In Tobacco in Aus- sive Episodes in Patients with Chronic Kidney Disease and tralia: Facts and Issues, edited by Elizabeth M. Greenhalgh, Initiation of Dialysis, Hospitalization, or Death.” JAMA 303 (19): Michelle M. Scollo, and Margaret H. Winstanley. Melbourne, 1946. https://doi.org/10.1001/jama.2010.619. Australia: Cancer Council Victoria. https://www.tobaccoinaus- 111. Heidari-Beni, Motahar. 2019. “Early Life Nutrition and Non tralia.org.au/chapter-9-disadvantage/9-4-the-relationship-be- Communicable Disease.” Advances in Experimental Medi- tween-tobacco-smoking-and-f. cine and Biology, 33–40. https://doi.org/10.1007/978-3-030- 100. Grossman, Michael. 2006. “Education and Nonmarket Out- 10616-4_4. comes.” In Handbook of the Economics of Education, edited 112. Heide, Iris van der, Rogier M van Rijn, Suzan JW Robroek, Alex by Eric Hanushek and Finis Welch, 1:577–633. Elsevier. https:// Burdorf, and Karin I Proper. 2013. “Is Retirement Good for Your ideas.repec.org/h/eee/educhp/1-10.html. Health? A Systematic Review of Longitudinal Studies.” BMC Pub- 101. Gupta, Nidhi, Kashish Goel, Priyali Shah, and Anoop Misra. lic Health 13 (1). https://doi.org/10.1186/1471-2458-13-1180. 2012. “Childhood Obesity in Developing Countries: Epidemiol- 113. Helliwell, John F, Richard Layard, and Jeffrey Sachs. 2017. ogy, Determinants, and Prevention.” Endocrine Reviews 33 (1): World Happiness Report 2017. New York, NY: Sustainable 48–70. https://doi.org/10.1210/er.2010-0028. Development Solutions Network. https://worldhappiness. 102. Haakenstad, Annie, Matthew Coates, Gene Buhkman, Margaret report/ed/2017/. McConnell, and Stéphane Verguet. 2022. “Comparative Health 114. Helliwell, John F., and Robert D. Putnam. 1999. “Education and Systems Analysis of Differences in the Catastrophic Health Ex- Social Capital.” NBER Working Paper No. W7121. Cambridge, penditure Associated with Non-Communicable versus Commu- MA: National Bureau of Economic Research. nicable Diseases among Adults in Six Countries.” Health Policy 115. Henkel, Dieter. 2011. “Unemployment and Substance Use: A and Planning, July. https://doi.org/10.1093/heapol/czac053. Review of the Literature (1990-2010).” Current Drug Abuse 103. Hackett, Ruth A., and Andrew Steptoe. 2017. “Type 2 Diabetes Reviewse 4 (1): 4–27. https://doi.org/10.2174/18744737111 Mellitus and Psychological Stress — a Modifiable Risk Factor.” 04010004. Nature Reviews Endocrinology 13 (9): 547–60. https://doi. 116. Höfelmann, Doroteia A, David A Gonzalez-Chica, Karen Glazer org/10.1038/nrendo.2017.64. Peres, Antonio Fernando Boing, and Marco Aurelio Peres. 2017. 104. Handa, Sudhanshu, Silvio Daidone, Amber Peterman, Benja- “Chronic Diseases and Socioeconomic Inequalities in Quality of min Davis, Audrey Pereira, Tia Palermo, and Jennifer Yablonski. Life among Brazilian Adults: Findings from a Population-Based 2018. “Myth-Busting? Confronting Six Common Perceptions Study in Southern Brazil.” European Journal of Public Health about Unconditional Cash Transfers as a Poverty Reduction 28 (4): 603–10. https://doi.org/10.1093/eurpub/ckx224. Strategy in Africa.” The World Bank Research Observer 33 (2): 117. Hokayem, Charles, and James P. Ziliak. 2014. “Health, Human 259–98. https://doi.org/10.1093/wbro/lky003. Capital, and Life Cycle Labor Supply.” American Economic Re- 105. Hansen, Casper Worm. 2013. “Life Expectancy and Human view 104 (5): 127–31. https://doi.org/10.1257/aer.104.5.127. Capital: Evidence from the International Epidemiological Tran- 118. Horioka, Charles Yuji, and Akiko Terada-Hagiwara. 2012. “The sition.” Journal of Health Economics 32 (6): 1142–52. https:// Determinants and Long-Term Projections of Saving Rates in doi.org/10.1016/j.jhealeco.2013.09.011. Developing Asia.” Japan and the World Economy 24 (2): 128– 106. Hanson, M. A., and P. D. Gluckman. 2014. “Early Developmental 37. https://doi.org/10.1016/j.japwor.2012.01.006. Conditioning of Later Health and Disease: Physiology or Patho- 119. Hosseinpoor, Ahmad Reza, Nicole Bergen, Shanthi Mendis, physiology?” Physiological Reviews 94 (4): 1027–76. https:// Sam Harper, Emese Verdes, Anton Kunst, and Somnath Chat- doi.org/10.1152/physrev.00029.2013. terji. 2012. “Socioeconomic Inequality in the Prevalence of CHAPTER 2 73 Noncommunicable Diseases in Low- and Middle-Income https://doi.org/10.1186/1478-4505-11-31. Countries: Results from the World Health Survey.” BMC Public 130. Karakochuk, Crystal D., Kyly C. Whitfield, Tim J. Green, and Krae- Health 12 (1). https://doi.org/10.1186/1471-2458-12-474. mer Klaus, eds. 2018. “The Biology of the First 1,000 Days.” Boca 120. Hsieh, Chang-Tai, and Pete Klenow. 2010. “Development Ac- Raton, FL: CRC Press. counting.” American Economic Journal: Macroeconomics 2 131. Kaur, Damanpreet, Prasad Rasane, Jyoti Singh, Sawinder Kaur, (1): 207–23. https://doi.org/10.1257/mac.2.1.207. Vikas Kumar, Dipendra Kumar Mahato, Anirban Dey, Kajal Dha- 121. Hyclak, Thomas J., Christopher L. Skeels, and Larry W. Taylor. wan, and Sudhir Kumar. 2019. “Nutritional Interventions for 2016. “The Cardiovascular Revolution and Economic Perfor- Elderly and Considerations for the Development of Geriatric mance in the OECD Countries.” Journal of Macroeconomics Foods.” Current Aging Science 12 (1): 15–27. https://doi.org/ 50 (C): 114–25. https://ideas.repec.org/a/eee/jmacro/v50y- 10.2174/1874609812666190521110548. 2016icp114-125.html. 132. Kilian, Reinhold, and Thomas Becker. 2007. “Macro-Econom- 122. Izquierdo, Mikel, R. A. Merchant, J. E. Morley, S. D. Anker, I. Apra- ic Indicators and Labour Force Participation of People with hamian, H. Arai, M. Aubertin-Leheudre, et al. 2021. “International Schizophrenia.” Journal of Mental Health 16 (2): 211–22. Exercise Recommendations in Older Adults (ICFSR): Expert Con- https://doi.org/10.1080/09638230701279899. sensus Guidelines.” The Journal of Nutrition, Health & Aging 25 133. Kim, Kyeongkuk, Sang-Hyop Lee, and Timothy J. Halliday. 2018. (7): 824–53. https://doi.org/10.1007/s12603-021-1665-8. “Health Shocks, the Added Worker Effect, and Labor Supply in 123. Jan, Stephen, Tracey-Lea Laba, Beverley M Essue, Adrian Gheo- Married Couples: Evidence from South Korea.” Department rghe, Janani Muhunthan, Michael Engelgau, Ajay Mahal, et al. of Economics Working Paper No. 18-12. Honolulu: Univer- 2018. “Action to Address the Household Economic Burden of sity of Hawaii at Manoa. https://ideas.repec.org/p/hai/wpa- Non-Communicable Diseases.” The Lancet 391 (10134): 2047– per/201812.html. 58. https://doi.org/10.1016/s0140-6736(18)30323-4. 134. Kinugasa, Tomoko, and Andrew Mason. 2007. “Why Countries 124. Jaspers, Loes, Veronica Colpani, Layal Chaker, Sven J. van der Become Wealthy: The Effects of Adult Longevity on Saving.” Lee, Taulant Muka, David Imo, Shanthi Mendis, et al. 2014. “The World Development 35 (1): 1–23. https://doi.org/10.1016/j. Global Impact of Non-Communicable Diseases on Households worlddev.2006.09.002. and Impoverishment: A Systematic Review.” European Jour- 135. Knapp, Martin, and Gloria Wong. 2020. “Economics and Mental nal of Epidemiology 30 (3): 163–88. https://doi.org/10.1007/ Health: The Current Scenario.” World Psychiatry 19 (1): 3–14. s10654-014-9983-3. https://doi.org/10.1002/wps.20692. 125. Jayachandran, Seema, and Adriana Lleras-Muney. 2009. “Life 136. Knaul, Felicia Marie, Beverley M Essue, Héctor Arreola-Or- Expectancy and Human Capital Investments: Evidence from nelas, David Watkins, and Ana Langer. 2021. “Universal Health Maternal Mortality Declines*.” Quarterly Journal of Economics Coverage Must Become a Best Buy for Women.” The Lan- 124 (1): 349–97. https://doi.org/10.1162/qjec.2009.124.1.349. cet 398 (10318): 2215–17. https://doi.org/10.1016/s0140- 126. Jha, Prabhat, Mary MacLennan, Frank J. Chaloupka, Ayda Yurek- 6736(21)02755-0. li, Chintanie Ramasundarahettige, Krishna Palipudi, WItold 137. Kraay, Aart. 2018. “Methodology for a World Bank Human Cap- Zatonsky, Samira Asma, and Prakash C. Gupta. 2014. “Global ital Index.” Policy Research Working Paper 8593. Washington, Hazards of Tobacco and the Benefits of Tobacco Cessation D.C.: World Bank. https://papers.ssrn.com/sol3/papers.cfm?ab- and Tobacco Taxes.” In Disease Control Priorities (Volume 3): stract_id=3255311. Cancer, edited by Prabhat Jha and Rangaswamy Sankarana- 138. Kruger, Herculina S., and Naomi S. Levitt. 2017. “Fetal Origins of rayanan. Washington, D.C.: World Bank. Obesity, Cardiovascular Disease, and Type 2 Diabetes.” In The 127. Jin, Lei, Lin Tao, and Xiangqian Lao. 2022. “Diverging Trends and Biology of the First 1,000 Days, edited by Crystal D. Karako- Expanding Educational Gaps in Smoking in China.” Interna- chuk, Kyly C. Whitfield, Tim J. Green, and Klaus Kraemer. Boca tional Journal of Environmental Research and Public Health Raton, FL: CRC Press. https://doi.org/10.1201/9781315152950. 19 (8): 4917. https://doi.org/10.3390/ijerph19084917. 139. Lane, Chris, Amanda Glassman, and Eleni Smitham. 2021. “Us- 128. Kachan, Diana, Lora E. Fleming, Sharon Christ, Peter Muennig, ing Health Taxes to Support Revenue: An Action Agenda for the Guillermo Prado, Stacey L. Tannenbaum, Xuan Yang, Alberto J. IMF and World Bank.” Development Policy Paper 203. Washing- Caban-Martinez, and David J. Lee. 2015. “Health Status of Older ton, D.C.: Centre for Global Development (CGD). https://www. US Workers and Nonworkers, National Health Interview Sur- cgdev.org/sites/default/files/Lane-Health-Tax-COVID-19.pdf. vey, 1997–2011.” Preventing Chronic Disease 12 (September). 140. Lane, Christopher. 2022. “Meeting Health Challenges in Devel- https://doi.org/10.5888/pcd12.150040. oping Asia with Corrective Taxes on Alcohol, Tobacco, and Un- 129. Kankeu, Hyacinthe Tchewonpi, Priyanka Saksena, Ke Xu, and healthy Foods.” Manila, Philippines: Asian Development Bank. David B Evans. 2013. “The Financial Burden from Non-Com- https://www.adb.org/sites/default/files/institutional-doc- municable Diseases in Low- and Middle-Income Countries: A ument/782851/ado2022bp-health-challenges-asia-correc- Literature Review.” Health Research Policy and Systems 11 (1). tive-taxes.pdf. 74 CHAPTER 2 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E 141. Langer, Ana, Afaf Meleis, Felicia M Knaul, Rifat Atun, Meltem Training.” Working Papers 2021-10. Department of Economics, Aran, Héctor Arreola-Ornelas, Zulfiqar A Bhutta, et al. 2015. University of Connecticut. https://ideas.repec.org/p/uct/ucon- “Women and Health: The Key for Sustainable Development.” np/2021-10.html. The Lancet 386 (9999): 1165–1210. https://doi.org/10.1016/ 154. Malta, Deborah Carvalho, Regina Tomie Ivata Bernal, Maria de s0140-6736(15)60497-4. Fatima Marinho de Souza, Celia Landman Szwarcwald, Mar- 142. Laureys, Lien. 2021. “The Cost of Human Capital Deprecia- gareth Guimarães Lima, and Marilisa Berti de Azevedo Barros. tion during Unemployment.” The Economic Journal 131 2016. “Social Inequalities in the Prevalence of Self-Reported (634): 827–50. https://ideas.repec.org/a/oup/econjl/v131y- Chronic Non-Communicable Diseases in Brazil: National Health 2021i634p827-850..html. Survey 2013.” International Journal for Equity in Health 15 (1). 143. Layard, Richard, Andrew E. Clark, Francesca Cornaglia, Nattavudh https://doi.org/10.1186/s12939-016-0427-4. Powdthavee, and James Vernoit. 2014. “What Predicts a Success- 155. Manley, James, Yarlini Balarajan, Shahira Malm, Luke Harman, ful Life? A Life-Course Model of Well-Being.” The Economic Jour- Jessica Owens, Sheila Murthy, David Stewart, Natalia Elena nal 124 (580): F720–38. https://doi.org/10.1111/ecoj.12170. Winder-Rossi, and Atif Khurshid. 2020. “Cash Transfers and 144. Lee, John Tayu, Fozia Hamid, Sanghamitra Pati, Rifat Atun, Child Nutritional Outcomes: A Systematic Review and Me- and Christopher Millett. 2015. “Impact of Noncommunicable ta-Analysis.” BMJ Global Health 5 (12): e003621. https://doi. Disease Multimorbidity on Healthcare Utilisation and Out-Of- org/10.1136/bmjgh-2020-003621. Pocket Expenditures in Middle-Income Countries: Cross Sec- 156. Martín-María, Natalia, Marta Miret, Francisco Félix Caballero, tional Analysis.” Edited by Ignacio Correa-Velez. PLOS ONE 10 Laura Alejandra Rico-Uribe, Andrew Steptoe, Somnath Chat- (7): e0127199. https://doi.org/10.1371/journal.pone.0127199. terji, and José Luis Ayuso-Mateos. 2017. “The Impact of Sub- 145. Lee, Ronald, Andrew Mason, and Timothy Miller. 2000. “Life jective Well-Being on Mortality.” Psychosomatic Medicine 79 Cycle Saving and the Demographic Transition: The Case of Tai- (5): 565–75. https://doi.org/10.1097/psy.0000000000000444. wan.” Population and Development Review 26 (Supplemen- 157. Mattke, Soeren, Aruna Balakrishnan, Giacomo Bergamo, and tal): 194–219. https://www.jstor.org/stable/3115217. Sydne J. Newberry. 2007. “A Review of Methods to Measure 146. Li, Hongbin, Jie Zhang, and Junsen Zhang. 2007. “Effects of Health-Related Productivity Loss.” The American Journal of Longevity and Dependency Rates on Saving and Growth: Managed Care 13 (4): 211–17. https://pubmed.ncbi.nlm.nih. Evidence from a Panel of Cross Countries.” Journal of Devel- gov/17408341/. opment Economics 84 (1): 138–54. https://doi.org/10.1016/j. 158. McIntyre, Diane, Michael Thiede, Göran Dahlgren, and Mar- jdeveco.2006.10.002. garet Whitehead. 2006. “What Are the Economic Conse- 147. Li, Min, Xiao-Wei Zhang, Wen-Shang Hou, and Zhen-Yu Tang. quences for Households of Illness and of Paying for Health 2015. “Impact of Depression on Incident Stroke: A Meta-Analy- Care in Low- and Middle-Income Country Contexts?” Social sis.” International Journal of Cardiology 180 (February): 103– Science & Medicine 62 (4): 858–65. https://doi.org/10.1016/j. 10. https://doi.org/10.1016/j.ijcard.2014.11.198. socscimed.2005.07.001. 148. Liakos, Matthew, and Puja B. Parikh. 2018. “Gender Disparities 159. McKee-Ryan, Frances, Zhaoli Song, Connie R Wanberg, and in Presentation, Management, and Outcomes of Acute Myo- Angelo J Kinicki. 2005. “Psychological and Physical Well-Be- cardial Infarction.” Current Cardiology Reports 20 (8). https:// ing during Unemployment: A Meta-Analytic Study.” The doi.org/10.1007/s11886-018-1006-7. Journal of Applied Psychology 90 (1): 53–76. https://doi. 149. Lleras-Muney, Adriana. 2022. “Education and Income Gradients org/10.1037/0021-9010.90.1.53. in Longevity: The Role of Policy.” NBER Working Paper 29694. 160. Meijer, Anna, Henk Jan Conradi, Elisabeth H. Bos, Brett D. Cambridge, MA: National Bureau of Economic Research. Thombs, Joost P. van Melle, and Peter de Jonge. 2011. “Prog- https://doi.org/10.3386/w29694. nostic Association of Depression Following Myocardial Infarc- 150. Lorentzen, Peter, John McMillan, and Romain Wacziarg. 2008. tion with Mortality and Cardiovascular Events: A Meta-Analysis “Death and Development.” Journal of Economic Growth 13 of 25 Years of Research.” General Hospital Psychiatry 33 (3): (2): 81–124. https://doi.org/10.1007/s10887-008-9029-3. 203–16. https://doi.org/10.1016/j.genhosppsych.2011.02.007. 151. Low, Lee-Fay, Melvyn Yap, and Henry Brodaty. 2011. “A System- 161. Meng, Liping, Haiquan Xu, Ailing Liu, Joop van Raaij, Wanda atic Review of Different Models of Home and Community Care Bemelmans, Xiaoqi Hu, Qian Zhang, et al. 2013. “The Costs and Services for Older Persons.” BMC Health Services Research 11 Cost-Effectiveness of a School-Based Comprehensive Interven- (1). https://doi.org/10.1186/1472-6963-11-93. tion Study on Childhood Obesity in China.” Edited by Maarten 152. Lundberg, Shelly. 1985. “The Added Worker Effect.” Journal Postma. PLoS ONE 8 (10): e77971. https://doi.org/10.1371/ of Labor Economics 3 (1): 11–37. https://www.jstor.org/sta- journal.pone.0077971. ble/2535048. 162. Modini, Matthew, Sadhbh Joyce, Arnstein Mykletun, Helen 153. Ma, Xiao, Alejandro Nakab, and Daniela Vidart. 2022. “Human Christensen, Richard A Bryant, Philip B Mitchell, and Samuel Capital Investment and Development: The Role of On-The-Job B Harvey. 2016. “The Mental Health Benefits of Employment: CHAPTER 2 75 Results of a Systematic Meta-Review.” Australasian Psychiatry gramming, and Predisposition.” The Lancet Global Health 7 24 (4): 331–36. https://doi.org/10.1177/1039856215618523. (4): e404–5. https://doi.org/10.1016/S2214-109X(19)30051-8. 163. Moreira da Silva, Jorge. 2019. “Why You Should Care about Un- 174. O’Keefe, Phillip B., Puja V. Dutta, and Harry Moroz. 2022. “Di- paid Care Work.” OECD Development Matters (blog). March verse Paths: The Dynamic Evolution of Social Protection in Asia 18, 2019. https://oecd-development-matters.org/2019/03/18/ and the Pacific.” Washington, D.C.: World Bank. why-you-should-care-about-unpaid-care-work.. 175. Oldereid, Nan B., Ulla-Britt Wennerholm, Anja Pinborg, Anne Loft, 164. Mortensen, Laust H, Johan Rehnberg, Espen Dahl, Finn Dider- Hannele Laivuori, Max Petzold, Liv Bente Romundstad, Viveca ichsen, Jon Ivar Elstad, Pekka Martikainen, David Rehkopf, Lasse Söderström-Anttila, and Christina Bergh. 2018. “The Effect of Pa- Tarkiainen, and Johan Fritzell. 2016. “Shape of the Association ternal Factors on Perinatal and Paediatric Outcomes: A Systemat- between Income and Mortality: A Cohort Study of Denmark, ic Review and Meta-Analysis.” Human Reproduction Update 24 Finland, Norway and Sweden in 1995 and 2003.” BMJ Open 6 (3): 320–89. https://doi.org/10.1093/humupd/dmy005. (12): e010974. https://doi.org/10.1136/bmjopen-2015-010974. 176. Onarheim, Kristine Husøy, Johanne Helene Iversen, and David E. 165. Mushkin, Selma J. 1962. “Health as an Investment.” Journal of Bloom. 2016. “Economic Benefits of Investing in Women’s Health: Political Economy 70 (5): 129–57. https://www.jstor.org/sta- A Systematic Review.” Edited by Jodi Pawluski. PLOS ONE 11 (3): ble/1829109. e0150120. https://doi.org/10.1371/journal.pone.0150120. 166. Naoum, Panagiota, Elpida Pavi, and Kostas Athanasakis. 2021. 177. Oreopoulos, Philip, and Kjell G Salvanes. 2011. “Priceless: The “Economic Evaluation of Digital Health Interventions in Palli- Nonpecuniary Benefits of Schooling.” Journal of Economic Per- ative Care: A Systematic Review of the Literature.” Frontiers spectives 25 (1): 159–84. https://doi.org/10.1257/jep.25.1.159. in Digital Health 3 (November). https://doi.org/10.3389/ 178. Organisation for Economic Cooperation and Development fdgth.2021.730755. (OECD). 2017. “Preventing Ageing Unequally.” Paris: OECD Pub- 167. Nikoloski, Zlatko, Ada Mohammed Alqunaibet, Rasha Ab- lishing. https://www.oecd.org/social/preventing-ageing-un- dulrahman Alfawaz, Sami Saeed Almudarra, Christopher H. equally-9789264279087-en.htm. Herbst, Sameh El-Saharty, Reem Alsukait, and Abdullah Alg- 179. ———. 2019. “OECD Future of Education 2030: Making Phys- wizani. 2021. “Covid-19 and Non-Communicable Diseases: Evi- ical Education Dynamic and Inclusive for 2030 - International dence from a Systematic Literature Review.” BMC Public Health Curriculum Analysis.” Paris: OECD Publishing. https://www. 21 (1). https://doi.org/10.1186/s12889-021-11116-w. oecd.org/education/2030-project/contact/OECD_FUTURE_ 168. Nikoloski, Zlatko, Luc Christiaensen, and Ruth Hill. 2018. OF_EDUCATION_2030_MAKING_PHYSICAL_DYNAMIC_AND_ “Household Shocks and Coping Mechanism: Evidence from INCLUSIVE_FOR_2030.pdf. Sub-Saharan Africa.” In Agriculture in Africa: Telling Myths 180. ———. 2020. Who Cares? Attracting and Retaining Elderly from Facts. Directions in Development—Agriculture and Care Workers. OECD Health Policy Studies. Paris: OECD Pub- Rural Development, edited by Luc Christiaensen and Li- lishing. https://doi.org/10.1787/92c0ef68-en. onel Demery. Washington, D.C.: World Bank. https://doi. 181. Patel, Vikram, and Somnath Chatterji. 2015. “Integrating Mental org/10.1596/978-1-4648-1134-0. Health in Care for Noncommunicable Diseases: An Imperative 169. Noordt, Maaike van der, Helma IJzelenberg, Mariël Droom- for Person-Centered Care.” Health Affairs 34 (9): 1498–1505. ers, and Karin I Proper. 2014. “Health Effects of Employment: https://doi.org/10.1377/hlthaff.2015.0791. A Systematic Review of Prospective Studies.” Occupational 182. Patterson, Christopher C., Suvi Karuranga, Paraskevi Salpea, Pouya and Environmental Medicine 71 (10): 730–36. https://doi. Saeedi, Gisela Dahlquist, Gyula Soltesz, and Graham D. Ogle. 2019. org/10.1136/oemed-2013-101891. “IDF Diabetes Atlas: Worldwide Estimates of Incidence, Prevalence 170. Norström, Fredrik, Pekka Virtanen, Anne Hammarström, Per E and Mortality of Type 1 Diabetes in Children and Adolescents: Re- Gustafsson, and Urban Janlert. 2014. “How Does Unemploy- sults from the International Diabetes Federation Diabetes Atlas, ment Affect Self-Assessed Health? A Systematic Review Focus- 9th Edition.” Diabetes Research and Clinical Practice 157 (Sep- ing on Subgroup Effects.” BMC Public Health 14 (1). https:// tember): 107842. https://doi.org/10.1016/j.diabres.2019.107842. doi.org/10.1186/1471-2458-14-1310. 183. Paul, Karsten I., and Klaus Moser. 2009. “Unemployment Impairs 171. Nussbaum, Martha C. 2013. Creating Capabilities. Harvard Mental Health: Meta-Analyses.” Journal of Vocational Behav- University Press. Cambridge, MA: Harvard University Press. ior 74 (3): 264–82. https://doi.org/10.1016/j.jvb.2009.01.001. https://www.hup.harvard.edu/books/9780674072350. 184. Petrosino, Anthony, Claire Morgan, Trevor Fronius, Emily E. 172. Nyaradi, Anett, Jianghong Li, Siobhan Hickling, Jonathan Fos- Tanner-Smith, and Robert F. Boruch. 2014. “What Works in ter, and Wendy H. Oddy. 2013. “The Role of Nutrition in Chil- Developing Nations to Get Children into School or Keep Them dren’s Neurocognitive Development, from Pregnancy through There?” Research on Social Work Practice 25 (1): 44–60. Childhood.” Frontiers in Human Neuroscience 7. https://doi. https://doi.org/10.1177/1049731514524837. org/10.3389/fnhum.2013.00097. 185. Pizzi, Michael A., and Kerryellen Vroman. 2013. “Childhood Obe- 173. Nyirenda, Moffat J., and Peter Byass. 2019. “Pregnancy, Pro- sity: Effects on Children’s Participation, Mental Health, and Psy- 76 CHAPTER 2 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E chosocial Development.” Occupational Therapy in Health Care bank.org/en/publication/documents-reports/documentde- 27 (2): 99–112. https://doi.org/10.3109/07380577.2013.784839. tail/250981606928190510/pdf. 186. Png, May Ee, Joanne Yoong, Thao Phuong Phan, and Hwee Lin 198. Salmuth, Victoria von, Eilise Brennan, Marko Kerac, Marie Mc- Wee. 2016. “Current and Future Economic Burden of Diabetes Grath, Severine Frison, and Natasha Lelijveld. 2021. “Maternal-Fo- among Working-Age Adults in Asia: Conservative Estimates for cused Interventions to Improve Infant Growth and Nutritional Singapore from 2010-2050.” BMC Public Health 16 (1). https:// Status in Low-Middle Income Countries: A Systematic Review doi.org/10.1186/s12889-016-2827-1. of Reviews.” Edited by Mary Hamer Hodges. PLOS ONE 16 (8): 187. Pol, Marjon van der. 2010. “Health, Education and Time e0256188. https://doi.org/10.1371/journal.pone.0256188. Preference.” Health Economics 20 (8): 917–29. https://doi. 199. Satin, Jillian R., Wolfgang Linden, and Melanie J. Phillips. 2009. org/10.1002/hec.1655. “Depression as a Predictor of Disease Progression and Mortal- 188. Popkin, Barry M, Camila Corvalan, and Laurence M Grum- ity in Cancer Patients.” Cancer 115 (22): 5349–61. https://doi. mer-Strawn. 2020. “Dynamics of the Double Burden of org/10.1002/cncr.24561. Malnutrition and the Changing Nutrition Reality.” The Lan- 200. Savelyev, Peter A. 2014. “Psychological Skills, Education, and cet 395 (10217): 65–74. https://doi.org/10.1016/s0140- Longevity of High-Ability Individuals.” Department of Econom- 6736(19)32497-3. ics Working Papers 14-00007. Vanderbilt University. https:// 189. Powdthavee, Nattavudh, Warn N. Lekfuangfu, and Mark ideas.repec.org/p/van/wpaper/vuecon-14-00007.html. Wooden. 2015. “What’s the Good of Education on Our Overall 201. Schuring, Merel, Johan Mackenbach, Toon Voorham, and Alex Quality of Life? A Simultaneous Equation Model of Education Burdorf. 2010. “The Effect of Re-Employment on Perceived and Life Satisfaction for Australia.” Journal of Behavioral and Health.” Journal of Epidemiology & Community Health 65 (7): Experimental Economics 54 (February): 10–21. https://doi. 639–44. https://doi.org/10.1136/jech.2009.103838. org/10.1016/j.socec.2014.11.002. 202. Sen, Amartya. 1997. “Editorial: Human Capital and Human Ca- 190. Pozzan, Emanuela, and Umberto Cattaneo. 2020. “Wom- pability.” World Development 25 (12): 1959–61. https://www. en Health Workers: Working Relentlessly in Hospitals and at staff.ncl.ac.uk/david.harvey/AEF806/Sen1997.pdf. Home.” ILO Newsroom, April 7, 2020. https://www.ilo.org/ 203. Seuring, Till, Pieter Serneels, and Marc Suhrcke. 2019. “The global/about-the-ilo/newsroom/news/WCMS_741060/lang-- Impact of Diabetes on Labour Market Outcomes in Mex- en/index.htm. ico: A Panel Data and Biomarker Analysis.” Social Science 191. Proimos, Jenny, and Jonathan D. Klein. 2012. “Noncommunica- & Medicine 233 (July): 252–61. https://doi.org/10.1016/j. ble Diseases in Children and Adolescents.” Pediatrics 130 (3): socscimed.2019.05.051. 379–81. https://doi.org/10.1542/peds.2012-1475. 204. Sewdas, Ranu, Astrid de Wind, Sari Stenholm, Pieter Coenen, 192. Propper, Carol, John Rigg, and Simon Burgess. 2007. “Child Ilse Louwerse, Cécile Boot, and Allard van der Beek. 2020. “As- Health: Evidence on the Roles of Family Income and Maternal sociation between Retirement and Mortality: Working Longer, Mental Health from a UK Birth Cohort.” Health Economics 16 Living Longer? A Systematic Review and Meta-Analysis.” Jour- (11): 1245–69. https://doi.org/10.1002/hec.1221. nal of Epidemiology and Community Health 74 (5): 473–80. 193. Reis, Mauricio C. 2007. “Added Worker Effect: Evidence from https://doi.org/10.1136/jech-2019-213023. Health Shocks in the Brazilian Informal Labor Market.” Brasil: 205. Shekar, Meera, Jakub Kakietek, Julia Dayton Eberwein, and IPEA. https://www.econ.puc-rio.br/uploads/adm/trabalhos/ Dylan Walters. 2017. “An Investment Framework for Nutrition: files/seminario/2007/Mauricio.pdf. Reaching the Global Targets for Stunting, Anemia, Breastfeed- 194. Roberts, Helen C., Stephen E. R. Lim, Natalie J. Cox, and Kinda ing, and Wasting.” Directions in Development. Washington, Ibrahim. 2019. “The Challenge of Managing Undernutrition in D.C.: World Bank. https://doi.org/10.1596/978-1-4648-1010-7. Older People with Frailty.” Nutrients 11 (4): 808. https://doi. 206. Shrimpton, Roger, and Claudia Rokx. 2012. “The Double Burden org/10.3390/nu11040808. of Malnutrition: A Review of Global Evidence.” Health, Nutri- 195. Rohwedder, Susann, and Robert J Willis. 2010. “Mental Re- tion, and Population (HNP) Discussion Paper. Washington, tirement.” Journal of Economic Perspectives 24 (1): 119–38. D.C.: World Bank. http://hdl.handle.net/10986/27417. https://doi.org/10.1257/jep.24.1.119. 207. Si, Chengyu, Denis Nadolnyak, and Valentina Hartarska. 2021. 196. Romer, Paul M. 1990. “Endogenous Technological Change.” “The Gender Wage Gap in Developing Countries.” Applied Eco- Journal of Political Economy 98 (5). https://www.jstor.org/ nomics and Finance 8 (1): 1–12. https://ideas.repec.org/a/rfa/ stable/2937632. aefjnl/v8y2021i1p1-12.html. 197. Saavedra, Jaine C., Mario Cristian Aedo Inostroza, Omar S. Arias 208. Smith, James P. 1999. “Healthy Bodies and Thick Wallets: The Diaz, Adelle Pushparatnam, Marcela Gutierrez Bernal, and F. Dual Relation between Health and Economic Status.” Jour- Halsey Rogers. 2022. “Realizing the Future of Learning : From nal of Economic Perspectives 13 (2): 145–66. https://doi. Learning Poverty to Learning for Everyone, Everywhere.” Wash- org/10.1257/jep.13.2.145. ington, D.C.: World Bank Group. https://documents.world- 209. Smith, James P. 2004. “Unraveling the SES: Health Connection.” CHAPTER 2 77 Population and Development Review 30: 108–32. https:// don: Oxford Health Alliance. https://www.uv.es/~atortosa/ www.jstor.org/stable/3401465. chronicdis-ecoperspective. 210. Snilstveit, Birte, Jennifer Stevenson, Radhika Menon, Daniel R 221. Suhrcke, Marc, and Dieter Urban. 2010. “Are Cardiovascular Dis- Phillips, Emma Gallagher, Maisie Geleen, Hannah Jobse, Tanja eases Bad for Economic Growth?” Health Economics 19 (12): Schmidt, and Emmanuel Jimenez. 2016. “The Impact of Educa- 1478–96. https://doi.org/10.1002/hec.1565. tion Programmes on Learning and School Participation in Low- 222. Sum, Grace, Chris Salisbury, Gerald Choon-Huat Koh, Rifat Atun, and Middle-Income Countries.” Systematic Review Summary Brian Oldenburg, Barbara McPake, Sukumar Vellakkal, and John 7. International Initiative for Impact Evaluation (3ie). https:// Tayu Lee. 2019. “Implications of Multimorbidity Patterns on doi.org/10.23846/srs007. Health Care Utilisation and Quality of Life in Middle-Income 211. Sommer, Isolde, Ursula Griebler, Peter Mahlknecht, Kylie Thaler, Countries: Cross-Sectional Analysis.” Journal of Global Health Kathryn Bouskill, Gerald Gartlehner, and Shanti Mendis. 2015. 9 (2). https://doi.org/10.7189/jogh.09.020413. “Socioeconomic Inequalities in Non-Communicable Diseases 223. Summan, Amit, Nicholas Stacey, Johanna Birckmayer, Evan and Their Risk Factors: An Overview of Systematic Reviews.” Blecher, Frank J Chaloupka, and Ramanan Laxminarayan. 2020. BMC Public Health 15 (1). https://doi.org/10.1186/s12889- “The Potential Global Gains in Health and Revenue from In- 015-2227-y. creased Taxation of Tobacco, Alcohol and Sugar-Sweetened 212. Stacey, Nicholas, Ijeoma Edoka, Karen Hofman, Elizabeth C. Beverages: A Modelling Analysis.” BMJ Global Health 5 (3): Swart, Barry Popkin, and Shu Wen Ng. 2021. “Changes in Bev- e002143. https://doi.org/10.1136/bmjgh-2019-002143. erage Purchases Following the Announcement and Imple- 224. Surkan, Pamela J, Caitlin E Kennedy, Kristen M Hurley, and Mau- mentation of South Africa’s Health Promotion Levy: An Obser- reen M Black. 2011. “Maternal Depression and Early Childhood vational Study.” The Lancet Planetary Health 5 (4): e200–208. Growth in Developing Countries: Systematic Review and Me- https://doi.org/10.1016/S2542-5196(20)30304-1. ta-Analysis.” Bulletin of the World Health Organization 89 (8): 213. Stephen, Ruth, Mariagnese Barbera, Ruth Peters, Nicole Ee, 608–15E. https://doi.org/10.2471/blt.11.088187. Lidan Zheng, Jenni Lehtisalo, Jenni Kulmala, et al. 2021. “De- 225. Sweeny, Kim, Bruce Rasmussen, and Peter Sheehan. 2015. “The velopment of the First WHO Guidelines for Risk Reduction of Impact of Health on Worker Attendance and Productivity in Cognitive Decline and Dementia: Lessons Learned and Future Twelve Countries.” Project Report. Victoria Institute of Strate- Directions.” Frontiers in Neurology 12 (October). https://doi. gic Economic Studies (VISES). Melbourne, Australia: Victoria org/10.3389/fneur.2021.763573. University. https://vuir.vu.edu.au/32605/. 214. Steptoe, Andrew, ed. 2006. Depression and Physical Illness. 226. Terada-Hagiwara, Akiko, Shiela Camingue-Romance, and Jo- Cambridge University Press. seph Zveglich Jr. 2018. “Gender Pay Gap: A Macro Perspective.” 215. ———. 2019. “Happiness and Health.” Annual Review of ADB Economics Working Paper Series No. 538. Manilla, Phil- Public Health 40 (1): 339–59. https://doi.org/10.1146/an- ippines: Asian Development Bank. http://dx.doi.org/10.22617/ nurev-publhealth-040218-044150. WPS189255-2. 216. Steptoe, Andrew, Angus Deaton, and Arthur A Stone. 227. The Task Force on Fiscal Policy for Health. 2019. “Health Taxes 2015. “Subjective Wellbeing, Health, and Ageing.” The Lan- to Save Lives: Employing Effective Excise Taxes on Tobacco, cet 385 (9968): 640–48. https://doi.org/10.1016/s0140- Alcohol, and Sugary Beverages.” New York, NY: Bloomberg 6736(13)61489-0. Philanthropies. https://www.issup.net/files/2019-04/Health- 217. Steptoe, Andrew, and Jane Wardle. 2012. “Enjoying Life and Liv- Taxes-to-Save-Lives-Report.pdf. ing Longer.” Archives of Internal Medicine 172 (3): 273. https:// 228. Tilvis, Reijo S., Venla Laitala, Pirkko Routasalo, Timo E. Strand- doi.org/10.1001/archinternmed.2011.1028. berg, and Kaisu H. Pitkala. 2012. “Positive Life Orientation Pre- 218. Stuck, Andreas E, Albert L Siu, G. Darryl Wieland, Laurence Z dicts Good Survival Prognosis in Old Age.” Archives of Geron- Rubenstein, and John L. Adams. 1993. “Comprehensive Geri- tology and Geriatrics 55 (1): 133–37. https://doi.org/10.1016/j. atric Assessment: A Meta-Analysis of Controlled Trials.” The archger.2011.06.030. Lancet 342 (8878): 1032–36. https://doi.org/10.1016/0140- 229. Verguet, Stéphane, Cindy L Gauvreau, Sujata Mishra, Mary 6736(93)92884-v. MacLennan, Shane M Murphy, Elizabeth D Brouwer, Rachel 219. Sudharsanan, Nikkil, Yuan Zhang, Collin F. Payne, William Dow, A Nugent, Kun Zhao, Prabhat Jha, and Dean T Jamison. 2015. and Eileen Crimmins. 2020. “Education and Adult Mortality in “The Consequences of Tobacco Tax on Household Health and Middle-Income Countries: Surprising Gradients in Six Nation- Finances in Rich and Poor Smokers in China: An Extended ally-Representative Longitudinal Surveys.” SSM - Population Cost-Effectiveness Analysis.” The Lancet Global Health 3 (4): Health 12 (December): 100649. https://doi.org/10.1016/j.ss- e206–16. https://doi.org/10.1016/s2214-109x(15)70095-1. mph.2020.100649. 230. Verguet, Stéphane, Paulina Limasalle, Averi Chakrabarti, Arif 220. Suhrcke, Marc, Rachel A. Nugent, David Stuckler, and Lorenzo Husain, Carmen Burbano, Lesley Drake, and Donald A. P. Bun- Rocco. 2006. “Chronic Disease: An Economic Perspective.” Lon- dy. 2020. “The Broader Economic Value of School Feeding Pro- 78 CHAPTER 2 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E grams in Low- and Middle-Income Countries: Estimating the Mellitus and Epilepsy.” Psychology in the Schools 45 (1): 52– Multi-Sectoral Returns to Public Health, Human Capital, Social 62. https://doi.org/10.1002/pits.20278. Protection, and the Local Economy.” Frontiers in Public Health 241. World Bank. 2015. Global Monitoring Report 2015/2016. 8: 587046. https://doi.org/10.3389/fpubh.2020.587046. Washington, D.C.: World Bank. https://doi.org/10.1596/978- 231. Victora, Cesar G, Linda Adair, Caroline Fall, Pedro C Hallal, Rey- 1-4648-0669-8. naldo Martorell, Linda Richter, and Harshpal Singh Sachdev. 242. ———. 2018. “The State of Social Safety Nets 2018.” Washing- 2008. “Maternal and Child Undernutrition: Consequences for ton, D.C.: World Bank. http://hdl.handle.net/10986/29115. Adult Health and Human Capital.” The Lancet 371 (9609): 340– 243. ———. 2021. “The Changing Wealth of Nations 2021: Manag- 57. https://doi.org/10.1016/s0140-6736(07)61692-4. ing Assets for the Future.” Washington, D.C.: World Bank. http:// 232. Vila, Luis E. 2000. “The Non-Monetary Benefits of Education.” hdl.handle.net/10986/36400. European Journal of Education 35 (1): 21–32. https://doi. 244. World Health Organization (WHO). 2012. “Programme on org/10.1111/1467-3435.00003. Mental Health: WHOQOL User Manual, 2012 Revision.” Geneva, 233. Wagstaff, Adam, Gabriela Flores, Justine Hsu, Marc-François Switzerland: World Health Organization. https://www.who.int/ Smitz, Kateryna Chepynoga, Leander R. Buisman, Kim van publications/i/item/WHO-HIS-HSI-Rev.2012-3. Wilgenburg, and Patrick Eozenou. 2018. “Progress on Cata- 245. ———. 2016. “Report of the Commission on End- strophic Health Spending in 133 Countries: A Retrospective ing Childhood Obesity.” Geneva: Switzerland: World Observational Study.” The Lancet Global Health 6 (2): e169– Health Organization. https://iris.who.int/bitstream/han- 79. https://doi.org/10.1016/S2214-109X(17)30429-1. dle/10665/204176/9789241510066_eng.pdf?sequence=1. 234. Wagstaff, Adam, and Sven Neelsen. 2020. “A Comprehensive As- 246. ———. 2017. “Tackling NCDs: ‘Best Buys’ and Other Recom- sessment of Universal Health Coverage in 111 Countries: A Ret- mended Interventions for the Prevention and Control of Non- rospective Observational Study.” The Lancet Global Health 8 communicable Diseases.” Geneva, Switzerland: World Health (1): e39–49. https://doi.org/10.1016/S2214-109X(19)30463-2. Organization. https://www.who.int/publications/i/item/WHO- 235. Watkins, David A., Dean T. Jamison, Anne Mills, Rifat Atun, Kris- NMH-NVI-17.9. ten Danforth, Amanda Glassman, Susan Horton, et al. 2017. 247. ———. 2018. “Global Status Report on Road Safety 2018.” “Universal Health Coverage and Essential Packages of Care.” Geneva, Switzerland: World Health Organization. https://www. Edited by Dean T. Jamison, Hellen Gelband, Susan Horton, Pra- who.int/publications/i/item/9789241565684. bhat Jha, Ramanan Laxminarayan, Charles N. Mock, and Rachel 248. ———. 2019. “‘GHE: Life Expectancy and Healthy Life Ex- Nugent. 3rd ed. Washington, D.C.: World Bank. https://www. pectancy.’” Global Health Observatory. World Health Organi- ncbi.nlm.nih.gov/books/NBK525285/. zation. 2019. https://www.who.int/data/gho/data/themes/ 236. Watkins, David, Sali Ahmed, and Sarah Pickersgill. 2024. “Prior- mortality-and-global-health-estimates/ghe-life-expectan- ity Setting for NCD Control and Health System Investments.” cy-and-healthy-life-expectancy. In Unlocking the Power of Healthy Longevity: Compendium 249. ———. 2020. “Palliative Care.” Fact Sheet. World Health Orga- of Research for the Healthy Longevity Initiative. Washington, nization. https://www.who.int/news-room/fact-sheets/detail/ D.C.: World Bank. palliative-care. 237. Webb, Roger, Kathryn Abel, Andrew Pickles, and Louis Appleby. 250. ———. 2021. “Obesity and Overweight.” Fact Sheet. World 2005. “Mortality in Offspring of Parents with Psychotic Disor- Health Organization. https://www.who.int/news-room/fact- ders: A Critical Review and Meta-Analysis.” American Journal sheets/detail/obesity-and-overweight. of Psychiatry 162 (6): 1045–56. https://doi.org/10.1176/appi. 251. ———. 2022. “World Health Statistics 2022.” World Health ajp.162.6.1045. Organization. May 20, 2022. https://www.who.int/news/ 238. Williams, Julianne, Luke Allen, Kremlin Wickramasinghe, Bente item/20-05-2022-world-health-statistics-2022. Mikkelsen, Nia Roberts, and Nick Townsend. 2018. “A System- 252. Wu, Chenkai, Michelle C Odden, Gwenith G Fisher, and Robert atic Review of Associations between Non-Communicable Dis- S Stawski. 2016. “Association of Retirement Age with Mortality: eases and Socioeconomic Status within Low- and Lower-Mid- A Population-Based Longitudinal Study among Older Adults in dle-Income Countries.” Journal of Global Health 8 (2): 020409. the USA.” Journal of Epidemiology and Community Health https://doi.org/10.7189/jogh.08.020409. 70 (9): 917–23. https://doi.org/10.1136/jech-2015-207097. 239. Windle, Gill, Dyfrig Hughes, Pat Linck, Ian Russell, and Bob 253. Yiengprugsawan, Vasoontara S., and John Piggott. 2022. Shap- Woods. 2010. “Is Exercise Effective in Promoting Mental Well-Be- ing Long-Term Care in Emerging Asia. Taylor & Francis. ing in Older Age? A Systematic Review.” Aging & Mental Health 254. Yiengprugsawan, Vasoontara Sbirakos, Riana Rahmawati, 14 (6): 652–69. https://doi.org/10.1080/13607861003713232. Robert G. Cumming, and John Piggott. 2022. “Factors Relating 240. Wodrich, David L., and Melissa M. Cunningham. 2007. “School- to Depressive Symptoms of Caregivers for Older Care Recipi- Based Tertiary and Targeted Interventions for Students with ents in Indonesia.” Aging & Mental Health 26 (12): 2454–61. Chronic Medical Conditions: Examples from Type 1 Diabetes https://doi.org/10.1080/13607863.2021.1980858. CHAPTER 2 79 255. You, Emily C., David Dunt, Colleen Doyle, and Arthur Hsueh. 2012. 256. Yusuf, Salim, Sumathy Rangarajan, Koon Teo, Shofiqul Islam, “Effects of Case Management in Community Aged Care on Client Wei Li, Lisheng Liu, Jian Bo, et al. 2014. “Cardiovascular Risk and and Carer Outcomes: A Systematic Review of Randomized Trials Events in 17 Low-, Middle-, and High-Income Countries.” New and Comparative Observational Studies.” BMC Health Services England Journal of Medicine 371 (9): 818–27. https://doi. Research 12 (1). https://doi.org/10.1186/1472-6963-12-395. org/10.1056/nejmoa1311890. ANNEX 2.1 TABLE 2A.1  Mapping the human capital trajectory and its interactions with ncds across the life cycle Life stage Human Capital (HC) Trajectory stage/elements NCD Incidence and factors Impacts of NCD on HC (and feedback) Gestation/birth HC “at entry” • IUGR as risk factor for NCD in • IUGR impacts both child vulnerability to adulthood, incl. CVD, renal, CDs and adult NCDs, with negative HC lung, and diabetes. Maternal, acquisition and deployment implications placental or fetal origins (10- • Low birthweight increases risks of 15% of pregnancies) later life NCDs, incl. diabetes, obesity, • Health/nutrition at birth incl. hypertension, CVD and neurological birthweight, and at-birth disorders. And children or low BW disability. Mainly inter- mothers more likely to have low BW generational drivers, derivative children of maternal health/nutrition, • Compromised foundation for HC risk factors/behaviour, and development genetics Early years 0-14 HC intensive formation and building stage • Childhood NCD onset and Short to medium run impacts: • Early cognitive base, learning and NCD-related mortality. incl. • Short-circuit of future HC through foundational skills built asthma, obesity, congenital childhood deaths • Childhood health and formation of heart, mental/behavioural • Decreased education attendance and health behaviours health attainment • Childhood nutritional status as key HC • Derivative impacts of parental • Returns to HK compromised by lower life course determinant NCDs on children/adolescents attainment • Some risk factors and • Decreased educational attendance and behaviours for later NCDs start. attainment from parental NCD impacts Pollution, sugar and salt intake, on children diet Long-run impacts: • Nutritional deficits, esp. in early years, increase adult NCD risks • Horizon effect of lower societal life expectancy from NCDs on investments in education School to working HC trajectory diversifies, as HC accumu- • Intensive period of mental • Lower likelihood to progress in education age transition and lation continues, HC deployment stage health conditions onset due to lagged and current NCD impacts early career – 15-34 starts, and potential HC depreciation • Initiation of some NCD risk • Failure to enter labor force or withdrawal stage starts for some behaviours, e.g., smoking and from LF due to NCD mortality, disability or alcohol morbidity: multiple HC impact of no HC Tracks of accumulation: • Childhood NCD continuation stock deployment, depreciation of HC stock • Intensive formative stage + for those and exacerbation (e.g., obesity) (which had investment costs to build) and continuing education • Earlier onset of adulthood failure to accumulate further HC on-job. • Learning-by-doing HC accumulation NCDs, [with social gradient] • Compromised HC deployment/returns stage in work starts due to lower productivity within jobs from absenteeism and presenteeism Track of deployment: • HK depreciation due to work gaps/weak • Returns to stock+flow of HC initiated LM attachment • Reduced accumulation of OTJ HC and Track of depreciation: reduced incentives to deepen human • For some, disability or serious or capital recurrent NCD may prevent LM entry • Lower HC of parents/mothers starts to altogether or compromise participation, impact children resulting in active depreciation and • Correlation between education and failure to accumulate further HC on the own health outcomes, but casual link job disputed and may vary according to health variable and other factors 80 CHAPTER 2 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E Life stage Human Capital (HC) Trajectory stage/elements NCD Incidence and factors Impacts of NCD on HC (and feedback) Mid-career to end of Three track HC trajectory continues, but • Increased onset of NCDs, some • Withdrawal from LF due to NCD “standardized” work- relative balance between tracks may originating in childhood mortality, disability or morbidity: multiple ing life – 35-65/69 shift over this life stage: • Increase in co-morbidities HC impact of no HC stock deployment, • HC accumulation through learning by • NCDs of older parents increase depreciation of HC stock, and failure to doing continues, with maintenance or and rising ADL/IADL impacts accumulate further HC on-job increase of crystallized skills and tapering • Where acute, spillover effects on of fluid abilities. partners/families as carers and secondary • HC accumulation divergence in low and LM impacts, and income effects and SES high quality jobs association with lower HK accumulation • Continued deployment of HC/returns, for children with RoR to human capital diverging • Compromised HC deployment/returns over time for high/low skilled due to lower productivity within jobs • Employer biases against older workers from absenteeism and presenteeism compromise returns to HC from around • HK depreciation due to work gaps/weak 55 LM attachment • Normal depreciation of skills element • Reduced accumulation of OTJ HC and of HC may be more compromised with reduced incentives to deepen human time capital • Protection of human capital stock • Parental NCDs require LF withdrawal challenge increases or reduction (or transition to informal sector) for adult children, esp. women, with low/no RoR in HK deployment Later life: 65/70 till • LM participation divergence between • Co-morbidities increase sharply, • Impacts on HC of adult children, esp. death formal and informal workers and disabling ADL/IADL with women, more pronounced, both in LM/ • Decelerated and then terminated time economic terms and in terms of higher accumulation of HC • Enhanced risk factor of social NCD likelihood for carer also • Returns to HK decline and cease isolation and links to other risk • Impacts of grandparental HC decline on • Protection of stock of HK more crucial on factors and worse NCD status their care contribution and spillover to health and cognitive domains • Specific NCDs of age grandchild HC formation and increased onset (dementia/cognitive demand for parental child care deterioration/ higher depression rates?) TABLE 2A.2  Selected policies for human capital and ncds across the life course Life stage Policy and action areas Cross life course policies Gestation and Gestation and early years • Taxation policies to reduce NCD risks at early years 0-14 • Generalized nutritional interventions for adolescent girls and women of child- individual and societal levels, including bearing age pre-pregnancy public bads such as pollution, tobacco, • Nutritional interventions for pregnant and lactating women (PLW), including alcohol and sugars supplementation and SBCC for parental roles WRT nutrition • Ante-natal care, including growth monitoring of fetus • Accessible and affordable health care/ • Early years nutritional interventions for children 0-5, with priority focus on first UHC with enhanced NCD emphasis in 1,000 days, including zinc and Vitamin A supplementation basic package, including mental health • (Un)conditional cash transfers linked to PLW and early years cohort and (aligned with WHO best buys and DCP3) participation in nutrition interventions, including SBCC and parenting • Parental leave policies • Governments engaging with private • Early intervention for childhood NCD prevention and management including sector to control NCD risk factors, obesity and asthma promote healthy lifestyles, and • Pre-primary education interventions, including community-based approaches manage impacts, using sticks and • Free or subsidized childcare to facilitate parental LFP carrots (e.g., pharma, food, sports and fitness, insurance, banking, advertising, School years: entertainment, transport, and • Tuition free basic education infrastructure). Consider Public–private • Incentive policies to minimize school drop-out, including cash transfers, partnerships with appropriate statutory allowances for transport/uniforms, fee-free education etc. and regulatory frameworks + managing • School feeding programs conflicts of interest • Remedial programs to address lagging learning • Health education mainstreaming in school curricula and physical activity programs • Fiscal systems built on regular • Labor regulation to prohibit harmful child labor and accompanying enforcement demographic projections and stress mechanism testing CHAPTER 2 81 Life stage Policy and action areas Cross life course policies Youth, educa- • Free or subsidized post-secondary education, with targeted financial incentives • Deepening research on key aspects of tion-to-work tran- for low-income and vulnerable groups healthy longevity, NCDs, human capital sition and early • Adult/life-long learning programs to promote regular skills upgrading/refreshing and their interactions and economic, career, 15-35 • Tax/financial incentives to employers to invest in workforce training distributional and gender impacts, and • Incorporation of experiential learning in National Qualifications Frameworks evaluation of public interventions to • Social security programs for sickness, disability and work injury (non-contributory assess cost-effectiveness for informal sector) and financial subsidies to promote pension scheme participation by workers • Promotion of universal design through • Unemployment allowances and/or generalized social assistance and ALMP to planning and approval processes promote workforce re-entry and skills building to promote accessibility to the built environment, transport and other services for those with functional Mid-career to • Adult/life-long learning programs to promote regular skills upgrading/refreshing impairment end of “normal” across working life, adapted/targeted to meet particular needs and learning working age, strengths of late-career workers 36-69 • Tax/financial incentives to employers to invest in workforce training • Revised labor regulations to facilitate flexible work arrangements • Financial incentives to firms to retain or hire older workers • Public information/awareness campaigns to address older worker discrimination/ biases • Financial incentives to adapt workplaces to minimize work injury and maximize participation and productivity of people with health conditions and disabilities • Social security programs for sickness, disability and work injury (non-contributory for informal sector) and financial subsidies to promote pension scheme participation by workers • Unemployment allowances and/or generalized social assistance and ALMP to promote workforce re-entry and skills building • Reform of pension systems to incentivize longer working lives in formal sector and align male and female retirement ages • Option of carer allowances for care of older parents Young elderly • Adequate pensions/social pensions for elderly with wide coverage, and consider stage, 70-79 “soft” conditioning of pensions to periodic health checks • Learning and skill development interventions for those past working age, adapted for specific learning strengths/deficits of older people • Promotion of volunteer initiatives for older people to enhance community engagement and non-market contributions • Publicly subsidized aged care system with emphasis on home- and community- based care, and targeted support to most vulnerable populations (and consideration of financing model). Built on case management and age- integrated care centres • Option of carer allowances for care of older parents and enhanced care support for care-related NCDs • Community-based initiatives to address social isolation and promote social participation • Cross-sectoral mechanisms to promote coordination of care across healthcare and aged/social care systems, built around case management. • Dementia risk reduction and treatment programs • Further research and implementation of multi-mode interventions to delay or reverse cognitive decline in healthy older people • Incorporation of gerontology modules in GP, nurse and social work training • Public information/awareness campaigns to address ageism Older elderly • As for 65-75 stage, 80-death • Home and community-based interventions to address social isolation • Widespread and affordable availability of palliative care 82 CHAPTER 2 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E TABLE 2A.3  Human capital promoting policies during working life for accumulation, deployment and protection of HC Dimension of human capital Policy/intervention Further HC accumulation • Build second chance education pathways, both towards formal qualifications and directly to the labor market, including subsidies for poor/vulnerable participants • Targeted and timebound wage subsidy program to provide work experience and improve employability for youth, women and people with disabilities • Provide tax/financial incentives for employers to enhance in- and on-job training • Consider training levy to promote formal sector firm-level training • Ensure modularity of TVET, university and other formal adult learning offerings to allow workers to upgrade skills while employed • Enhance emphasis on digital skills development for older workers, and use of technology in LLL offerings and remote access • Incorporate adult/lifelong learning into National Qualifications Frameworks • Ensure stipends or other support to ensure low-income people can access adult learning opportunities • Promote accessible public-available information on career options, skills requirements, compensation and other features • Expand skill, entrepreneurship, and financial literacy programs for informal sector workers/groups, including through partnerships with non-governmental providers and community-based approaches • Encourage inter-generational work teams to maximize diverse skill exchange • Sustain funding for agricultural extension services and increase use of technology in their provision to promote reach HC deployment/utilization • Implement Recognition of Prior Learning (RPL) system to recognize and increase returns to skills and competencies acquired through non-formal channels • Expand state subsidies for parental leave and childcare to increase LFPR, especially for women • Promote personal tax regimes that incentivize, or at minimum do not disincentivize, labor force participation by couples • Expand state support to aged care to reduce mid-late career labor force withdrawal by those providing care • Pension system reforms to incentivize longer working lives and remove mandates or financial incentives to retire prematurely • Review taxation systems to ensure that tax force to retire is minimized • Phase out wage setting practices which reward age/tenure over worker productivity • Facilitate access to credit for firms and particularly for informal sector workers/enterprises through diverse channels, including banks, MFIs, group credit schemes, etc. • Ensure non-discrimination in hiring, through labor and/or general anti-discrimination laws and their enforcement • Consider employer incentives/subsidies for retention of older workers • Adaptation of workplaces to ensure sustained productive work HC depreciation • Strengthen employment services/ALMPs to reduce human capital depreciation during unemployment and accelerate LM re-entry • Provide unemployment insurance/assistance for laid-off workers • Provide financial incentives for workplace adaptations to allow for longer working lives with lower physical stress on workers • Provide incentives in pension systems for longer working lives to realized health benefits of longer working lives • Promote and incentivize volunteerism for benefit of volunteers and recipients of services HC protection • Enforcement of legislation outlawing harmful child labor • Expand social security coverage of HI, sickness/disability, and pension social insurance, with full/partial subsidization of premiums for informal sector workers • Strengthen workplace occupational health and safety compliance and provide incentives for workplace health promotion/wellness programs • Labor regulations on unfair dismissal and workplace discrimination, with grievance redress mechanisms • General social assistance in event of need/poverty COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 3 83 Economic Cost of Avoidable Mortality Chapter 3 The Economic Value of Avoidable Mortality Chapter 4 The Economic Value Associated with Avoidable Mortality: A Systematic Assessment by Cause of Death across World Regions Chapter 5 Rates of Progress in Mortality Decline, 2000–2019 3 84 CHAPTER 3 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E The Economic Value of Avoidable Mortality Angela Y. Chang a,b,c, Gretchen A. Stevens d, Diego S. Cardoso e, Bochen Cao f, Dean T. Jamison g a Danish Institute for Advanced Study, University of Southern Denmark b Department of Clinical Research, University of Southern Denmark c The Interdisciplinary Centre on Population Dynamics (CPop), University of Southern Denmark d Independent Researcher e Department of Agricultural Economics, Purdue University f Department of Data Analytics, World Health Organization g Department of Epidemiology and Biostatistics and Institute for Global Health Sciences, University of California, San Francisco COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 3 85 INTRODUCTION Disease burden is commonly measured using a variety of health statistics, including deaths by cause, injury and disease prevalence and incidence, and disability-adjusted life years. While these metrics are meaningful, it is also evident that preventing injuries and diseases has practical economic and social consequences beyond health which have economic value. As part of the World Bank’s Healthy Longevity Initiative, this chapter aims to estimate the eco- nomic value of avoidable mortality by world regions, sex, and age, between 2000 and 2019, with projection to 2050. This work lies at the conjunction of three important recent, comprehensive pictures of a now substan- strands of literature – estimation of the global burden tially empirical literature are now available (Schelling of disease (GBD); estimation of the proportion of that 1968; Hammitt and Herrera-Araujo 2019; Viscusi burden that is amenable to intervention or can other- 1993; Viscusi and Aldy 2003; Robinson et al. 2019; wise be avoided; and estimation of the economic value Hammitt 2020). Usher introduced methods for associated with reducing this burden (OECD 2019; assessing the contribution of mortality change to Rutstein 1976; Nolte & McKee 2004). Key to defin- growth in economic welfare and provided estimates ing the GBD was correcting for the overestimation of for selected countries (Usher 1973; Bloom et al. deaths by individual causes that resulted from 2004). The Lancet Commission on Investing in individual disease communities providing estimates Health provided updated estimates of the contri- for their own disease. In the 1980s, researchers at bution of mortality reduction to economic welfare the World Health Organization (WHO) initiated change in low- and middle-income countries (Jami- systematic correction by imposing demographically son 2013; Chang et al. 2017). defined death totals as an envelope within which The purpose of this chapter is to bring together disease-specific estimates must lie (Hakulinen et al. these two literatures with the economists’ valuation 1986). This work was extended to include a more of mortality change and the demographers and ep- detailed list of causes of death and estimates of idemiologists’ definition of an overall envelope of nonfatal health burden and initially published in the mortality within which the sum of deaths from indi- World Bank’s 1993 World Development Report, vidual diseases must lie. The result is an estimate of Investing in Health, Appendix B (World Bank 1993; the economic value of avoidable mortality that points to Murray et al. 1994). Earlier than the work on burden the monetized benefits from future improvements in of disease, economists had initiated efforts to place health. With our framework, one could answer value on small changes in mortality probabilities (albeit tentatively) questions such as: “For a typical and, quite separately, on estimating the value of female, age 45, from Colombia: What percentage of mortality reductions relative to increases in GDP in their income would they be willing to forgo to live the providing a comprehensive estimate of growth in next year under the lowest possible mortality risk?”. national eco-nomic welfare. Schelling provided early By systematically applying our framework to a approaches to valuing changes in mortality rates; variety of populations, we then characterize the eco- nomic value to various groups and regions. CONCEPTUAL FRAMEWORK Two fundamental concepts underlie our effort: the definition of avoidable mortality and the economic valuation of mortality risks. This section outlines the conceptual framework adopted in our methodology. Avoidable mortality Avoidable mortality comprises deaths that may be mortality is typically estimated as the number of prevented through public health or prevention inter- deaths assigned to set of specific causes of death, ventions that reduce incidence (preventable mortal- when they occur below a threshold age (OECD 2019; ity) and those that can be avoided through curative Rutstein 1998; Nolte and McKee 2004). These deaths health care interventions that reduce case-fatality are categorized as either preventable or treatable. (treatable or amenable mortality) (OECD 2019; This list-based approach has several drawbacks for Nolte and McKee 2004; Rutstein 1976). Avoidable our purposes: first, it requires comparable, detailed, 86 CHAPTER 3 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E and accurate cause-of-death data, which are not far smaller than the 5-year gap between male and currently available for most countries (WHO 2020). female frontier life expectancy in 2019 (Luy 2016). Second, deaths cannot be categorized as exclusive- ly preventable or amenable as some deaths may fit Economic value in both categories: a death from road traffic injury of mortality reduction could, in many cases, have been prevented (e.g., by traffic calming measures) or avoided by better care Reductions in mortality risks lead to improved (e.g., in the emergency room). We have addressed wellbeing and higher living standards in societies. these issues by implementing a simplified method As such, these reductions are perceived as a benefit which is suitable for countries with varying avail- to the populations that experience them. But con- ability and quality in cause-of-death information. sistent mortality reductions are often the result of Our method estimates avoidable mortality without costly public interventions. Therefore, governments disaggregation into preventable and treatable mor- seeking to make more objective decisions frequent- tality. Specifically, we estimate avoidable mortality ly perform quantitative comparisons of the benefits as the difference between current (estimated or pro- and costs of various interventions before establish- jected) mortality levels and unavoidable, or frontier ing their priorities. While estimating the cost of po- mortality levels, which are the lowest mortality lev- tential interventions is a relatively straightforward els that can be obtained for each age given past and accounting exercise, quantifying the benefits of current technologies and knowledge. mortality reductions requires a conceptual frame- Frontier mortality levels are estimated for each work of economic value. time t as the lowest projected or observed mortality Leading scholars have developed and applied rates at each age in either sex at time t. We apply a estimates of the value per statistical life (VSL) to single age-sex-year-specific frontier to all countries quantify the benefits of mortality reductions (Viscu- and both sexes. In other words, all countries are si 1992; Sunstein 2002; Revesz 2008; Sunstein 2014; compared to the single best-practice frontier, and Viscusi 2018; Robinson et al. 2019; Chang et al. 2018; both females and males are compared to the same Khadka and Verguet 2021). The VSL measures the frontier. This is built on our belief that all countries, willingness to pay (or to accept compensation) for a with the right level of financial investment, political small reduction (increase) in mortality risk. The fol- will, policies, and technological advancements, have lowing example illustrates the concept. Suppose we the opportunity to reach the frontier, even though know that a group of workers is willing to pay $1,000 it may be more challenging (and take longer) for on average to reduce their individual annual fatali- some countries than others. ty risk by 1/10,000. Then, if exactly 10,000 of those We compared both male and female mortal- workers reduce their fatality risk by this amount, they ity to the lowest sex-specific mortality rate, which statistically avoid one death among them. Therefore, in all cases is a female mortality rate. Most of the they collectively spend $10 million to save one sta- sex differences in life expectancy has been shown tistical life – or, equivalently they imply a VSL of $10 to be due to non-biological factors, namely gen- million. The VSL as a unit is, empirically, on the order der differences in health behavior and risks, such of 10,000 times the value actually measured. as smoking, alcohol use, and injuries (Luy 2016). The VSL approach is widely adopted by govern- In high-income settings, from which our frontier ments and regulatory agencies in countries that re- mortality rates are drawn, excess male mortality is quire an economic analysis of potential public sector caused by injuries and non-communicable diseas- investments and regulations, including the United es (WHO 2020). Excess injury deaths in males are States, Norway, United Kingdom, and Canada (Vis- likely a result of socio-cultural factors rather than cusi and Aldy 2003). Before the theoretical and em- innate biological susceptibility. A higher proportion pirical advances that led to the consolidation of the of male non-communicable disease deaths can be VSL framework, the value of saving lives was usual- attributed to modifiable risk factors, again pointing ly equated to savings in medical costs and forgone to the importance of socio-cultural differences caus- wages. However, this alternative approach based on ing higher male mortality (GBD 2019). Specialized economic losses had substantial limitations for its studies of populations where the gender differences use in benefit-cost analyses. For instance, because in health exposures are minimized have found that of the list-based quantification, the calculated value the female longevity advantage is around 1-2 years, is highly sensitive to omissions and often leads to COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 3 87 substantial underestimation. Moreover, measures of changes are not appreciably larger than values em- cost savings are inherently disconnected from the pirically measured. As indicated above, the VSL is idea of economic benefits, which are conceptually defined and estimated based on small risk changes. linked to willingness to pay in the economic theo- When applied to large risk changes, the VSL multi- ry. As the studies by Viscusi have indicated, policies plication approach can lead to substantial overesti- that incorporate the VSL are much more protective mation and even to values that exceeds one’s income of lives and health than policies that adopt the eco- – a direct contradiction with the concept of the in- nomic losses approach (Viscusi 2018). come-mortality trade-off that underlies the VSL. The VSL can be estimated based on two types of The VSL multiplication (or fixed VSL) ap- data: observed behavior (the revealed preference ap- proach can overestimate economic values because it proach) or surveys about subjects’ willingness to pay is a linear approximation which holds constant the (the stated preference approach). The empirical VSL economic value of risk change. Such approximation literature generally favors estimates based on ob- results in negligible errors when risk changes are served behavior because of the significant challeng- small and of the same order used for the empirical es in eliciting truthful and consistent answers about estimation of the VSL. However, as the mortality risk willingness to pay from hypothetical scenario.1 How- increases, each subsequent risk reduction plausibly ever, inference from observed behavior also poses has a smaller value – a principle commonly referred relevant limitations, such as biases from individuals’ to as the decreasing marginal value. Intuitively, this misperception of risks or selection into specific be- principle also holds because as individuals trade off haviors.2 Estimates following the revealed preference more and more of their resources for reduced mor- approach frequently infer the VSL based on market tality, they need to forgo other uses of their income data. For instance, these studies can statistically es- in increasing order of value. The linear extrapolation timate the VSL based on the labor markets, using in the fixed VSL approach fails to account for the the differences between wages in occupations with changing nature of this substitution rate. Therefore, different fatality risks, or safety equipment markets in cases where mortality risk changes can be large, (such as helmets and airbags), using the differences such as the case of eliminating avoidable mortality, in the demand for various devices with different de- it is necessary to adjust the value of risk reductions grees of risk protection. These inherent uncertainties to stay consistent with the underlying VSL frame- in empirical estimation imply that findings based on work (Hammitt and Herrera-Araujo 2019; Cardoso the literature should be viewed more as importantly and Dahis 2024; Bressler 2022). suggestive than conveying quantitative decision. In this chapter, we apply the more plausible ap- The VSL approach is increasingly applied in the proach that regards estimates of the value of mortal- global health literature (Robinson et al. 2019; Chang ity change as reasonable approximation for modest et al. 2018). These applications typically calculate the changes around the baseline mortality levels. Then, economic value of reducing mortality by multiplying we approximate the value of larger changes based the VSL by the number of expected lives saved (Jami- on assumed functions that respect income and son et al. 2013 Khadka and Verguet, 2021). However, resource constraints, thus never exceeding one’s this approach is appropriate only when mortality risk means in their valuation of mortality reduction. METHODS Estimating avoidable mortality We define avoidable mortality as the difference between current (estimated or projected) mortality levels and unavoidable, or frontier mortality levels, which are the lowest mortality levels that can be obtained for each age given current technologies and knowledge. Frontier mortality rates are selected from countries selecting incorrectly estimated mortality rates. with high-quality death registration data to avoid Specifically, we selected the frontier mortality rates 1   For further details, see discussions in Hammitt, James and Graham (1999); and Andersson and Treich (2011). 2   For an extensive discussion of these limitations, see Viscusi and Aldy (2003) 88 CHAPTER 3 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E from lifetables included in the Human Mortality from the 1980-1984 to 2015-2019. This is a com- Database (HMD), which are derived from popula- monly used approach for projecting mortality that tions where death registration and census data are results in a projected rate of increase of frontier life virtually complete (Barbieri et al. 2015; Wilmoth et expectancy of 1.8 years/decade, lower than predicted al. 2021; Human Mortality Database 2023). We took by Vaupel and colleagues but consistent with histor- two steps to reduce the effect of stochastic variabil- ical increases in our derived frontier life expectancy ity on frontier mortality: (1) we selected the lowest (webappendix Figure 1) (Oeppen and Vaupel 2002). mortality rates for each age range and year from 5x5 Frontier mortality rates for each age were then inter- mortality tables (5 years of age by 5 calendar years), polated to obtain annual estimates for the years 2000- and we excluded countries with a population of 1 2021. An implication of this approach is that frontier million or less in 2019. mortality rates continue to decline in 2020 and be- yond, despite the COVID-19 pandemic. This is con- Formally, we define sistent with continued declines in mortality in some of the lowest-mortality countries in 2020. In 2020,    [1] Japanese women’s life expectancy increased 0.3 years to 87.7, and South Korean women’s life expectancy where is avoidable mortality rate in also increased 0.2 years to 86.5 years in 2020 (Human deaths/1000 population in the age range [x,x+n) in Mortality Database, 2023). These increases were sim- country i, sex s, and year t, nmx,i,s,t is the observed/ ilar to the historical annual increase in frontier life ex- projected mortality rate in deaths/1000 population pectancy, and they occurred in the populations that in the age range [x,x+n) in country i, sex s, and year set the frontier for some age groups prior to 2020. t from the World Population Prospects (WPP) 2022, Avoidable deaths are computed as the product and is the lowest reliably observed mortality of age-country-sex-year avoidable mortality rates rate of any population-sex unit, obtained from pop- ( ) and age-country-sex-year population totals ulations included in the HMD of age range [x,x+n) from WPP2022. in year t or projected as described below (Human To provide context and asses the feasibility of mortality database, 2023; UNDESA, 2022). achieving the frontier mortality rates, we calculated We compared both male and female mortality to the historical average annual rate of change for all the lowest sex-specific mortality rate, which in near- country-year-sex-age combinations between 2000 ly all cases is a female mortality rate. In high-income and 2019 and identified the top 10th percentile per- settings, from which our frontier mortality rates are formers. We consider these observed changes to be drawn, excess male mortality is caused by injuries fast but plausible, and created a scenario (named and by non-communicable diseases (2). Excess inju- “rapid progress”) in which countries experience such ry deaths in males are likely a result of socio-cultural rates from the present to 2050. We compare this sce- factors rather than innate biological susceptibility. nario to the frontier for 2050 to assess the feasibility A higher proportion of male non-communicable of achieving the frontier mortality rates by 2050. disease deaths can be attributed to modifiable risk factors (3), again pointing to the importance of so- Estimating the economic cio-cultural differences causing higher male mor- value of avoidable mortality tality. We carry out a sensitivity analysis where we construct separate frontiers for males and females, The economic value of avoidable mortality in a given however, this sensitivity analysis is expected to un- year is measured as the percentage of annual income derestimate avoidable mortality in males. an individual is willing to forgo to live that year at the The HMD includes historical data as soon as frontier survival probabilities. We assign economic they are available, up to the 2022 calendar year, but values to changes in the mortality risk using VSL does not include projections. Frontier (or highest ob- estimates from the literature (Robinson et al. 2019). served) life expectancy has been shown to increase We follow the literature in assuming that indi- linearly at around 2.5 years per decade (Oeppen and viduals are willing to forgo part of their available re- Vaupel 2002; Vaupel et al. 2021). We projected log sources—usually consumption, wealth, or income— age-specific frontier mortality rates linearly for the for a lower mortality risk (Schelling 1968; Viscusi years 2022-2050 on the basis of observed mortality 1993). Specifically, we assume that the value of income rates from all complete 5-year periods in the HMD is modeled with a logarithmic function – the simplest COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 3 89 nonlinear member of the family of functions typical- Equation (2) characterizes the key elements of ly used to relate utility to income. Function our framework to calculate the economic valuation represents the discounted expected value of remain- of large mortality risk changes. We can check that ing lifetime income for an individual at age , where when a country is at the frontier, there is no mortality represents the vector of current change ( ), so the value is zero. As the distance annual survival probabilities at ages between the frontier mortality ( ) and a country’s and is annual income. For simplicity, we model one mortality ( ) increases, the value also increases. At country-representative individual for each age group. first, increases faster, with a rate governed primar- Moreover, annual income is constant over time and ily by the VSL-based parameter . As the distance equal to the gross national income per capita (denot- between and becomes larger, goes towards 1 ed by ). Then, we can write , (full income) but never exceeds it (see webappendix where is the value attributed to A1 for a detailed discussion of this equation). living one year with income level and is a nor- We closely followed the recommendations made malization constant, is by the Harvard Benefit Cost Analysis Reference Case analogous to a discounted remaining life expectancy, for conducting benefit cost analysis in global health is the annual discount rate, and is the probabil- in choosing the base VSL, setting the level of income ity of surviving from age until age . elasticity, placing a floor constraint on VSLr — the Let represent the frontier one-year sur- ratio between VSL and annual income per capita—, vival probability at age a. Then, the economic val- and conducting the recommended sensitivity anal- ue of avoidable mortality is implicitly defined by ysis scenarios (Viscusi and Aldy 2003). As a starting point, we set the VSLr at 160 (the ratio comes from a . This equality says that an individual at age is indif- United States VSL of $9.4 million and GNI per capita ferent between continuing under the status quo and a of $57,900), and income elasticity of 0.8 when extrapo- hypothetical scenario where she forgoes a proportion lating across countries with higher GNI per capita than of this year’s income in exchange for higher surviv- the United States, and 1.2 for countries with lower GNI al probability (no avoidable mortality) that year. per capita. We apply a lower bound constraint for the Using the equality above, we can obtain the ex- VSLr at 20. We choose to estimate income using GNI pression to calculate : the proportion per capita expressed in 2017 international dollars and of income that corresponds to the value of avoid- adjusted for purchasing power parity (PPP), as recom- able mortality. However, this procedure solves for a mended by the reference case. Data on GNI per capita family of functions. To pin down an exact solution, (PPP constant 2017 International $) between 2000- we use empirical estimates of the VSL. This is pos- 2021 came from the World Bank (World Bank 2021); sible because the VSL contains information about income levels for 2050 are projected using OECD’s pro- the value of income relative to small mortality risk jected country-specific growth rates between 2021 and changes (see webappendix A1 for details). Solving 2050 for listed countries (OECD and G20 countries), for allows us to derive an expression for the value and the world average growth rate during the same of eliminating avoidable mortality, which is given by time period for all remaining countries. The initial VSLr for all ages are set as equal. The annual discount rate is 3 percent. The economic value of avoidable mor- [2] tality is presented as percentage of annual income for both sexes, and females and males separately. Follow- ing the common approach in the literature, we used where is a country-specific parameter grounded on the same VSL for both sexes due to two reasons: lack empirical estimates of the VSL and background risk. of good data on sex-specific VSL, and that females Parameter is defined as , where likely have lower VSL due to lower labor income. For denotes reference age for the typical worker and example, Aldy & Smyth estimated about 13 percent is the estimated VSL.3 This parameter is analogous lower VSL for American females than males, reflecting to the Value per Statistical Life-Year (VSLY) for the a combination of females’ higher life expectancy (in- typical worker adjusted for background risk. crease in VSL) but lower labor market compensation 3   Using a reference value of 40 years based on VSL estimates for the United States, the VSLY = VSL / 40. 90 CHAPTER 3 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E (decrease in VSL) than males (Aldy and Smyth 2014). of major international investment in global health, We present the estimated avoidable mortali- current conditions just prior to the COVID-19 pan- ty and its associated economic value by the World demic, and future projections. We also discuss es- Bank’s geographic regions, with China and India pre- timates for 2021, which is the latest observed year sented separately. We focus on the years 2000, 2019, available from WPP2022 and reflects the unique and 2050, which represent the beginning of the era mortality profile during the COVID-19 pandemic. Sensitivity analyses Six sets of sensitivity analyses were conducted. In WPP2022 projections and the 2050 frontier. In es- estimating avoidable mortality, we first replaced the timating the value of mortality reduction, in addi- mortality frontier with sex-specific frontiers. In oth- tion to the three sets of standard sensitivity analyses er words, we applied the lowest frontier mortality proposed by the Harvard Benefit Cost Analysis Ref- levels for each sex separately to estimate avoidable erence Case (on income elasticities, discount rates, mortality by sex. Second, we compared mortality un- baseline VSL-to-income ratio), we further test differ- der the rapid progress scenario to business-as-usual ent value-of-income functions (linear, reciprocal). RESULTS During the four five-year periods between 2000-2019 and considering 19 age groups (76 age-period categories), Japanese women had the lowest observed mortality rate for 22 age-period categories; the next most frequently appearing population was Hong Kong SAR, China women (15 age-period categories; webappendix Table A1). Other populations that set the frontier for at least two age-period categories were women from Denmark, Finland, Italy, Slovenia, South Korea, Spain, Sweden, and Switzerland. Combining these age-specific mortality rates, frontier life expectancy was estimated to increase from 84.8 in 2000 to 88.5 in 2019, and was projected to reach 94.2 in 2050 (webappendix Figure A7), an increase of 1.8 years per decade. Frontier mortality rates for years 2000, 2019, and 2050 are depicted in Figure 3.1, and the gap between the frontier life expectancy at birth and observed or projected life expectancy is listed in Table 3.1. FIGURE 3.1 Frontier mortality rates from 2000, 2019, and 2050 Note: The scale of the y-axes for the two graphs are different. SMU = standardized mortality unit, deaths per 10-4 per year COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 3 91 TABLE 3.1 Gap between observed and frontier life expectancy at birth in years, 2000, 2019 and 2050, globally and by region Euroasia & Latin America & World China Mediterranean High-income India Caribbean Sub-Saharan Africa Males 2000 20.7 15.0 21.9 10.1 23.0 17.1 35.4 2019 18.3 13.2 19.7 9.7 18.9 16.9 29.5 2050 19.4 11.9 20.0 10.5 18.0 16.2 29.8 Females 2000 15.9 10.5 15.7 3.8 21.1 10.5 32.4 2019 13.1 7.5 13.4 4.3 16.0 10.5 25.8 2050 14.4 8.1 14.0 6.2 14.1 10.9 24.9 Note: estimated life expectancies at birth were compared to frontier life expectancies at birth, which were 84.8 in 2000, 88.5 in 2019, and 94.2 in 2050. Avoidable mortality as percentage of total mortality For the oldest age group – adults aged 80 years and was approximately 69 percent in 2019, equivalent to older – avoidable mortality as a percentage of to- 40 million deaths. Nearly half of avoidable deaths tal mortality ranged from 26 percent of deaths in occurred in Euroasia & Mediterranean (10.7 mil- High-income to 64 percent in sub-Saharan Africa. lion) and Sub-Saharan Africa (8.9 million). The per- Avoidable mortality in the oldest ages account for centage of deaths that were avoidable was lowest and an increasing proportion of overall avoidable mor- highest in High-income countries (42 percent) and tality, with avoidable mortality in ages 80 and over sub-Saharan Africa (91 percent), respectively (Fig- contributing 17 percent of avoidable mortality glob- ure 3.2, Table 3.2). More than 70 percent of mortal- ally in 2019 and anticipated to contribute 34 percent ity in every age group under age 80 was avoidable. of avoidable mortality in 2050. FIGURE 3.2 Avoidable mortality as percentage of total mortality, year 2019 92 CHAPTER 3 TABLE 3.2 Total mortality and percentage of mortality that is avoidable in 2019, globally and by region, for all ages and by age World China Euroasia & Mediterranean High-income India Latin America & Caribbean Sub-Saharan Africa total deaths total deaths total deaths total deaths total deaths total deaths total deaths (millions) % avoidable (millions) % avoidable (millions) % avoidable (millions) % avoidable (millions) % avoidable (millions) % avoidable (millions) % avoidable All ages 58.1 69 10 62 14.2 75 10.6 42 9.3 78 4.1 69 9.7 91 Infants 4.0 95 0.1 77 1.0 94 0.1 64 0.7 95 0.1 90 2.0 97 1-9 years 2.1 96 0.05 79 0.3 93 0.02 54 0.2 94 0.04 87 1.4 99 10-19 years 1.1 91 0.06 79 0.2 88 0.03 64 0.2 88 0.06 85 0.6 97 20-39 years 4.0 88 0.3 72 0.9 86 0.2 71 0.7 87 0.3 86 1.5 95 40-59 years 9.2 77 1.4 64 2.5 79 1.0 61 2.0 83 0.7 76 1.5 89 60-79 years 21.4 75 4.5 71 5.8 80 3.7 58 3.8 81 1.6 74 2.0 86 80 and over 16.3 43 3.7 48 3.5 55 5.6 26 1.7 50 1.2 51 0.7 64 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 3 93 Across regions, mortality rates have converged because average ages are increasing everywhere, in absolute terms over time, but the percentage of and the percentage of deaths that are avoidable de- mortality that is avoidable has changed little. Avoid- clines with age (Figure 3.3). Avoidable mortality as able mortality rates were estimated and projected to percentage of total mortality was approximately 72 decline, with a few exceptions (webappendix Figure percent in 2000, 69 percent in 2019, and projected A9). Regional trends in the percentage of deaths to be 70 percent in 2050 (Figure 3.3). Across all re- that are avoidable in each age group are a mixture gions, the lowest proportions of avoidable mortality of stable or, especially after 2010, increasing, in- were found in the high-income region, with 42, 42, dicating that relative gains in frontier mortality and 45 percent in years 2000, 2019, and 2050, re- rates have been and are projected to outpace gains spectively. The highest proportions were found in in mortality in many countries (Figure A10). De- sub-Saharan Africa (92, 91, and 92 percent in years spite increases in the percentage of deaths avoidable 2000, 2019, 2050), India (82, 78, 75 percent), and at each age, the percentage of all deaths that were Euroasia & Mediterranean (77, 75, 76 percent). avoidable changed little globally and in each region FIGURE 3.3 Avoidable mortality globally and by region, year, and age group Figure 3.4 depicts the gap between regional life tions, including women in high-income countries, expectancy and frontier life expectancy. From 2000- this gap is projected to grow as women in some 2010, this gap was decreasing for most regions. countries fall behind the frontier. Because males However, the gaps between frontier and projected have higher mortality rates, avoidable mortality in life expectancy are expected to change little after males is larger than in females, both in percentage 2025 due to the projected deceleration of reductions of deaths and number of deaths, in all regions and in avoidable mortality at all ages. In some popula- years (webappendix Figures A9 and A10). 94 CHAPTER 3 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E FIGURE 3.4 Gap in life expectancy at birth between observed/projected life expectancy and frontier life expectancy Based on the estimated levels of avoidable mortality decline in marginal economic value assigned in Figure presented above, we calculated the economic value of 3.5, using the example of age group 0-1 in the United avoidable mortality as percentage of annual income by States in 2019. As the decay curve shows, our model region, year, and sex. First, we illustrate the relation- assigns lower marginal economic value (VSMU) with ship between the size of mortality risk change and the increasing size of mortality risk reduction. FIGURE 3.5 Ratio of average VSMU to initial VSMU, for United States age group 0-1 Note: An SMU (standardized mortality unit) is a mortality rate of 10-4 per year. The y-axis represents the ratio of the value of SMU (VSMU) associated with small increment from initial mortality level to applied VSMU for active change in mortality level. In this example, we show the decline in the average VSMU with the change in SMUs for age group 0-1 in the United States. COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 3 95 TABLE 3.3 Value of avoidable mortality in 2019 globally and by region and age group Euroasia & Latin America & Sub-Saharan World China Mediterranean High-income India Caribbean Africa Economic value (2017 29320 4249 5080 12344 2077 1878 1380 international billion dollars, PPP) Economic value as percentage 23% 19% 22% 21% 23% 22% 34% of annual income Distribution by each age group (%) Infants 6 3 7 2 7 6 10 1-9 years 10 2 7 1 8 4 27 10-19 years 6 3 5 1 6 5 12 20-39 years 18 11 18 11 20 22 25 40-59 years 26 28 29 26 30 29 16 60-79 years 28 43 28 44 26 28 9 80 and over 6 10 5 15 4 6 1 FIGURE 3.6 Value of avoidable mortality as percentage of annual income for years 2000, 2019, 2021, and 2050 (projected) Globally in 2019, the economic value of avoid- cause older ages tend to have higher mortality) and able mortality (mean value, weighted by popula- relatively higher rates of avoidable mortality across tion) is estimated at about 29.4 trillion international all ages. As shown in the figure, at the same income dollars (2017, PPP-adjusted), representing approx- level, countries with higher avoidable mortality rate imately 23 percent of annual global income (Table have higher value of avoidable mortality as percent- 3.3, Figure 3.6). We estimated the highest percent- age of annual income. ages in sub-Saharan Africa (34 percent annual in- Different time trends are observed across re- come), followed by India (23 percent), Euroasia & gions (Figure 3.6). In sub-Saharan Africa, Euroasia Mediterranean (22 percent), Latin America & Ca- & Mediterranean, and India, we estimate the value ribbean (22 percent), High-income (21 percent), (as percentage of annual income) to decrease over and China (19 percent). The relationships between time. In sub-Saharan Africa, for example, the value the levels of avoidable mortality, income, and the was 46 percent in 2000, decreased to 34 percent in value of avoidable mortality are shown in Figure 3.7. 2019, and projected to decline further to 27 percent Higher avoidable mortality rate reflects the com- by 2050—nearly halved in 50 years. In Euroasia & bined effect of older population age structure (be- Mediterranean and India, large declines were seen 96 CHAPTER 3 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E FIGURE 3.7 Relationship between avoidable mortality rate, GNI per capita, and value of avoidable mortality Parameters: discount rate of 3%, baseline income elasticity, logarithmic valuation, 2019 data between 2000 and 2019, but the value is projected to olescents, and younger adults (ages 10 to 39) com- increase slightly by 2050. In comparison, the values prise almost half of the global population (47 per- for the other regions increased over time. The larg- cent) but only account for 24 percent of the total est increases are seen in China: from 17 percent in value of avoidable mortality. 2000, 19 percent in 2019, to 26 percent in 2050, re- Regional differences in population age struc- flecting the large projected increases in income and ture led to substantially different age group con- the proportion of older population. While China tributions to the total value of avoidable mortality, had the lowest values in 2000 and 2019, it is expect- as evidenced in Figure 3.8. In sub-Saharan Africa, ed to surpass many other regions in 2050. In con- approximately 37 percent of the value comes from trast, Euroasia & Mediterranean is expected to have ages 0–9, which accounts for 30 percent of the pop- the lowest value (at 23 percent) in 2050. ulation. Moreover, ages 60 and above represent Global and regional values of avoidable mor- about 10 percent of the value and only 5 percent of tality are unevenly distributed among different age the population. The distribution of contributions groups in all years. The two left-most columns in is markedly different in regions with older popula- Figure 3.8 show that infants (age 0) and ages 60 and tions. For instance, in High-Income countries, ages above account for a proportion of the total value 60 and above account for 60 percent of the total val- that is larger than their population weights. Glob- ue and represent 24 percent of the population. Fur- ally in 2019, ages 60 and above account for about thermore, with low infant mortality rates, age group 34 percent of the total value despite representing 13 0–9 accounts for only 3 percent of the value while percent of the population. Similarly, age 0 contrib- being 11 percent of the population. utes to 6 percent of the value while being less than 2 Looking at the differences by sex, Figure 3.9 percent of global population. The disproportionate shows that the global value of avoidable mortality contribution to value from these age groups reflects in 2019 is estimated at 19 and 27 percent of annual both their high avoidable mortality rates, especially income for females and males, respectively. Hence, for infants, and large absolute reductions in elderly the value of avoidable mortality is approximately mortality rates if avoidable mortality is eliminated 40 percent higher for males, reflecting the higher (see Figure 3.3). On the other hand, children, ad- avoidable mortality rates of males relative to females COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 3 97 FIGURE 3.8 Age-group contributions to the total value of avoidable mortality and population composition by age in year 2019 FIGURE 3.9 Value of avoidable mortality for females and males as percentage of annual income in year 2019 98 CHAPTER 3 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E (see Figure 3.4). Among females, we estimated the factors rather than biological differences (e.g., road lowest values in China (14 percent), High-income traffic injuries rather than prostate cancer; see meth- (15 percent), and Latin America & Caribbean (16 ods). This sensitivity analysis reclassified the largest percent), and the highest values in sub-Saharan Af- proportion of deaths in high-income countries and rica (32 percent), India (21 percent), and Euroasia the lowest proportion in sub-Saharan Africa, where & Mediterranean (18 percent). Among males, the around 90 percent of deaths are avoidable regardless lowest values were found in China (23 percent), of the method used to set the frontier. Nevertheless, India (25 percent), and Euroasia & Mediterranean regional patterns were similar in the sensitivity anal- (25 percent), and the highest values in sub-Saharan ysis vs. the baseline analysis. Africa (35 percent), Latin America & Caribbean (27 Second, to determine the extent to which the frontier percent), and High-income (26 percent). The larg- mortality levels are achievable, we considered a rap- est relative differences between the sexes were found id progress scenario in which countries experience in High-income (male’s value is 77 percent higher), fast and plausible (historically observed) rates of Latin America & Caribbean (64 percent), China (61 mortality change. Between 2000 and 2019, countries percent), and the smallest differences in sub-Saharan with the fastest 10th percentile decline in age-specific Africa (11 percent), India (17 percent), and Eurasia mortality rate experienced reductions between 3.4 & Mediterranean (39 percent). percent (ages 85+) and 12.2 percent (ages 5-9) an- We observe some changes in the estimated re- nually (webappendix Table A.2). Between 2019 and sults for 2021 during the COVID-19 pandemic. 2050, under this scenario, we would expect to see Globally, avoidable mortality as percentage of total 57 percent lower overall mortality compared to the mortality increased from 69 percent in 2019 to 74 baseline 2050 mortality rates projected by WPP2022 percent in 2021, and the gap between frontier and es- (webappendix Table A.3). In comparison, if all timated life expectancy at birth increased from 15.7 countries had the same mortality rates as the 2050 to 17.9 years. Across regions, the largest increases in frontier, we would expect 70 percent lower mortal- the share of avoidable mortality were in Latin Amer- ity level. Under the 10th percentile scenario, females ica & Caribbean (from 69 to 76 percent), India (78 from 30 high-income countries are projected to have to 84 percent), and High-income (42 to 47 percent), mortality rates lower than the 2050 frontier in 80 and lowest in sub-Saharan Africa (91 to 92 percent) percent of the age groups. and China (62 to 62 percent). The economic value Third, as expected, applying a fixed VSL (linear of avoidable mortality increased globally from 23 to extrapolation) yields much larger values of avoid- 25 percent, with the largest increases in Latin Amer- able mortality compared to using our parameters ica & Caribbean (22 to 27 percent), India (23 to 27 methods that adjust for the effect of large changes percent), and High-income (21 to 24 percent), and in risk. For instance, the global value of avoidable very little change in sub-Saharan Africa (34 to 34 mortality in 2019 using a fixed VSL is 35 percent of percent) and China (19 to 20 percent) (Figure 3.6). annual income (versus 23 percent with a logarith- mic function). The overestimation with fixed VSL Sensitivity analyses is more pronounced in regions with higher mor- tality rates, such as sub-Saharan Africa, where this Here, we highlight three key results from our sensi- method yields a value of 58 percent (versus 34 per- tivity analyses; for additional results and details, see cent with a logarithmic function). These differences webappendix A2. First, replacing the mortality fron- are particularly striking within high-mortality age tier with sex-specific frontier reduced the estimat- groups and countries, where values estimated with a ed avoidable mortality among males (webappendix fixed VSL can far exceed annual income–and some- Figure A1). Globally, this reduced the percentage times reach up to 1600 percent of income. In com- mortality considered avoidable in 2019 from ap- parison, applying a negative reciprocal function— proximately 69 percent to 59 percent. As previously another widely-used non-linear adjustment— only discussed, this likely underestimates male avoidable slightly reduces the values; for example, the global mortality because causes of excess male mortality in value for 2019 is 20 percent of income (versus 23 frontier countries are associated with socio-cultural percent with a logarithmic function). COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 3 99 DISCUSSION We estimated the economic value of avoidable mortality from 2000-2019 and projected to 2050 by region, age, and sex. In other words, we estimated the perceived economic benefits, measured as percentage of annual income, if countries were to live at the lowest possible (frontier) mortality levels. Using the frontier approach, we first identified countries with the lowest all-cause mortality by age and estimated avoidable mortality for all countries. Then, using the VSL approach with non-linear adjustment, we estimated the proportion of annual income one is willing to forgo to live at the mortality frontier. We highlight five key messages from our findings. First, a high percentage of mortality is avoidable: es in mortality rates between the frontier countries Our comparisons of frontier mortality to estimated and the rest, approximately 70 percent of all mor- mortality show that, globally, 69 percent of deaths tality was deemed avoidable, and this number does would have been postponed in 2019 if all countries not change much across time. Second, in line with experienced frontier mortality rates, with the low- the literature, we placed high monetary value on re- est and highest percentages in the High-income ducing mortality risk (the VSL-to-income ratio at region (42 percent) and sub-Saharan Africa (91 160). The combination of these two factors resulted percent), respectively. While the absolute level of in high values of avoidable mortality. mortality is projected to decline globally over time, The levels and changes in economic value re- the proportion of avoidable mortality has changed flect multiple factors beyond the levels and changes little. Although past illness, injuries, and risk expo- in mortality. First, it reflects national income and sures mean that a rapid shift to frontier mortality how it is projected. For example, Poland and Tajik- rates would not be possible, this analysis demon- istan show approximately the same proportion of strates that there are large opportunities for health avoidable mortality (81.3 and 82.6 percent, respec- improvements worldwide, even in higher income tively) and annual survival probabilities for males countries and among older adults. at ages 50–54. Nevertheless, Poland’s per capita in- Second, it is plausible to significantly reduce come is 8.5 times larger than Tajikistan’s ($31,800 the level of avoidable mortality by 2050: Although versus $3,700 in 2017 USD PPP). As a result, the many countries do not reach frontier mortality value of avoidable mortality for this segment of levels by 2050 under a fast but plausible mortali- their populations is 47.6 percent of annual income ty decline scenario, the gap between the projected for Poland but only 34.8 percent for Tajikistan. and frontier mortality would narrow significantly. Hence, Tajikistan’s value is lower both in absolute If countries experienced the historical 10th percen- dollars and relative to national income levels. Sec- tile declines in mortality, by 2050, they will have re- ond, the estimates are weighed by the age distribu- duced avoidable mortality by 81 percent. We also tion of the population. For illustration, we compare highlight that historical fast declines were frequent- Japan and Morocco, where the economic values of ly observed in some age groups in non-high-income avoidable mortality are estimated at 14.5 and 16.7 countries, suggesting that these rates are achievable percent of annual income in 2019, respectively. Mo- regardless of income level. rocco has lower population-weighted economic val- Third, the economic value of avoidable mortal- ue of avoidable mortality than Japan despite having ity is large: globally in 2019, the economic value of much higher level of avoidable mortality across all avoidable mortality represented approximately 23 ages, and especially in older ages, than Japan, where percent of global annual income, with the lowest avoidable mortality is close to the frontier for all ex- and highest percentages in China (19 percent) and cept among older males. On the other hand, even sub-Saharan Africa (34 percent), respectively. The though Japan’s older age groups are assigned low- economic value is approximately 4 percent higher er values (as percentage of income) than Morocco for males than females, reflecting the higher avoid- because of its relatively lower mortality, Japan has able mortality rates of males. Hence, these results in- a much higher proportion of older adults, whose dicate that, on average, regional populations would values are higher than younger people. Thus, coun- be willing to give up about one-fifth to one-third tries with larger proportions of older populations of their current income in exchange for a year liv- (like Japan) have higher economic values of avoid- ing at the lowest possible mortality rate, even when able mortality than younger countries (like Mo- considering resource constraints. Two main factors rocco) even though the former has lower levels of drive these results. First, due to the large differenc- avoidable mortality across all ages. Third, baseline 100 CHAPTER 3 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E mortality rates should also be considered. Reducing First, similar to the GBD’s approach in creat- avoidable mortality by 50 percent in a low-mortality ing an “envelope” to ensure the sum of cause-spe- country yields a lower value than the same propor- cific d eaths d o n ot e xceed a ll-cause d eaths, o ur tional change in a high-mortality country. estimates on the economic value of avoidable mor- Fourth, avoidable mortality and its econom- tality could be used as an envelope for the value of ic values vary substantially across age groups: as mortality reduction from specific causes, interven- shown in Figure 3.8, some age groups (ages 0 and tions or programs. 60 and above) contribute disproportionately high- Second, avoidable mortality is typically estimat- er value than others (relative to their population ed as the number of deaths assigned to set of specific share), as a result of the higher rates of avoidable causes of death, when they occur below a threshold mortality in those age groups. Previous literature on age (OECD 2019; Rutstein 1998; Nolet and McKee estimating amenable or preventable mortality that 2004). This list-based approach requires compa- use list-based approaches exclude deaths among age rable, detailed, and accurate cause-of-death data, 70 or 75 and above (OECD 2019;Barber et al. 2017). which are not currently available for most countries In comparison, our approach estimates avoid- (WHO 2020). We propose a simplified method by able deaths in all age groups. For example, using a computing avoidable mortality as the difference list-based approach, OECD found that more than between current observed or estimated mortality a quarter of deaths in OECD countries—mainly and an unavoidable, or frontier, mortality (Mathers high-income countries with high-quality cause-of- et al. 2015). This has several advantages, including death information—were avoidable in 2019 (OECD application to countries with limited cause-of-death 2021). In our analysis, we found that 42 percent of information, allowing for a time-varying frontier deaths in high-income countries were avoidable based on improvements in knowledge and technol- in 2019, with more than half of avoidable deaths ogy, and inclusion of avoidable mortality in older occurring in ages 75 and over. Thus, our analysis adults, and avoiding artificial categorization into demonstrates that exclusion of avoidable deaths in preventable and treatable deaths. older adults results in a substantial undercount of Third, the global health economics literature overall avoidable mortality. using the VSL approach has traditionally relied on Fifth, it is necessary to adjust the VSL for large existing valuation studies to value mortality risk re- risk change. Existing global health literature sel- duction in various settings. This process, referred to dom considers the issues associated with using lin- as “benefit transfer” or “value transfer”, typically in- ear approximation when extrapolating values with volve extrapolating the VSL from the United States larger mortality changes. For example, we find that or western Europe to another setting based on diffe- linear extrapolation yields a value that is on aver- rences in the income levels and the assumed level of age 54 percent higher than estimates using a log- income elasticity (Robinson et al. 2019; Robinson et al. arithmic value of income function. For certain age 2019). This step implicitly assumes that income is the groups in countries with high avoidable mortality only difference between the settings, such that income rates, linear extrapolation results in values that adjustment to the VSL is sufficient. However, several are nearly ten times larger than our estimates. The other key differences may exist between the original logarithmic approach also ensures that valuation research from which the VSL estimate was derived of mortality reduction does not exceed resource and the new setting. They may include population constraints, and we believe future research should age distribution, size of the change in mortality consider adopting this method. risk, baseline mortality, and whether the risk change is one-time or permanent. There is a lack of empirical Contributions to the literature evidence on the relationship between these factors and the VSL (with the exception of population We aim to make three important contributions to age), which may have led to the omission of the literature: (1) creating an envelope for global quantitative adjustments or discussions around economic value of mortality reduction; (2) imple- them in the global health benefit cost literature. In menting a simplified method to estimate avoidable this chapter, we rely on the theoretical literature and mortality; and (3) incorporating risk-size adjust- introduce a novel and relatively easy approach to ments that are necessary to the economic valuation incorporate these factors into the VSL estimate framework in this context. (Hammitt 2020; Cardoso and Dahis 2024). This is COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 3 101 especially important in global health research, of disease has taken different approaches. One ap- which often focuses on a wide range of baseline proach is using a list-based approach, for example and size of risk reduction across time and region. work by Alkire and colleagues provide estimates of amenable mortality, i.e., mortality that can be re- Limitations duced by the health care system, for 130 low- and middle-income countries between 2015-30 (Alkire This study has s everal l imitations. F irst, mortality et al. 2018). The second approach looked at what and population projections from WPP2022 are un- specific packages of interventions could be expect- certain. For many countries, WPP2022 estimates ed to do if more fully implemented than they now mortality rates by country, age and sex from incom- are. These assessments have been undertaken in plete death registration data and population-based the context of estimating the impact of universal census or survey data, such as reports of child or health coverage on mortality (and on health care sibling survival, as available. The s ame l imitation costs). Two analytic programs, from the WHO and holds for country-level income data. We therefore the third edition of the Disease Control Priorities focus on the results at the regional level in the main (DCP) Project, respectively, have generated esti- paper to mitigate the effect of outlying country esti- mates, and both point to a substantial volume of mates. Second, the medium- and long-term impact amenable mortality (and are thus consistent with of COVID-19 on levels and trends in mortality is the main conclusion of this study) (Stenberg et al. uncertain. As populations gain immunity (from 2019; Stenberg et al. 2017; Jamison and Alwan 2018; past infection and vaccination) and treatments Watkins et al. 2020). In terms of valuing these mor- continue to improve, we expect that mortality tality reductions, Alkire et al. estimate the value of rates will resume their declines in all populations amenable mortality at $6.0 trillion in 2015 (inter- and our projections for 2050 may be considered national dollars) for all low- and middle-income informative. Third, w e o nly f ocused o n m ortality countries, and approximately 5 to 20 percent of and not morbidity, suggesting that the results likely GDP in different regions Our methodological dif- underestimate the actual economic value of pre- ferences – the three most important being how and venting and treating diseases and injuries. Fourth, which mortality is valued, inclusion or exclusion of our projected results for 2050 are sensitive to some ages over 75, and the use of linear versus logarith- key assumptions, one of which is the projected GNI mic value-to-income functions – make it difficult to per capita in 2050. While we relied on the project- compare the estimates directly. Furthermore, put- ed growth rates of a reliable source (OECD), the ting our numbers into context, while not directly results are highly sensitive to how income levels comparable, in 2019 national health expenditure in 2050 are projected. Fifth, t he VSL approach t o (as percentage of GDP) in all countries were much assigning monetary values to risk reduction is not less than our estimate, suggesting that people may without its limitations (Viscusi and Aldy 2003). be willing to invest much more in health promotion While we have introduced important adjustments and health care or other means in exchange for bet- when extrapolating the VSL from the United States ter survival probabilities (World Bank 2022). Studies to other countries, having empirical studies from on primary health care in low- and middle-income countries with different i ncome a nd r isk l evels i s countries reported that an additional 3.3 percent of preferred. Finally, our results estimate one’s will- GDP on top of current levels of health expenditure ingness to forgo income for a one-year change in could lead to large gains in health (Stenberg et al. survival. An alternative perspective and approach 2019). Another study estimated an additional 4.6 would be to estimate the exchange between lifetime percent of GDP is required to achieve the Sustain- income and lifetime changes in survival, which able Development Goal 3 by 2030 (Stenberg et al. could also be insightful for policymaking. 2017). The DCP studies concluded costs would be somewhat higher. These numbers are much lower Comparison to previous literature than our estimates of value. Previous work on estimating the economic burden 102 CHAPTER 3 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CONCLUSION As governments worldwide engage in policy dialogues on how and how much to invest in improving population health, our work provides supportive evidence on the high economic value placed on improving health. Further- more, the disproportionally higher distribution of the economic value of avoidable mortality among older adults also lends support to policies related to healthy aging and interventions targeting non-communicable diseases. DEDICATION We dedicate this study to Professor James W. Vaupel (1945-2022) who inspired us and many others with his intellect, vision, and passion for great research. REFERENCES 1. Aldy JE, Smyth SJ. Heterogeneity in the Value of Life. National 15. Institute for Health Metrics and Evaluation (IHME). GBD Results Bureau of Economic Research, 2014. Tool | GHDx. http://ghdx.healthdata.org/gbd-results-tool (ac- 2. Alkire BC, Peters AW, Shrime MG, Meara JG. The Economic Con- cessed May 20, 2021). sequences of Mortality Amenable To High-Quality Health Care 16. Jamison DT, Alwan A, Mock CN, et al. Universal health coverage In Low- And Middle-Income Countries. Health Aff (Millwood) and intersectoral action for health: key messages from Disease 2018; 37: 988–96. Control Priorities. The Lancet 2018; 391: 1108–20. 3. Barber RM, Fullman N, Sorensen RJ, et al. Healthcare Access 17. Jamison DT, Summers LH, Alleyne G, et al. Global health 2035: and Quality Index based on mortality from causes amenable a world converging within a generation. The lancet 2013; 382: to personal health care in 195 countries and territories, 1990– 1898–955. 2015: a novel analysis from the Global Burden of Disease Study 18. Khadka A, Verguet S. The economic value of changing mor- 2015. The lancet 2017; 390: 231–66. tality risk in low- and middle-income countries: a systematic 4. Barbieri M, Wilmoth JR, Shkolnikov VM, et al. Data resource breakdown by cause of death. BMC Med 2021; 19: 156. profile: the human mortality database (HMD). Int J Epidemiol 19. Luy M. The impact of biological factors on sex differences in life 2015; 44: 1549–56. expectancy: insights gained from a natural experiment. 2016. 5. Bloom DE, Canning D, Jamison DT. Health, wealth, and welfare. 20. Mathers CD, Stevens GA, Boerma T, White RA, Tobias MI. Causes Finance Dev 2004; 41. of international increases in older age life expectancy. The Lan- 6. Bressler RD, Heal G. Valuing Excess Deaths Caused by Climate cet 2015; 385: 540–8. Change. National Bureau of Economic Research, 2022. 21. Murray CJ, Lopez AD, Jamison DT. The global burden of disease 7. Cardoso DS, Dahis R. Calculating the economic value of in 1990: summary results, sensitivity analysis and future direc- non-marginal mortality risk reductions. Economics Letters. tions. Bull World Health Organ 1994; 72: 495. 2024 Mar 21:111673. 22. Nolte E, McKee M. Measuring the health of nations: analysis 8. Chang AY, Horton S, Jamison DT. Benefit-cost analysis in dis- of mortality amenable to health care. J Epidemiol Community ease control priorities. 2018. Health 2004; 58: 326–326. 9. Chang AY, Robinson LA, Hammitt JK, Resch SC. Economics in 23. Oeppen J, Vaupel JW. Broken limits to life expectancy. Ameri- “Global Health 2035”: a sensitivity analysis of the value of a life can Association for the Advancement of Science, 2002. year estimates. J Glob Health 2017; 7. 24. Organization for Economic Co-operation and Development 10. GBD 2019 Risk Factors Collaborators. Global burden of 87 risk (OECD). Avoidable mortality: OECD/Eurostat lists of prevent- factors in 204 countries and territories, 1990–2019: a system- able and treatable causes of death. 2019; published online atic analysis for the Global Burden of Disease Study 2019. The Nov. https://www.oecd.org/health/health-systems/Avoid- Lancet 2020; 396: 1223–49. able-mortality-2019-Joint-OECD-Eurostat-List-preventable- 11. Hakulinen T, Hansluwka H, Lopez AD, Nakada T. Global and treatable-causes-of-death.pdf (accessed July 29, 2021). Regional Mortality Patterns by Cause of Death in 1980. Int J 25. _____. Health at a Glance 2021. 2021. https://www.oecd-ili- Epidemiol 1986; 15: 226–33. brary.org/content/publication/ae3016b9-en. 12. Hammitt JK, Herrera-Araujo D. Using non-marginal risk reduc- 26. Revesz RL, Livermore MA. Retaking rationality: How cost-bene- tions to elicit VSL: 28. fit analysis can better protect the environment and our health. 13. _____. Valuing mortality risk in the time of COVID-19. J Risk Oxford University Press, 2008. Uncertain 2020; 61: 129–54. 27. Robinson LA, Hammitt JK, Cecchini M, et al. Reference Case 14. Human Mortality Database. Max Planck Institute for Demo- Guidelines for Benefit-Cost Analysis in Global Health and De- graphic Research (Germany), University of California, Berkeley velopment. 2019; : 126. (USA), and French Institute for Demographic Studies (France). 28. Robinson LA, Hammitt JK, Jamison DT, Walker DG. Conducting www.mortality.org (accessed March 23, 2023). Benefit-Cost Analysis in Low- and Middle-Income Countries: COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 3 103 Introduction to the Special Issue. J Benefit-Cost Anal 2019; 10: 118: e2019536118. 1–14. 39. Viscusi WK, Aldy JE. The Value of a Statistical Life: A Critical Re- 29. Robinson LA, Hammitt JK, O’Keeffe L. Valuing Mortality Risk view of Market Estimates Throughout the World. J Risk Uncer- Reductions in Global Benefit-Cost Analysis. J Benefit-Cost Anal tain 2003; 27: 5–76. 2019; 10: 15–50. 40. Viscusi WK. Fatal tradeoffs: Public and private responsibilities 30. Rutstein DD, Berenberg W, Chalmers TC, Child CG, Fishman for risk. Oxford University Press, 1992. AP, Perrin EB. Measuring the quality of medical care. A clinical 41. _____. The Value of Risks to Life and Health. J Econ Lit 1993; method. N Engl J Med 1976; 294: 582–8. 31: 1912–46. 31. Schelling TC. The life you save may be your own. Probl Public 42. _____. Pricing lives. In: Pricing Lives. Princeton University Expend 1968: 127–62. Press, 2018. 32. Stenberg K, Hanssen O, Bertram M, et al. Guide posts for in- 43. Watkins DA, Qi J, Kawakatsu Y, Pickersgill SJ, Horton SE, Jamison vestment in primary health care and projected resource needs DT. Resource requirements for essential universal health cover- in 67 low-income and middle-income countries: a modelling age: a modelling study based on findings from Disease Control study. Lancet Glob Health 2019; 7: e1500–10. Priorities. Lancet Glob Health 2020; 8: e829–39. 33. Stenberg K, Hanssen O, Edejer TT-T, et al. Financing transforma- 44. Wilmoth JR, Andreev K, Jdanov D, Glei DA, Riffe T. Methods pro- tive health systems towards achievement of the health Sus- tocol for the Human Mortality Database. 2021 https://www. tainable Development Goals: a model for projected resource mortality.org/File/GetDocument/Public/Docs/MethodsProto- needs in 67 low-income and middle-income countries. Lancet colV6.pdf (accessed Oct 4, 2023). Glob Health 2017; 5: e875–87. 45. World Bank. World Development Report 1993: Investing in 34. Sunstein CR. Risk and reason: Safety, law, and the environment. Health. Washington, D.C.: The World Bank, 1993 https://open- Cambridge university press, 2002. knowledge.worldbank.org/handle/10986/5976. 35. _____. Valuing life. In: Valuing Life. University of Chicago Press, 46. World Bank. World Development Indicators: GNI per capita, 2014. PPP (current international $) 2021. https://data.worldbank. 36. United Nations Department of Economic and Social Affairs org/indicator/NY.GNP.PCAP.PP.CD (accessed Feb 24, 2021). (UNDESA), Population Division. World population prospects 47. World Bank. World Development Indicators: Current health 2022. https://population.un.org/wpp/Download/Standard/ expenditure (% of GDP). https://data.worldbank.org/indicator/ CSV/ (accessed Oct 4, 2023). SH.XPD.CHEX.GD.ZS (accessed July 7, 2022). 37. Usher D. An imputation to the measure of economic growth 48. World Health Organization (WHO). Global Health Estimates. for changes in life expectancy. In: The measurement of eco- 2020. https://www.who.int/data/global-health-estimates (ac- nomic and social performance. NBER, 1973: 193–232. cessed July 29, 2021). 38. Vaupel JW, Villavicencio F, Bergeron-Boucher M-P. Demograph- 49. _____. WHO methods and data sources for country-level ic perspectives on the rise of longevity. Proc Natl Acad Sci 2021; causes of death 2000-2019. Geneva, Switzerland, 2020. 4 104 CHAPTER 4 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E The Economic Value Associated with Avoidable Mortality A systematic assessment by cause of death across world regions Stéphane Verguet a, Sarah Bolongaita a, Angela Y. Chang b,c, Diego S. Cardoso d, Gretchen A. Stevens e a Department of Global Health and Population, Harvard T.H. Chan School of Public Health b Danish Institute for Advanced Study, University of Southern Denmark c Interdisciplinary Centre on Population Dynamics (CPop), University of Southern Denmark d Department of Agricultural Economics, Purdue University e Independent Researcher © Harvard University COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 4 105 INTRODUCTION Before the onset of the COVID-19 pandemic in 2020, life expectancy at birth had been steadily increasing globally (UN 2019); and the advancements to older ages and frontiers of longevity were still regularly progressing (Vau- pel et al. 2021). In many countries around the world, populations are rapidly ageing, immediately following the combination of major declines in fertility and important progresses in longevity (Christensen et al. 2009). Such rapid ageing transitions are often accompanied with the development of disability and dependence, including increasing prevalence of diseases of old age. This imposes new and substantial challenges to health systems, neces- sitating the design of novel technologies, the implementation of innovative health services delivery platforms, and the strengthening of financial architectures (Partridge et al. 2018; Spasova et al. 2018; Tynkkynnen et al. 2022). Facing limited resources and an increasing burden of non-communicable diseases (NCDs), health systems must make difficult trade-offs when allocating resources toward improving the healthy longevity of their populations. Researchers and policymakers utilize a variety of tors (e.g., health, education, agriculture), including analytical methods for generating evidence to in- health interventions in low- and middle-income form decision-making on which health policies countries (LMICs) (Jamison et al. 2013; Jha et al. and interventions should be prioritized, funded (or 2013). Likewise, in their seminal work (Jamison not), and implemented. Commonly used approach- et al. 2013) in which they used the value of a sta- es include various types of multi-criteria decision tistical life (VSL) to estimate the monetary gains analyses (MCDA) (Baltussen et al. 2018; Verguet from reducing under-five mortality rates in LMICs, et al. 2021) and economic evaluations, traditionally Jamison and colleagues demonstrated the utility of drawn from cost-effectiveness analysis (CEA) meth- monetized health outcomes. Most recently, Khad- ods, of the health policies and interventions under ka & Verguet (2021) adapted such VSL methods to consideration (Drummond et al. 2015; Neumann et assign an economic value to life years and devel- al. 2016). Such CEA studies often use a constructed op further value of a statistical life years, and ap- metric, like the quality-adjusted life year (QALY) plied methodologies to assess the economic value or the disability-adjusted life year (DALY) (Wein- of changing mortality risks associated with NCDs. stein et al. 2009; Murray & Acharia 1997), which In doing so, they pointed to the vast welfare losses encompass both mortality and morbidity outcomes associated with elevated NCD mortality rates, espe- of the disease(s) addressed by the interventions un- cially cardiovascular diseases (CVDs) and cancers. der consideration. However, using non-monetary In this chapter, we build on this previous VSL valuations in health intervention assessments limits scholarship. We compute the economic impact of their broader utility, preventing comparisons with disease, by first quantifying the avoidable mortality assessments of interventions outside of the health by cause of death (with a focus on major NCD and sector (for example, education and social protec- injury causes) and second applying novel economic tion) (Robinson et al. 2019). valuation methods. Specifically, we compute a mon- Yet, an attractive alternative analytical method etary value for hypothetical reductions (over 2000- that can address this intersectoral comparison issue 2050) in cause-specific mortality: first, we bench- is benefit-cost analysis (BCA; Robinson et al. 2019; mark these reductions against a mortality frontier Robinson, Hammitt & O’Keeffe 2019), where both (lowest mortality rates achievable) by cause; and the numerator (i.e., the health benefits procured second, we assign a willingness to pay value to these by a given intervention) and the denominator (i.e., reductions in cause-specific mortality. In sum, we the costs incurred by a given intervention) of the estimate the percent of income an individual would benefit-cost ratio are expressed in monetary terms be willing to forgo to live one year under the lowest (say in USD gained or spent). As an example, the possible mortality rate for a given cause of death. Copenhagen Consensus Center1 has long used BCA Our assessment is pursued for six large analytical approaches to consistently compare the returns on regions (e.g., China, Eurasia & the Mediterranean, investment of a variety of policies globally across High-income, India, Latin America & the Carib- multiple key human capital and development sec- bean, and sub-Saharan Africa) and major disease 1  www.copenhagenconsensus.com 106 CHAPTER 4 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E areas (e.g., cancers, CVD), with findings presented globally and regionally, and to derive a metric di- for the years 2000, 2019, and 2050. Our objective is rectly comparable to annual incomes (e.g., gross to provide a systematic monetary assessment of the national income (GNI)), so as to enable both health economic value associated with mortality by cause sector and multisectoral priority setting. of death (with a focus on major NCDs and injuries) METHODS General approach comprehensive list of causes to a smaller list of mu- We proceeded in three steps. First, we projected tually exclusive, collectively exhaustive set of causes age-cause-sex-specific mortality rates into the fu- primarily focused on NCDs and injuries (Table 4.1; ture (2020-2050) for each country included in the webappendix Table A2 gives a mapping between analysis (webappendix Table A1). We estimated GHE causes of death and the causes of death used in age-cause-specific mortality frontiers ( ; a the analysis). These causes of death were selected as = age, K = disease cause, s = sex*)2 for the period they were leading causes of mortality and morbidity 2000-2019, with the mortality frontiers being the according to burden of disease assessments (WHO lowest (10th percentile) observed mortality rates for 2020a; IHME 2022). For cancers, we focused on a given cause and age group, globally. We then pro- tobacco-related (e.g., lung, mouth, esophagus) and jected these frontiers over 2020-2050. They can be infection-related (e.g., cervix, liver, stomach) can- thought of as “aspirational” mortality rates (for both cers; breast cancer was selected as it was the leading females and males) for countries that experience cancer among females (WHO 2020a; IHME 2022). mortality rates greater than the frontier. Age refers to five-year age groups (except for the Second, we compared country- and cause-spe- first two and final age groups): 0-1, 1-4, 5-9, 10-14, 15- cific mortality rates with frontier mortality rates to 19, …, 75-79, 80-84, and 85+ years. Age- and sex-spe- yield mortality differentials (i.e., “avoidable mortal- cific population and all-cause mortality estimates and ity”, the gap between a country’s mortality rate and projections for all countries for the period 2000-2050 the frontier mortality rate). We then aggregated were sourced from the United Nations (UN) 2022 these differentials across six analytical regions (in- World Population Prospects (WPP) (UN 2022). spired by the classification proposed by The Lancet Commission on Investing in Health (Watkins et al. 2018)): China; Eurasia & the Mediterranean; India; Estimating the mortality frontiers High-income; Latin America & the Caribbean; and The mortality frontiers (hereafter referred to as the sub-Saharan Africa (webappendix Table A1). “frontiers”) were defined as the 10th percentile of the Third, we assigned a monetary value to the age-cause-specific mortality rates of eligible coun- avoidable mortality. This gave an estimated percent tries in a given year (Mathers et al. 2015; Chang et al. of annual income an individual would be willing to 2024). Countries were included in the frontier anal- forgo to live one year under the lowest possible mor- ysis (webappendix Table A1) if they had populations tality rate for a given cause of death; and a metric of at least 5 million in 2019; had available income directly comparable to national incomes (e.g., GNI). (GNI per capita, current international dollars) data These three steps are further detailed below. for 2019; had high-quality vital registration data (as defined in WHO 2020b); and were included in GHE Data sources estimates for the year 2019. The frontiers were age- and cause-specific for all causes of death except for Estimates on age-cause-sex-specific death counts for breast cancer and cervix uteri cancer, for which age- all countries for 2000-2019 were sourced from the sex-cause-specific frontiers were computed given World Health Organization’s (WHO) Global Health the sex-specific epidemiology of these two condi- Estimates (GHE) (WHO 2020a). We adapted WHO’s tions.3 The frontier was defined as the 10th percentile 2  The frontier was only sex-specific for the following two sex-specific causes: breast cancer and cervix uteri cancer. For all other causes, the frontier was selected from among sex-specific mortality rates of both sexes. 3  Male-specific cancers, such as testicular or prostate cancers were not included as they are less common and affect older ages, so their overall burden is much lower. COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 4 107 of eligible country mortality rates, as opposed to the summed to the frontiers of higher (“parent”) level minimum, to ensure stability and the absence of sto- causes (Table 4.1), we utilized a nested, level-wise ap- chastic variation in the resulting computed frontiers proach. We first extracted the frontier (i.e., the low- (see details in webappendix Figure A1). est 10th percentile of eligible country mortality rates) For the period 2000-2019, frontiers were cal- for all-cause mortality, that is the highest level (level culated directly using GHE estimates. To ensure 0) cause. The frontiers for the next level of causes that the frontiers of lower-level causes appropriately (level 1: communicable, maternal, perinatal and nu- TABLE 4.1 Causes of death Level Cause of death 1 Communicable, maternal, perinatal and nutritional conditions 2 Infectious and parasitic diseases 2 Maternal and neonatal conditions 2 Nutritional deficiencies 1 Noncommunicable diseases 2 Cardiovascular diseases 3 Ischaemic heart disease 3 Stroke 3 Other cardiovascular diseases 2 Diabetes mellitus 2 Digestive diseases 3 Cirrhosis of the liver 3 Other digestive diseases 2 Malignant neoplasms 3 Breast cancer 3 Cervix uteri cancer 3 Liver cancer 3 Mouth and oropharynx cancers 3 Oesophagus cancer 3 Stomach cancer 3 Trachea, bronchus, lung cancers 3 Other malignant neoplasms 2 Respiratory diseases 3 Chronic obstructive pulmonary disease 3 Other respiratory diseases 2 Other noncommunicable diseases 1 Injuries 2 Intentional injuries 2 Unintentional injuries 3 Road injury 3 Other unintentional injuries Note: The causes of death utilized in this analysis were adapted from the WHO’s Global Health Estimates (WHO 2020a). Level 1 and 2 causes of death are mutually exclusive and collectively exhaustive; within level 2 causes of death, level 3 causes are mutually exclusive and collectively exhaustive. Webappendix Table A2 gives a mapping between GHE causes of death and the causes of death used in the analysis. 108 CHAPTER 4 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E tritional conditions; NCDs; and injuries) were then able mortality developed under Chang et al. 2024), extracted. These level 1 frontiers were then summed our all-cause frontier mortality rates (immediately to calculate the parent level (level 0: all-cause) fron- described above) were scaled to these all-cause fron- tier. The resulting calculated level 0 all-cause fron- tiers. We divided our age- and year-specific all-cause tier was then compared with the extracted level 0 frontiers by the corresponding external age- and all-cause frontier to determine a scaling factor. This year-specific all-cause frontier mortality rates to pro- scaling factor was then applied to the component duce a set of age-year-specific scaling factors between level 1 frontiers. This process was continued for level our all-cause frontiers and the Chang and colleagues’ 2 causes (e.g., CVD, malignant neoplasms) and level frontiers. The resulting scalars were then applied to 3 causes (e.g., stroke, stomach cancer). our cause-specific frontiers (i.e., cause levels 1, 2, and For the period 2020-2050, the frontiers were 3) to ensure consistency with the avoidable mortality calculated based on the trends underlying the 2010- frontiers put forward by Chang et al. (2024). 2019 frontiers (immediately described above). Log-linear models were used to project the frontier Estimating mortality projections mortality rates using the following equation: for the six regions ,  [1] First, we projected country- and cause-specific mor- tality rates for the period 2020-2050 using the same where is the mortality rate for age group in methods as for the frontier mortality rate projections year (2000, 2019, and 2050), and cause of death described immediately above. Second, we aggregated is a constant term, and is the year. is these country-specific mortality rate projections with- the estimated annual rate of change (decline) of the in each region studied: China; Eurasia & the Medi- mortality rate associated with each cause and broad terranean; India; High-income; Latin America & the age group (0-4, 5-14, 15-29, 30-44, 45-59, 60-69, Caribbean; and sub-Saharan Africa. For that purpose, 70-84, and 85+ year-olds). The projections (based population weights (within each age group) based on and [1]) were conducted separately by cause on country- and age-specific populations using 2022 and broad age group to ensure similar mortal- WPP (UN 2022) were used to eventually yield region- ity trends across neighboring five-year age groups, age-sex-cause-specific mortality rate projections. consistent with previous projection analyses (Ma- thers & Loncar 2006). Economic value associated with cause- When the 2010-2019 frontier mortality rates specific avoidable mortality contained a zero-value (i.e., an observation of 0 deaths per 100,000 population) or when the average We denote the mortality rate among the age number of deaths (averaged over the 2010-2019 fron- group , where is the size of the age group tiers) for a given age-cause group was less than 100, ( , except for 0-1, 1-4, and 85+ age groups). The then the average mortality rate (for the period 2010- probability of dying at age , , is expressed as 2019) was used for the 2020-2050 projection for all ; using a discrete formulation, five-year age groups within the corresponding broad . age group category (yielding a constant frontier over We compare a country’s age group’s probabil- 2020-2050). Additionally, if the regression ([1]) gave ity of dying (“C” for country) to a reference prob- a positive trendline ( , which would indicate ability of dying (“ ” for frontier), that is to increasing frontier rates over time), then the aver- . The difference in those probabilities can be age mortality rate (across years) was also used for all obtained by the following “avoidable mortality”: five-year age groups within the corresponding broad .4 Now take the population age group category (webappendix Table A3). of country C with an age distribution / , where is the size of the age Consistency with all-cause frontiers group and is the oldest age group). In ad- dition, denote the monetary value assigned to the To ensure consistency with all-cause mortality fron- change in the probability of dying ( over one tiers (i.e. the all-cause mortality frontiers of avoid- year) at a given age . Then, for a given year, the aggre- 4  In what follows, we examine differentials over one-year periods, denoted , where is then set to 1 year. COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 4 109 gate monetary value associated with avoidable mortal- We then derived the following value (per year): ity (comparing “C” and “ ”) will be given by: , aggregat- ed per region R: .  [2] [3] We then assign a value to that depends on the extent of the differential in mortality risk and on the VSL in country C ( ). We apply From [3], we obtained an economic value directly the methods developed by Chang et al. 2024, where comparable to annual incomes, which captures the increases with and its marginal rate of increase percent of income an individual would be willing to decreases with (consistent with Cardoso & Dahis forgo to live one year under the lowest possible mor- (2020) and Hammitt (2020)). The value of can be tality rate for a given cause of death (in the region R). related to the VSL in the United States ( ) using an income elasticity : , Sensitivity analyses where YC and YUS are the GNI per capita of C and the United States, respectively. First, we estimated the mortality frontiers in using the To derive , we set = 0.8 for countries with lowest cause-specific mortality rates, in place of the 10th GNI per capita greater than the United States and = percentile of the mortality rates. Second, when quan- 1.2 for countries with lower GNI per capita (Viscusi tifying the economic values associated with avoidable & Materman 2017; Robinson, Hammitt, & O’Keefe mortality, the resulting estimates are sensitive to the VSL 2019)5; we also set a floor constraint for the ratio be- values assigned. Though this will not necessarily affect tween VSL and GNI per capita of 20. GNI per capita the relative distribution of the cause-specific allocations was expressed in current international dollars (Pur- and associated economic values and we report eco- chasing Power Parity (PPP); World Bank 2022) for nomic values in terms of multiples of annual income, the period 2000-2021; and were projected to 2050 we applied three sets of standard sensitivity analyses re- using OECD’s projected country-specific growth lated to VSL estimates (Robinson et al. 2019), in varying rates over 2021-2050 for listed countries, and the (i) income elasticities, to either 1.0 or 1.5; (ii) discount world average growth rate during the same time pe- rates, to either 1 or 5 percent per year; and using (iii) an riod for all other countries. We set = 160 x alternative baseline VSL-to-income ratio of 100 relative US GNI per capita (Robinson, Hammitt, & O’Keeffe to the average income among OECD countries (in place 2019); and a discount rate of 3 percent per year. of the VSL-to-income ratio for the United States). Therefore, per year and country, we comput- All computations and simulations were con- ed the following proportions per cause K and age ducted using R software (version 2022.02.3). All group : = / . codes are available on GitHub. RESULTS We first report on the estimation of the mortality frontiers per cause of death and display the mortality trajec- tories and avoidable mortality with respect to the frontiers for the six regions. Second, we exhibit the economic values associated with avoidable mortality for each cause of death and region. Mortality frontiers and region-specific mortality trajectories following broad causes of death: all causes; commu- nicable, maternal, perinatal and nutritional condi- The estimated mortality frontiers (for 2000-2050) tions; CVD; malignant neoplasms (cancers); and are displayed for ages 20 years and older (the prin- injuries (Figure 4.1).6 As expected, the mortali- cipal age groups over which NCDs manifest) for the ty frontiers are higher (corresponding to higher 5  Alternative values of income elasticity of 1.0 and 1.5 are tested in sensitivity analyses. 6  Note that some of the slight fluctuations observed for the frontiers are due to the scaling with the all-cause frontiers estimated by Chang et al. (2024). Likewise, for the mortality trajectories, the fluctuations observed are due to the method used for calculating regional estimates (i.e., aggregation of population-weighted country-specific estimates) as well as the scaling with the all-cause frontiers. 110 CHAPTER 4 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E mortality rates) for older age groups. All estimated older (Figure 4.1). Likewise, as expected, the trajec- frontiers for all causes of death are available in the tories show higher mortality rates for the older age webappendix (Figure B1). groups for all causes. In general (to a few rare ex- Mortality trajectories for the six regions were ceptions), the mortality trajectories are also higher estimated by sex for all age groups and causes of for males than for females. All mortality trajectories death (over 2000-2050). As an illustration, we show for regions and causes of death are collected in the the trajectories for the Latin America & Caribbe- webappendix (Figures B2-B8). an region for females and males, ages 20 years and COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 4 111 FIGURE 4.1 Mortality frontiers (females and males) for the period 2000-2050; and mortality trajectories for Latin America and the Caribbean over the period 2000-2050; by sex, age group (20-39 years, 40-59 years, 60-79 years, 80+ years), and cause of death: all causes; communicable, maternal, perinatal and nutritional conditions; cardiovascular diseases; malignant neoplasms; and injuries. Note: Points correspond to WHO’s Global Health Estimates (WHO 2020) for the period 2000-2019; the lines are projections. The impact of COVID-19 is not accounted for in these projections and therefore years 2020-2023 are not displayed. 112 CHAPTER 4 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E Economic value of avoidable mortality (5.2 percent for females, 7.2 percent for males, of For the six regions and calendar years 2000, 2019, and annual income in 2019) and China (3.8 percent and 2050, we display the economic valuations (summed 6.8 percent) (webappendix Table B1). Yet, these eco- across all age groups) tied to specific causes of death nomic values would be quite smaller in Eurasia & the (Figure 4.2 displays level 2 causes; the fractions dis- Mediterranean (2.5 percent and 3.1 percent), India played are percentages of annual income per cause, (1.6 percent and 1.3 percent), Latin America & the region, and year; Table 4.2). Alternative disaggre- Caribbean (2.9 percent and 2.5 percent), and sub-Sa- gations (level 1 and level 3 causes) are shown in haran Africa (2.2 percent and 1.1 percent). webappendix Figures B9 and B10, along with rela- As for injuries, they seem to be more geographi- tive distributions of these economic valuations by cally located. Indeed, injuries display substantial values cause (webappendix Figures B11, B12, and B13). for males in Latin America & the Caribbean (around We observe important variations in the cause val- 8.8 percent of annual income in 2019; compared with uations across regions, as well as some slight differenc- 1.6 percent for females) and are concentrated within es when comparing females and males within the same intentional injuries (around 5.1 percent of male annual region. CVD contributes to a very substantial econom- income in 2019); as well as for males in sub-Saharan ic value; for example, in 2019: around 6.8 percent of an- Africa (around 7.4 percent of annual income in 2019, nual income in Eurasia & the Mediterranean, 7.0 per- 3.2 percent for females) with a concentration within cent in China, 5.2 percent in India, 3.9 percent in Latin unintentional injuries (around 5.1 percent of male an- America & the Caribbean, 3.7 percent in High-income, nual income in 2019) (webappendix Table B1). On the and 2.5 percent in sub-Saharan Africa (Table 4.2). contrary, for the other four regions, the economic val- Furthermore, cancers (i.e., malignant neo- ues for injuries were lower: between 4.1 and 4.8 percent plasms) also largely participate to the economic value of annual income for males, between 1.3 and 2.6 per- of avoidable mortality, especially so in High-income cent of annual income for females (in the year 2019). TABLE 4.2 Economic value of avoidable mortality as a percentage of annual income by region and cause of death, for the years 2000, 2019, and 2030 Cause of death 2000 2019 2050 China Communicable, maternal, perinatal and nutritional conditions 2.5% 0.9% 0.2% Infectious and parasitic diseases 1.8% 0.7% 0.1% Maternal and neonatal conditions 0.6% 0.3% 0.1% Nutritional deficiencies 0.0% 0.0% 0.1% Noncommunicable diseases 10.5% 14.9% 23.6% Cardiovascular diseases 3.7% 7.0% 12.4% Ischaemic heart disease 0.6% 2.5% 6.2% Stroke 2.7% 3.9% 4.4% Other cardiovascular diseases 0.4% 0.4% 1.7% Diabetes mellitus 0.1% 0.3% 0.6% Digestive diseases 0.5% 0.4% 0.1% Cirrhosis of the liver 0.4% 0.3% 0.1% Other digestive diseases 0.1% 0.1% 0.0% Malignant neoplasms 3.5% 5.3% 8.2% Breast cancer 0.0% 0.1% 0.3% Cervix uteri cancer 0.1% 0.2% 0.3% Liver cancer 0.8% 0.5% 0.8% Mouth and oropharynx cancers 0.1% 0.1% 0.1% Oesophagus cancer 0.5% 0.6% 0.4% Stomach cancer 0.8% 1.0% 0.5% Trachea, bronchus, lung cancers 0.9% 1.9% 3.6% Other malignant neoplasms 0.6% 1.0% 2.2% Respiratory diseases 1.9% 1.4% 0.8% Chronic obstructive pulmonary disease 1.8% 1.4% 0.9% Other respiratory diseases 0.0% 0.0% 0.0% Other noncommunicable diseases 0.7% 0.6% 1.4% COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 4 113 Cause of death 2000 2019 2050 Injuries 4.2% 2.9% 1.7% Intentional injuries 0.9% 0.4% 0.1% Unintentional injuries 3.3% 2.5% 1.6% Road injury 1.5% 1.3% 0.6% Other unintentional injuries 1.8% 1.2% 0.9% Eurasia & Mediterranean Communicable, maternal, perinatal and nutritional conditions 9.1% 5.2% 2.8% Infectious and parasitic diseases 7.3% 3.9% 2.0% Maternal and neonatal conditions 1.5% 1.1% 0.6% Nutritional deficiencies 0.2% 0.1% 0.1% Noncommunicable diseases 12.8% 13.7% 17.8% Cardiovascular diseases 6.4% 6.8% 8.7% Ischaemic heart disease 3.3% 3.5% 4.8% Stroke 2.3% 2.4% 2.8% Other cardiovascular diseases 0.8% 0.8% 1.0% Diabetes mellitus 0.5% 0.8% 1.3% Digestive diseases 1.3% 1.3% 1.3% Cirrhosis of the liver 0.9% 0.9% 0.9% Other digestive diseases 0.3% 0.3% 0.4% Malignant neoplasms 2.4% 2.8% 3.6% Breast cancer 0.2% 0.3% 0.3% Cervix uteri cancer 0.2% 0.2% 0.1% Liver cancer 0.2% 0.3% 0.3% Mouth and oropharynx cancers 0.2% 0.2% 0.3% Oesophagus cancer 0.1% 0.1% 0.2% Stomach cancer 0.2% 0.2% 0.2% Trachea, bronchus, lung cancers 0.4% 0.5% 0.6% Other malignant neoplasms 0.9% 1.1% 1.7% Respiratory diseases 0.8% 0.8% 1.0% Chronic obstructive pulmonary disease 0.5% 0.5% 0.7% Other respiratory diseases 0.3% 0.3% 0.3% Other noncommunicable diseases 1.5% 1.4% 1.9% Injuries 4.7% 3.0% 2.2% Intentional injuries 1.5% 0.9% 0.7% Unintentional injuries 3.1% 2.2% 1.5% Road injury 1.2% 1.1% 0.8% Other unintentional injuries 1.9% 1.0% 0.7% High-income Communicable, maternal, perinatal and nutritional conditions 1.6% 1.6% 1.8% Infectious and parasitic diseases 1.2% 1.2% 1.4% Maternal and neonatal conditions 0.3% 0.2% 0.1% Nutritional deficiencies 0.1% 0.1% 0.1% Noncommunicable diseases 14.1% 16.2% 19.2% Cardiovascular diseases 4.1% 3.7% 4.3% Ischaemic heart disease 2.8% 2.2% 2.2% Stroke 0.6% 0.5% 0.7% Other cardiovascular diseases 0.8% 1.1% 1.7% Diabetes mellitus 0.4% 0.4% 0.5% 114 CHAPTER 4 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E Cause of death 2000 2019 2050 Digestive diseases 0.9% 1.0% 1.1% Cirrhosis of the liver 0.5% 0.5% 0.5% Other digestive diseases 0.3% 0.4% 0.6% Malignant neoplasms 6.2% 6.2% 5.8% Breast cancer 0.6% 0.5% 0.4% Cervix uteri cancer 0.1% 0.1% 0.0% Liver cancer 0.3% 0.3% 0.3% Mouth and oropharynx cancers 0.2% 0.2% 0.2% Oesophagus cancer 0.2% 0.3% 0.2% Stomach cancer 0.4% 0.3% 0.1% Trachea, bronchus, lung cancers 2.1% 2.2% 1.5% Other malignant neoplasms 2.6% 2.7% 3.1% Respiratory diseases 0.9% 1.5% 2.3% Chronic obstructive pulmonary disease 0.6% 0.9% 1.4% Other respiratory diseases 0.3% 0.5% 0.8% Other noncommunicable diseases 1.8% 3.7% 5.5% Injuries 3.4% 2.9% 2.5% Intentional injuries 1.3% 1.4% 1.0% Unintentional injuries 1.9% 1.4% 1.4% Road injury 1.1% 0.6% 0.5% Other unintentional injuries 0.8% 0.8% 0.9% India Communicable, maternal, perinatal and nutritional conditions 16.4% 7.8% 3.3% Infectious and parasitic diseases 13.6% 6.4% 2.7% Maternal and neonatal conditions 2.3% 1.2% 0.6% Nutritional deficiencies 0.4% 0.1% 0.0% Noncommunicable diseases 8.2% 11.6% 18.9% Cardiovascular diseases 3.3% 5.2% 8.2% Ischaemic heart disease 1.8% 3.2% 5.3% Stroke 0.9% 1.3% 2.0% Other cardiovascular diseases 0.6% 0.6% 0.8% Diabetes mellitus 0.3% 0.6% 1.2% Digestive diseases 1.4% 1.4% 1.2% Cirrhosis of the liver 0.8% 0.9% 0.8% Other digestive diseases 0.6% 0.4% 0.3% Malignant neoplasms 0.7% 1.4% 3.4% Breast cancer 0.0% 0.2% 0.3% Cervix uteri cancer 0.1% 0.2% 0.2% Liver cancer 0.0% 0.0% 0.1% Mouth and oropharynx cancers 0.2% 0.4% 0.8% Oesophagus cancer 0.1% 0.1% 0.2% Stomach cancer 0.1% 0.2% 0.2% Trachea, bronchus, lung cancers 0.1% 0.2% 0.4% Other malignant neoplasms 0.2% 0.4% 1.4% Respiratory diseases 1.6% 2.0% 3.2% Chronic obstructive pulmonary disease 1.0% 1.4% 2.6% Other respiratory diseases 0.5% 0.6% 0.6% Other noncommunicable diseases 1.1% 1.1% 1.9% Injuries 4.2% 3.6% 3.3% Intentional injuries 1.3% 1.1% 0.9% COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 4 115 Cause of death 2000 2019 2050 Unintentional injuries 2.8% 2.5% 2.4% Road injury 0.8% 0.9% 0.8% Other unintentional injuries 2.0% 1.6% 1.5% Latin America & Caribbean Communicable, maternal, perinatal and nutritional conditions 6.0% 3.6% 3.2% Infectious and parasitic diseases 4.2% 2.6% 2.6% Maternal and neonatal conditions 1.3% 0.8% 0.3% Nutritional deficiencies 0.5% 0.2% 0.2% Noncommunicable diseases 11.1% 12.8% 16.9% Cardiovascular diseases 3.3% 3.9% 5.1% Ischaemic heart disease 1.4% 1.9% 2.9% Stroke 1.2% 1.1% 1.0% Other cardiovascular diseases 0.7% 0.8% 1.1% Diabetes mellitus 0.9% 1.3% 1.9% Digestive diseases 1.5% 1.6% 1.9% Cirrhosis of the liver 0.8% 0.7% 0.8% Other digestive diseases 0.7% 0.9% 1.2% Malignant neoplasms 2.3% 2.7% 3.8% Breast cancer 0.2% 0.3% 0.4% Cervix uteri cancer 0.3% 0.3% 0.2% Liver cancer 0.1% 0.2% 0.2% Mouth and oropharynx cancers 0.1% 0.1% 0.1% Oesophagus cancer 0.1% 0.1% 0.1% Stomach cancer 0.3% 0.3% 0.3% Trachea, bronchus, lung cancers 0.3% 0.4% 0.5% Other malignant neoplasms 1.1% 1.3% 2.2% Respiratory diseases 1.0% 0.9% 1.0% Chronic obstructive pulmonary disease 0.5% 0.5% 0.7% Other respiratory diseases 0.5% 0.4% 0.4% Other noncommunicable diseases 2.0% 2.5% 3.2% Injuries 6.7% 5.1% 2.8% Intentional injuries 3.4% 2.9% 1.6% Unintentional injuries 3.3% 2.3% 1.2% Road injury 1.6% 1.3% 0.6% Other unintentional injuries 1.7% 1.0% 0.6% Sub-Saharan Africa Communicable, maternal, perinatal and nutritional conditions 33.5% 20.0% 9.5% Infectious and parasitic diseases 29.0% 16.6% 7.2% Maternal and neonatal conditions 3.6% 3.0% 1.8% Nutritional deficiencies 0.9% 0.4% 0.4% Noncommunicable diseases 7.5% 8.5% 11.4% Cardiovascular diseases 2.4% 2.5% 3.0% Ischaemic heart disease 0.7% 0.8% 1.2% Stroke 1.0% 1.0% 1.1% Other cardiovascular diseases 0.7% 0.6% 0.7% Diabetes mellitus 0.4% 0.5% 0.7% Digestive diseases 1.4% 1.4% 1.4% Cirrhosis of the liver 0.8% 0.8% 0.8% Other digestive diseases 0.5% 0.6% 0.7% 116 CHAPTER 4 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E Cause of death 2000 2019 2050 Malignant neoplasms 1.0% 1.7% 3.3% Breast cancer 0.1% 0.3% 0.5% Cervix uteri cancer 0.3% 0.3% 0.5% Liver cancer 0.1% 0.1% 0.2% Mouth and oropharynx cancers 0.0% 0.1% 0.1% Oesophagus cancer 0.1% 0.1% 0.1% Stomach cancer 0.0% 0.1% 0.1% Trachea, bronchus, lung cancers 0.0% 0.0% 0.1% Other malignant neoplasms 0.4% 0.8% 1.6% Respiratory diseases 0.6% 0.5% 0.6% Chronic obstructive pulmonary disease 0.2% 0.2% 0.3% Other respiratory diseases 0.4% 0.3% 0.4% Other noncommunicable diseases 1.7% 2.0% 2.4% Injuries 6.0% 5.3% 5.7% Intentional injuries 2.4% 1.5% 1.7% Unintentional injuries 3.6% 3.8% 4.0% Road injury 1.8% 2.1% 2.5% Other unintentional injuries 1.7% 1.6% 1.5% World Communicable, maternal, perinatal and nutritional conditions 10.1% 6.3% 3.9% Infectious and parasitic diseases 8.4% 5.1% 3.0% Maternal and neonatal conditions 1.5% 1.1% 0.7% Nutritional deficiencies 0.3% 0.1% 0.2% Noncommunicable diseases 11.0% 13.1% 17.5% Cardiovascular diseases 4.2% 5.2% 7.0% Ischaemic heart disease 1.9% 2.6% 3.8% Stroke 1.6% 1.9% 2.1% Other cardiovascular diseases 0.7% 0.7% 1.1% Diabetes mellitus 0.4% 0.6% 1.0% Digestive diseases 1.1% 1.1% 1.2% Cirrhosis of the liver 0.7% 0.7% 0.7% Other digestive diseases 0.4% 0.4% 0.5% Malignant neoplasms 2.8% 3.4% 4.4% Breast cancer 0.2% 0.2% 0.4% Cervix uteri cancer 0.1% 0.2% 0.2% Liver cancer 0.3% 0.3% 0.3% Mouth and oropharynx cancers 0.2% 0.2% 0.3% Oesophagus cancer 0.2% 0.2% 0.2% Stomach cancer 0.4% 0.4% 0.2% Trachea, bronchus, lung cancers 0.7% 0.9% 1.0% Other malignant neoplasms 1.0% 1.2% 1.9% Respiratory diseases 1.2% 1.2% 1.5% Chronic obstructive pulmonary disease 0.8% 0.9% 1.1% Other respiratory diseases 0.3% 0.3% 0.4% Other noncommunicable diseases 1.4% 1.7% 2.5% Injuries 4.6% 3.6% 3.2% Intentional injuries 1.6% 1.1% 1.0% Unintentional injuries 3.0% 2.4% 2.2% Road injury 1.3% 1.2% 1.1% Other unintentional injuries 1.7% 1.2% 1.1% COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 4 117 FIGURE 4.2 Economic values of avoidable mortality, measured as proportion of annual income, assigned to selected causes of death for the six regions, in the years 2000, 2019, and 2050 We note that the economic value attributable males) and China (0.8 percent for females, 1.1 per- to the cause category “communicable, maternal, cent for males), geographical locations that are most perinatal and nutritional conditions” would be very advanced into the epidemiological transition from substantial in sub-Saharan Africa (20.4 percent for communicable to non-communicable diseases. females, 19.6 percent for males, of annual income in Lastly, in almost all regions, we could point to the 2019), and quite large, yet to a lesser extent in In- rapid expansion of avoidable deaths associated with dia (8.1 percent and 7.5 percent). This is notably due NCDs, in particular CVD and cancers, and to the fast to large young populations, especially in sub-Saha- compression of avoidable mortality from communi- ran Africa (UN 2022), as well as to high mortality cable, perinatal, maternal and nutritional diseases, rates for these causes of death. The economic values into the future (from 2000 to 2019 and 2050; Figure for these causes of death would however be much 4.3). This is a testimony of the still advancing epidemi- smaller in the other four regions, particularly so in ological transition from communicable to non-com- high-income (1.5 percent for females, 1.8 percent for municable diseases and of the ageing of populations. 118 CHAPTER 4 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E FIGURE 4.3 Rates of change (percent per year) of the economic values of avoidable mortality assigned to selected causes of death (communicable, maternal, perinatal and nutritional conditions; cardiovascular diseases; malignant neoplasms; injuries) for the six regions, over the time periods 2000-2019 and 2019-2050. COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 4 119 Specifically, from 2000 to 2050, we observe steady de- expansions would be smaller, especially so in High-in- clines in the economic value associated with avoidable come, which displays a rather stagnant picture. Over- mortality from communicable, perinatal, maternal all, this dynamic evolution into the future crucially and nutritional diseases coupled with steady expan- emphasizes the paramount challenges and economic sions of the economic value of avoidable mortality losses ahead tied to the expansion of CVD and can- from CVD (except for sub-Saharan Africa) and can- cers, replacing communicable, perinatal, maternal cers in China, India, and sub-Saharan Africa. For the and nutritional diseases, particularly in the emerging other regions, the extents of these CVD and cancer economies of China, India, and sub-Saharan Africa. Sensitivity analyses When using the lowest mortality rates (in place of decrease (increase) with a higher (lower) income the 10th percentile of the mortality rates; webap- elasticity; or varied discount rates, to 1 and 5 per- pendix Figure C1) or when applying standard BCA cent per year (webappendix Figures C5, C6), where sensitivity analyses (Robinson et al. 2019), we did we observed modest changes; or we set an alterna- not observe qualitatively distinct findings, though tive baseline VSL-to-income ratio of 100 relative quantitatively the findings would be impacted. to OECD countries (instead of 160 relative to the This was the case, whether we prescribed differ- United States; webappendix Figure C4), where eco- ent income elasticities, of 1.0 and 1.5 (webappen- nomic values would decrease. dix Figures C2, C3), where economic values would DISCUSSION We described in this chapter analytical methods to compute the monetary values linked with mortality risk re- ductions associated with specific causes of death, across six regions. The economic framework presented provides a systematic monetary assessment of the burden of avoidable mortality, which can be useful in quantifying the costs and benefits of both health and non-health interventions, and allow comparisons within the health sector as well as beyond across sectors (e.g., health, education, transport, agriculture). Our assessment shows that the implications of con- tiveness of interventions (say as potential “best-buys”). trolling NCDs and injuries could be substantial in We also estimated sizeable monetary values for the terms of economic value and that the costs of inac- mortality reduction gains that could be obtained with tion would be enormous, especially for CVD. Indeed, scaling up key interventions and benefit packages ad- we found that the economic value associated with dressing NCDs and injuries (e.g., health taxes, primary avoidable CVD mortality would be universally large, and secondary prevention of CVD; NCD Countdown in the order of 3 to 7 percentage points of annual in- 2030 collaborators 2022). In sum, our dual focus on come in the year 2019 for example. We also observed mortality frontiers and modulated monetary valua- that the economic value of avoidable cancer mortal- tion of the size of mortality risk reductions provides ity would be large in high-Income and China; and, a priority-setting dimension complementary to cause- for injury, it would be substantial for males in Latin of-death distributions (webappendix Figure D1). America & the Caribbean (intentional injuries) and Relatedly, a few similar studies (in spirit) have males in sub-Saharan Africa (unintentional injuries). been conducted in the literature. Most importantly, While the extents of these hypothetical mortal- Bloom and colleagues have pioneered comparable ity reductions needed not be monetized, important- BCA methods and published seminal works as part ly, in computing these economic values, we assigned of making the investment case for NCDs (Bloom et al. willingness to pay values for cause-specific mortality 2011). They have also led macroeconomic modeling reductions and provided a metric (i.e., percent of an- to assess the economic losses attributable to NCDs nual income) that is directly comparable to GNIs and in a few Asian nations including China, Japan, and annual budgetary allocations. As such, we produced South Korea (Bloom et al. 2020) and to cancers glob- potential willingness to pay estimates (e.g., multipliers ally (Chen et al. 2023). These works, like ours, demon- of GNIs per capita; Bertram et al. 2016, Woods et al. strated the very substantial welfare impact of NCDs, 2016) against which one could gauge the cost-effec- at the global and regional levels. Also, Khadka & 120 CHAPTER 4 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E Verguet (2021) further developed VSL approaches to hearing impairments). We likely underestimate the quantify the monetary values associated with NCDs economic value of non-fatal conditions by looking in a systematic manner, and like our assessment, es- only at avoidable mortality. timated that the welfare losses associated with NCDs, Third, there are major limitations with the im- particularly for CVD and cancers, would be very sub- plementation of VSL approaches for economic val- stantial in the coming decades for LMIC economies. uation, in particular the high sensitivity of the VSL Similarly, Arias and colleagues (2022) pointed to the estimates to income elasticity parameters (Viscusi potentially large economic losses tied to mental disor- & Aldy 2003; Robinson, Hammitt & O’Keeffe 2019; ders using somewhat comparable approaches. webappendix Figures C2, C3). Importantly, we have Nevertheless, our analysis presents import- made a number of argued prescriptions regarding ant limitations. First, we used global estimates of the parametrization of our estimations, including the cause-specific mortality, which are based on incom- specification of a modulating function varying with plete and uncertain data, particularly in low- and the extent of mortality risk reductions (using Chang lower-middle income countries (WHO 2020a). et al. (2024), consistent with Hammitt (2020) and Second, we forecasted future trends (over the Cardoso & Dahis (2020)). In this respect, caution period 2020-2050) in age-cause-sex-specific frontier should be exercised when comparing the economic mortality rates and country groupings based on pre- values associated with avoidable mortality across re- vious trends (over the period 2010-2019) using sim- gions. Furthermore, to many, VSL approaches raise ple log-linear extrapolations into the future. A more ethical dilemmas, that is, notably, to eventually yield complex modeling approach could be pursued, yield- a monetary value for life years. Yet, often, VSL esti- ing probabilistic distributions of a space of plausible mates are misunderstood as economic values one so- future outcomes (Kontis et al. 2017). Yet, this would ciety assigns to one life; while they rather capture the likely bring additional complications for interpre- trade-offs individuals are willing to make in giving tation without, however, being further grounded up a portion of their income in exchange for small within empirical data. Also, we did not consider reductions in their mortality risk over a given year the short- or long-term impacts of the COVID-19 (Robinson et al. 2019). Also, BCA is only one (among pandemic on age-cause-specific mortality rates in many other possible) method for priority setting, our forecasts. In addition, for the estimation of the nonetheless attractive for making multisectoral com- mortality frontiers, we mostly used high-income parisons. Critically, the quantitative values extracted country data, because high-quality vital registration from VSL estimates are sensitive to income elastic- data for LMICs outside of the Latin America & Ca- ities and the income differentials across countries. ribbean region remain scarce and because high-in- Fourth, while we were able to compute monetary come countries would likely achieve the lowest mor- values associated with longevity losses due to specific tality rates (WHO 2020a; IHME 2022). Such frontier causes of death, this only provides one element (i.e., a trajectories could however be revisited as more data valuation of health gains) for use in BCAs. Evidently, become available, and alternative mortality frontiers we would also need the costing of interventions and could be constructed (e.g., region-specific mortality the valuation of morbidity reduction and other non- frontiers in place of global mortality frontiers). Like- health benefits (e.g., financial protection) (Verguet, wise, different percentiles (5th, 20th, 30th or lowest) Laxminaryan & Jamison 2015; Verguet & Norheim of mortality rates could be tested in place of the 10th 2022), so as to fully assess the returns on investment percentiles used here. The changes might quantita- from controlling NCDs and injuries. tively affect our findings, yet only likely to a modest In summary, we produced a systematic econom- level if, for example, we used the lowest mortality ic valuation of the burden of avoidable mortality by rates in place of the 10th percentiles (webappendix cause of death, with a focus on major NCDs and in- Figures A1, C1). Relatedly, differences in cause-of- juries, in LMICs. This economic framework can point death patterns across geographies and over time may to the very large economic value linked to NCDs relate to quality and practices in cause-of-death as- and injuries, amounting to several percentage points signments as opposed to real epidemiological differ- of annual incomes. It can also be further adapted to ences, especially for related causes (e.g., diabetes and estimate the economic returns from investing in the CVD). Our methods are based on quantification of health sector and, in this particular study, toward the economic value of avoidable mortality, which by NCD and injury interventions. In doing so, it allows design excludes non-fatal conditions, such as mental both economic comparisons of such investments both conditions and sensory disorders (e.g., vision and within the health sector as well as across other sectors. COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 4 121 REFERENCES 1. Arias D, Saxena S, Verguet S. Quantifying the global burden of 17. Kontis V, Bennett JE, Mathers CD, et al. Future life expectancy in mental disorders and their economic value. eClinical Medicine 35 industrialised countries: projections with a Bayesian model 2022; 54:101675. ensemble. Lancet 2017; 389:1323-1335. 2. Baltussen R,  Marsh K, Thakola P,  et al. Multicriteria decision 18. Mathers CD, Loncar D. 2006. Projections of Global Mortality and analysis to support health technology assessment agen- Burden of Disease from 2002 to 2030. PLoS Medicine 3 (11):e442. cies: benefits, limitations, and the way forward Value Health 19. Mathers CD, Stevens GA, Boerma T, White RA, Tobias MI. Causes 2019; 22(11):1283-1288. of international increases in older age life expectancy. Lancet 3. Bertram MY, Lauer JA, De Joncheere K, Edejer T, Hutubessy R, 2015; 385(9967):540-548. Kieny M-P, Hill SR. Cost-effectiveness thresholds: pros and cons. 20. Murray CJ, Acharya AK. Understanding DALYs (disability-adjust- Bulletin of the World Health Organization 2016; 94(12):925-930. ed life years). Journal of Health Economics 1997; 16(6):703-730. 4. Bloom, D.E., Cafiero, E.T., Jané-Llopis, E, Abrahams-Gessel, S, 21. NCD Countdown 2030 collaborators. NCD Countdown 2030: effi- Bloom, L.R., Fathima, S, et al. The Global Economic Burden of cient pathways and strategic investments to accelerate progress Non-communicable Diseases [Internet]. Geneva: World Eco- towards the Sustainable Development Goal target 3.4 in low- and nomic Forum; Available from: http://www3.weforum.org/ middle-income countries. Lancet 2022; 399:P1266-1278. docs/WEF_Harvard_HE_GlobalEconomicBurdenNonCommu- 22. Neumann PJ, Sanders GD, Russell LB, Siegel JE, Ganiats TG. nicableDiseases_2011.pdf. Cost-Effectiveness in Health and Medicine (2nd ed.). Oxford, 5. Bloom DE, Chen S, Kuhn M, McGovern ME, Oxley L, Prettner K. UK: Oxford University Press; 2015. The economic burden of chronic diseases: Estimates and pro- 23. Robinson LA, Hammitt JK, O’Keeffe L. Valuing mortality risk re- jections for China, Japan, and South Korea. The Journal of the ductions in global benefit-cost analysis. Journal of Benefit Cost Economics of Ageing. 2020 Oct 1;17:100163. Analysis 2019; 10(suppl 1):15-50. 6. Cardoso DS, Dahis R. 2020. Value of a Statistical Life Under Large 24. Robinson LA, Hammitt JK, Cecchini M, Chalkidou K, Claxton K, Mortality Risk Change: Theory and an Application to COVID-19’. Hoang-Vu Eozenou P, de Ferranti D, et al. 2019. Reference Case SSRN Electronic Journal. https://doi.org/10.2139/ssrn.3599529. Guidelines for Benefit-Cost Analysis in Global Health and De- 7. Chang AY, Stevens GA, Cardoso DS, Cao B, Jamison DT. The eco- velopment, May, 126. nomic burden of avoidable mortality, 2000-2050. In Unlocking 25. Partridge L, Deelen J, Slagboom PE. Facing up to the global the Power of Healthy Longevity: Compendium of Research for the challenges of ageing. Nature 2018; 561:45-56. Healthy Longevity Initiative. Washington, DC: World Bank. 2024. 26. Spasova S, Baeten R, Vanhercke B. Challenges in long-term care 8. Chen S, Cao Z, Prettner K, et al. Estimates and projections of the in Europe. Eurohealth Observer 2018; 24(4):7-12. global economic cost of 29 cancers in 204 countries and terri- 27. Tynkkynnen L-K, Pulkki J, Tervonen-Goncalves, et al. Health tories from 2020 to 2050. JAMA Oncology 2023; 9(4):465-472. system reforms and the needs of the ageing population – an 9. Christensen K, Doblhammer G, Rau R, Vaupel JW. Ageing pop- analysis of recent policy paths and reform trends in Finland and ulations: the challenges ahead. Lancet 2009; 374:1196-1208. Sweden. European Journal of Ageing 2022; 19:221-232. 10. Drummond MF, Sculpher MJ, Claxton K, Stoddart GL, Torrance 28. United Nations. Department of Economic and Social Affairs. GW. Methods for the economic evaluation of health care pro- World Population Prospects 2022. New York, NY: United Na- grammes (4th ed.). Oxford, UK: Oxford University Press; 2015. tions, 2022. Available from: https://population.un.org/wpp/ 11. Hammitt JK. 2020. Valuing Mortality Risk in the Time of (accessed April 26, 2022). COVID-19’. Journal of Risk and Uncertainty 61 (2):129-54. 29. Vaupel JW, Villavicencio F, Bergeron-Boucher M-P. Demo- 12. Institute for Health Metrics and Evaluation (IHME). GBD 2019. graphic perspectives on the rise of longevity. PNAS 2021; Seattle: University of Washington, 2022. Available from: https:// 118(9):e2019536118. vizhub.healthdata.org/gbd-compare/. 30. Verguet S, Laxminarayan R, Jamison DT. Universal public fi- 13. Jamison DT, Summers LH, Alleyne G, et al. Global health 2035: a nance for tuberculosis treatment in India: an extended cost-ef- world converging within a generation. Lancet 2013; 382:1898-1955. fectiveness analysis. Health Economics 2015; 24(3):318-332. 14. Jamison DT, Jha P, Laxminarayan R, Ord T. Infectious disease. In: 31. Verguet S, Hailu A, Eregata GT, et al. Toward universal health Global problems, smart solutions: costs and benefits, Lomborg coverage in the post-COVID-19 era. Nature Medicine 2021; B (editor). Cambridge, UK: Cambridge University Press, 2013. 27(3):380-387. 15. Jha P, Hum R, Nugent R, Verguet S, Bloom D. Chronic disease. In: 32. Verguet S, Norheim OF. Estimating and comparing health and Global problems, smart solutions: costs and benefits, Lomborg financial risk protection outcomes in economic evaluations. B (editor). Cambridge, UK: Cambridge University Press, 2013. Value in Health 2022; 25(2):238-246. 16. Khadka A, Verguet S. The economic value of changing mor- 33. Viscusi K, Aldy JE. The value of a statistical life: a critical review tality risk in low- and middle-income countries: a systematic of market estimates throughout the world. Journal of Risk and breakdown by cause of death. BMC Medicine 2021; 19:156. Uncertainty 2003; 27:5-76. 122 CHAPTER 4 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E 34. Viscusi K, Masterman CJ. Income elasticity and the global val- fectiveness thresholds: initial estimates and the need for fur- ue of a statistical life. Journal of Benefit-Cost Analysis 2017; ther research. Value in Health 2016; 19(8):929-935. 8(2):226-250. 39. World Bank. World Development Indicators. Available from: 35. Watkins DA, Yamey G, Schaferhoff M, Adeyi O, Alleyne G, Alwan https://datatopics.worldbank.org/world-development-indica- A, et al. Alma-Ata at 40 years: reflections from the Lancet Com- tors/ (accessed May 13, 2022). mission on Investing in Health. Lancet 2018; 392:P1434-1460. 40. World Health Organization. 2020a. Global Health Estimates. 36. Watkins DA, Qi J, Kawakatsu Y, Pickergill SJ, Horton SE, Jami- 2020. Available from: https://www.who.int/data/glob- son DT. Resource requirements for essential universal health al-health-estimates. coverage: a modelling study based on findings from Dis- 41. Word Health Organization. 2020b. WHO methods and data ease Control Priorities, 3rd edition. Lancet Global Health 2020; sources for country-level causes of death 2000-2019. Avail- 8(6):e829-e839. able from: https://cdn.who.int/media/docs/default-source/ 37. Weinstein MC, Torrance G, McGuire A. QALYs: the basics. Value gho-documents/global-health-estimates/ghe2019_cod_ in Health 2009; 12(2)(suppl 1):S5-S9. methods.pdf?sfvrsn=37bcfacc_5. 38. Woods B, Revill P, Sculpher M, Claxton K. Country-level cost-ef- 5 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 5 123 Rates of progress in mortality decline, 2000-2019 Daphne C. Wu a, Debapriya Chakraborty a, Ryan Hum a, Prabhat Jha a, Dean T. Jamison b a Centre for Global Health Research, University of Toronto, Unity Health Toronto b University of California, San Francisco 124 CHAPTER 5 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E INTRODUCTION Globally, mortality rates have been generally declining over time. The rates of decline, however, vary between ages and countries or country income levels (Norheim and others 2015). Mortality rates among children have declined remarkably since the 1990s, mostly owing to global as well as country-level efforts under the UN Mil- lennium Development Goals (MDGs) (Cha 2017; Lozano and others 2011; Tanner and others 2016). Compared to child mortality, progress in reducing adult mortality has been much slower (Figure 5.1) (Norheim and others 2015; Wang and others 2017). The gap in mortality rates between low- and middle-income countries (LMICs) and high-income countries is converging for deaths at younger ages but diverging for older ages. In 2019, about 80 percent of deaths above 15 years globally are due to non-communicable diseases (NCDs); this proportion has been increasing over time across all ages (World Health Organization (WHO) 2020). Recognizing the increasing disease burden due to NCDs, the Sustainable Development Goals (SDGs), adopted in 2015 by the UN General Assembly, included target 3.4 to reduce mortality due to NCDs by one-third by 2030 (United Nations General Assembly 2015). However, progress has been slow and insufficient to meet the target, even prior to the COVID-19 pandemic (United Nations, Department of Economic and Social Affairs, Sustainable Development 2023). The progress in mortality reduction not only var- income levels, during 2010-09 and compare them ies by age, but also over time (NCD Countdown to the rates during 2000-09. In the next section, 2030 collaborators 2018). As shown in Figure 5.1, we examine countries’ performance in the rate of mortality rates decreased notably faster for ages un- decline in mortality during 2010-19 compared to der 70 during 1990-2010 compared to after 2010, during 2000-09. The following section presents the at least in LMICs. The first section of this study progress by disease groups and age. Finally, the last seeks to examine and quantify the rates of progress section examines the trend in cost of saving a child in mortality reduction, by age group and country vs. an adult life from 2000 to 2019. FIGURE 5.1 Change in risk of death at start of age range, by age group and country income level, 1990-2019 Source: Original calculations for this publication COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 5 125 RATES OF PROGRESS OVER 2000-09 AND 2010-19 As we mentioned earlier, global progress on mortality reduction has been varying with time, country income levels, and age. Here, we use interrupted time series (ITS) analysis to examine and quantify the difference in the rates of progress in mortality reduction, by age and country income levels between 2010-19 and 2000-09. Annex 5.1 provides details of the ITS methodology. Actual mortality rates, by age, sex, and country income levels are presented in Annex Figure 5A.2. Table 5.1 shows the results of the ITS analysis of the rose substantially in middle-income countries in change in mortality rates from 2000-09 to 2010- 2010. For almost all country income levels, the rate 19, by age and sex, and by country income level. of decline in mortality between 2010 and 2019 was Between 2000 and 2010, mortality rates declined significantly slower than during 2000-09 for ages significantly across all age groups irrespectively of under age 70. The slowing of progress in the second country income level, except for males aged 70- decade was seen at all ages above 50, with the excep- 84 years in lower-middle-income countries where tion for ages 70-84 years, where the rate of decline the mortality rates increased but not significantly. was significantly faster during 2010-19 for males Compared to ages 0-14, the mortality rate among in low-income countries and males and females in those 50-69 years fell in low-income countries but middle-income countries. TABLE 5.1  Results from interrupted time series analysis of the change in mortality rates (per 100,000 population) from 2000-10 to 2010-19, by country income, age, and sex Annual rate of change during Level change in Trend change in Number of 2000-09 2010 2010-19 vs 2000-09 Constant observations Under 15 years Low-income countries Males -66.67 (0.000) 30.51 (0.016) 32.00 (0.000) 1671.36 (0.000) 20 Females -63.44 (0.000) 26.79 (0.013) 31.13 (0.000) 1514.51 (0.000) 20 Lower-middle-income countries Males -25.81 (0.000) 0.62 (0.897) 7.50 (0.000) 827.21 (0.000) 20 Females -26.90 (0.000) 2.37 (0.676) 8.33 (0.000) 812.14 (0.000) 20 Upper-middle-income countries Males -10.54 (0.000) -1.10 (0.611) 1.67 (0.001) 286.77 (0.000) 20 Females -8.49 (0.000) 0.19 (0.934) 1.36 (0.010) 233.64 (0.000) 20 High-income-countries Males -1.62 (0.000) -1.94 (0.003) 0.48 (0.000) 70.66 (0.000) 20 Females -1.19 (0.000) -1.47 (0.005) 0.24 (0.008) 57.03 (0.000) 20 15-49 years Low income-countries Males -18.72 (0.000) 9.36 (0.113) 13.09 (0.000) 703.94 (0.000) 20 Females -18.17 (0.000) -11.10 (0.110) 5.32 (0.000) 618.29 (0.000) 20 Lower-middle-income countries Males -3.75 (0.000) -5.14 (0.055) -2.92 (0.000) 377.95 (0.000) 20 Females -6.23 (0.000) -5.04 (0.164) 1.49 (0.012) 307.49 (0.000) 20 Upper-middle-income countries Males -5.06 (0.000) -6.47 (0.053) 0.12 (0.788) 301.19 (0.000) 20 Females -2.23 (0.000) -6.09 (0.007) -0.98 (0.002) 153.82 (0.000) 20 126 CHAPTER 5 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E Annual rate of change during Level change in Trend change in Number of 2000-09 2010 2010-19 vs 2000-09 Constant observations High income-countries Males -3.82 (0.000) -8.28 (0.017) 3.01 (0.000) 186.52 (0.000) 20 Females -1.32 (0.000) -2.77 (0.010) 0.96 (0.000) 89.38 (0.000) 20 50-69 years Low income-countries Males -38.62 (0.000) -54.43 (0.004) 8.06 (0.003) 2659.56 (0.000) 20 Females -38.75 (0.000) -25.33 (0.060) 5.82 (0.016) 2125.66 (0.000) 20 Lower-middle-income countries Males -31.11 (0.000) 14.51 (0.249) 9.13 (0.002) 2204.67 (0.000) 20 Females -30.21 (0.000) 14.49 (0.122) 21.05 (0.000) 1583.90 (0.000) 20 Upper-middle-income countries Males -36.55 (0.000) 29.34 (0.002) 22.77 (0.000) 1768.85 (0.000) 20 Females -26.82 (0.000) 24.03 (0.009) 16.68 (0.000) 992.49 (0.000) 20 High income-countries Males -22.76 (0.000) 5.59 (0.350) 15.72 (0.000) 1236.70 (0.000) 20 Females -10.91 (0.000) 7.88 (0.010) 9.18 (0.000) 641.12 (0.000) 20 70-84 years Low income-countries Males -70.49 (0.000) -149.27 (0.004) -53.94 (0.000) 9818.88 (0.000) 20 Females -104.62 (0.000) 27.86 (0.138) 41.76 (0.000) 7792.70 (0.000) 20 Lower-middle-income countries Males 8.87 (0.106) -182.17 (0.000) -123.27 (0.000) 8116.32 (0.000) 20 Females -25.50 (0.000) -86.54 (0.006) -67.10 (0.000) 6147.70 (0.000) 20 Upper-middle-income countries Males -64.09 (0.000) 22.59 (0.380) -13.57 (0.096) 7137.03 (0.000) 20 Females -64.32 (0.000) -24.61 (0.217) -18.41 (0.001) 4693.27 (0.000) 20 High income-countries Males -93.53 (0.000) -24.93 (0.388) 25.13 (0.000) 5457.30 (0.000) 20 Females -55.57 (0.000) -79.32 (0.112) 17.15 (0.051) 2917.58 (0.000) 20 p-values are given in parenthesis. Figures in bold indicate significance at 95 percent confidence interval. The “annual rate of change during 2000-09” refers to the annual rate of change in mortality rates (per 100,000 population) during 2000 to 2009, “level change in 2010” refers to the change (increase or decrease) in mortality rate in 2010, and “trend change in 2010-19 vs 2000-09” refers to change in the mortality rate during 2010-19 from that during 2000-09. A positive trend change indicates a slower decline in mortality during 2010-19 compared to during 2000-09, while a negative trend change indicates a more rapid decline during 2010-19 than during 2000-09. COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 5 127 COUNTRIES’ RATES OF PROGRESS DURING 2000-09 AND 2010-19 In this section, we examine performance, at the country level, in the rate of decline in mortality between 2000 and 2019. We used the framework of Verguet and Jamison (2014) to estimate performance in the rate of decline in under-five mortality in LMICs (Verguet and Jamison 2014). We define performance as the difference in the rate of decline in actual mortality compared to that predicted by a set of contextual factors. The analysis includes 71 LMICs that have a population of more than 7 million, which covers about 83 percent of the world population in 2019 (United Nations, Department of Economic and Social Affairs, Population Division 2022). High-income countries (HICs) were excluded, as the mathematical relationship between mortality rates and social, economic, and demographic characteristics markedly differs between HICs and LMICs. Details of the methodology used in provided in Annex 5.2. Table 5.2 presents the top 10 and worst 10 LMICs in period. A notable number of lower income African the performance in the rate of decline in mortality countries, including Tanzania and Zambia, were during 2010-19 and their ranking during 2000-09, among the top 10 performers for mortality decline by age and sex. Ranking of all 71 LMICs are provid- at ages 50-69 in both males and females. Among the ed in Annex Table 5A.2. We note that, across all ages, worst performers at these ages were a number of Haiti emerges as one of the top performers during middle-income countries in East Asia (Viet Nam 2010-19 but as one of the worst performers during and Malaysia) and Eastern Europe (Bulgaria and 2000-10. This is due to the earthquake in 2010 that Uzbekistan) along with Kenya. In terms of gender killed hundreds of thousands in Haiti. Similar situ- differences in country performance, for both males ations of natural disasters and conflicts in Mozam- and females, the best and worst performer countries bique (flooding in 2000), and Afghanistan and Iraq correlate rather well for under 15 years (R2=0.94), (war during the 2000s) make them outliers in this but less at older ages (Annex Figure 5A.3). analysis. The crisis countries are shown in the ta- We also examined performance transitions bles in italics. Excluding the crisis countries, across from 2000-09 to 2010-19. Performance transitions most, if not, all age groups, among the top perform- for the 71 LMICs and the distribution are given in ers were Rwanda and Cambodia during 2000-09, Annex Table 5A.3. For ages 50-69 years, 64 out of Russia during 2010-19, while among the worst was the 71 LMICs (or 90 percent) fall in the moderate Zimbabwe during 2000-09, and Tunisia during transition categories of +2.4 percent to -2.5 percent 2010-19. In the 50-69 age group, among males, mortality decline per year. For ages under 70 years, South Africa had the best performance during Rwanda and Cambodia presented some of the 2010-19, even after considering the substantial re- poorest transitions of -2.5 percent to -5 percent per duction in HIV infection that occurred during this year. TABLE 5.2  Top 10 and worst 10 performance rankings among 71 LMICs in the rate of decline of mortality, by sex and age: 2000-10 and 2010-19 Country Mortality rate in Performance ranking Country Mortality rate in Performance ranking 2019 (per 100,000 (performance*) in 2019 (per 100,000 (performance*) in population) population) 2010-19 2000-10 2010-19 2000-10 Under 15 males Under 15 females Top 10 performance countries during 2010-19 Top 10 performance countries during 2010-19 Russian Federation 46 1 (5.6) 33 (-0.4) Haiti 474 1 (7.6) 71 (-10.2) Haiti 610 2 (5.6) 71 (-8.8) Russian Federation 37 2 (5.2) 52 (-1.8) Azerbaijan 157 3 (5.1) 63 (-2.3) Azerbaijan 121 3 (5.2) 64 (-2.6) Belarus 27 4 (4.6) 16 (0.7) Kazakhstan 74 4 (4.7) 51 (-1.8) Kazakhstan 105 5 (4.6) 42 (-0.9) Belarus 20 5 (4.5) 34 (-0.4) China 71 6 (3.8) 25 (0.4) China 61 6 (3.8) 24 (0.5) Ukraine 55 7 (3.8) 64 (-2.4) Ukraine 43 7 (3.8) 69 (-3.9) Bulgaria 55 8 (3.5) 67 (-3) Burundi 576 8 (3.4) 48 (-1.5) Malawi 392 9 (3.2) 8 (2.1) Malawi 337 9 (2.6) 7 (2.5) 128 CHAPTER 5 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E Country Mortality rate in Performance ranking Country Mortality rate in Performance ranking 2019 (per 100,000 (performance*) in 2019 (per 100,000 (performance*) in population) population) 2010-19 2000-10 2010-19 2000-10 Under 15 males (continued) Under 15 females (continued) Worst 10 performance countries during 2010-19 Worst 10 performance countries during 2010-19 Burundi 685 10 (2.8) 50 (-1.6) South Africa 271 10 (2.4) 63 (-2.6) Malaysia 72 71 (-3.6) 31 (-0.3) Malaysia 61 71 (-4.4) 31 (-0.2) Viet Nam 245 70 (-3.1) 68 (-3.1) Viet Nam 144 70 (-3.4) 67 (-3.2) Dominican Re- 272 69 (-2.7) 61 (-2.1) Dominican Republic 227 69 (-2.9) 65 (-2.8) public Guinea 1054 68 (-2.2) 43 (-1) Guinea 954 68 (-2.3) 36 (-0.6) Philippines 247 67 (-2) 51 (-1.6) Nigeria 1351 67 (-2.1) 57 (-2) Nigeria 1377 66 (-1.8) 54 (-1.7) Philippines 194 66 (-1.8) 55 (-1.9) Cote d’Ivoire 822 65 (-1.8) 40 (-0.9) Tunisia 125 65 (-1.7) 59 (-2.1) Benin 1005 64 (-1.7) 55 (-1.7) Cote d’Ivoire 659 64 (-1.7) 38 (-0.7) Myanmar 438 63 (-1.6) 48 (-1.5) Kenya 347 63 (-1.5) 23 (0.5) Tunisia 152 62 (-1.6) 57 (-2) Benin 889 62 (-1.5) 53 (-1.8) 15-49 males 15-49 females Top 10 performance countries during 2010-19 Top 10 performance countries during 2010-19 Haiti 423 1 (7) 71 (-9.9) Haiti 246 1 (11.4) 71 Iraq 193 2 (4.6) 70 (-3.2) South Africa 396 2 (6.9) 19 South Africa 575 3 (4.5) 13 (1.2) Mozambique 311 3 (4.7) 65 Tanzania 320 4 (3.6) 45 (-0.9) Tanzania 239 4 (3.1) 29 Mozambique 522 5 (2.2) 62 (-2) Ukraine 85 5 (2.9) 13 Zimbabwe 632 6 (2.2) 65 (-2.2) Zambia 310 6 (2.7) 7 Serbia 121 7 (2.1) 4 (1.6) Zimbabwe 478 7 (2.6) 70 Belarus 301 8 (2.1) 27 (0.1) Malawi 313 8 (2.3) 36 Russian Federation 384 9 (2) 9 (1.4) Iran 86 9 (2.3) 51 Sierra Leone 367 10 (1.7) 1 (2.8) Sierra Leone 318 10 (2.2) 2 Worst 10 performance countries during 2010-19 Worst 10 performance countries during 2010-19 Afghanistan 530 71 (-3.5) 12 (1.2) Kenya 443 71 (-3.3) 53 (-1.7) Mexico 340 70 (-2.9) 68 (-2.7) Viet Nam 112 70 (-2.3) 60 (-2.1) Kenya 592 69 (-2) 60 (-1.9) Mexico 121 69 (-2.1) 57 (-1.9) Viet Nam 314 68 (-2) 61 (-1.9) Cambodia 195 68 (-1.9) 3 (1.9) Bulgaria 241 67 (-1.9) 41 (-0.7) Nigeria 723 67 (-1.8) 63 (-2.7) Burkina Faso 489 66 (-1.7) 40 (-0.6) Cote d’Ivoire 525 66 (-1.7) 68 (-3.6) Tunisia 175 65 (-1.7) 63 (-2) Mali 407 65 (-1.7) 32 (-0.3) Ghana 471 64 (-1.7) 64 (-2.1) Benin 415 64 (-1.5) 66 (-3.2) Mali 450 63 (-1.4) 43 (-0.7) Bolivia 185 63 (-1.5) 31 (-0.1) Indonesia 253 62 (-1.3) 57 (-1.6) Dominican Republic 136 62 (-1.5) 18 (0.7) 50-69 males 50-69 females Top 10 performance countries during 2010-19 Top 10 performance countries during 2010-19 Haiti 2271 1 (4) 71 (-4.9) Haiti 1473 1 (5.8) 71 (-6.6) Iraq 1503 2 (2.7) 64 (-0.7) Mozambique 1803 2 (2.2) 70 (-2.5) South Africa 2453 3 (2.3) 43 (0.4) Jordan 608 3 (2.1) 18 (0.7) COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 5 129 Country Mortality rate in Performance ranking Country Mortality rate in Performance ranking 2019 (per 100,000 (performance*) in 2019 (per 100,000 (performance*) in population) population) 2010-19 2000-10 2010-19 2000-10 50-69 males (continued) 50-69 females (continued) Top 10 performance countries during 2010-19 (continued) Top 10 performance countries during 2010-19 (continued) Tanzania 1972 4 (1.8) 66 (-0.9) Zambia 1581 4 (2) 8 (1.2) Jordan 1076 5 (1.6) 19 (0.9) Malawi 1677 5 (1.9) 58 (-1.1) Zambia 2218 6 (1.3) 10 (1.5) Algeria 590 6 (1.9) 1 (2) Mozambique 2626 7 (1.3) 70 (-1.7) Bangladesh 814 7 (1.9) 26 (0.3) Bangladesh 1396 8 (1.1) 68 (-1.3) Tanzania 1463 8 (1.8) 55 (-1) Algeria 861 9 (1.1) 1 (3.1) Ukraine 861 9 (1.8) 50 (-0.5) Peru 1013 10 (0.8) 23 (0.8) South Africa 1554 10 (1.7) 32 (0.1) Worst 10 performance countries during 2010-19 Worst 10 performance countries during 2010-19 Viet Nam 1440 71 (-2.2) 18 (0.9) Viet Nam 635 71 (-2.1) 27 (0.3) Thailand 1343 70 (-1.7) 30 (0.6) India 1291 70 (-1.8) 25 (0.3) Mexico 1423 69 (-1.5) 44 (0.4) Bulgaria 886 69 (-1.5) 31 (0.2) Bulgaria 2047 68 (-1.4) 32 (0.6) Guatemala 1069 68 (-1.4) 59 (-1.1) Malaysia 1496 67 (-1.3) 47 (0.1) Cambodia 1051 67 (-1.3) 6 (1.4) Rwanda 2205 66 (-1.2) 7 (1.8) Cote d’Ivoire 2297 66 (-1.3) 65 (-1.5) Kenya 2733 65 (-1.2) 65 (-0.8) Malaysia 841 65 (-1.1) 21 (0.5) Tunisia 1239 64 (-1.2) 28 (0.6) Uzbekistan 1172 64 (-1) 19 (0.7) Philippines 1536 63 (-1.2) 31 (0.6) Nigeria 2289 63 (-0.9) 67 (-1.9) Uzbekistan 1956 62 (-1) 13 (1.4) Kenya 1942 62 (-0.9) 52 (-0.7) 70-84 males 70-84 females Top 10 performance countries during 2010-19 Top 10 performance countries during 2010-19 Haiti 9175 1 (1.7) 71 (-2.8) Haiti 7390 1 (2.6) 71 (-3.7) Sri Lanka 6941 2 (1.6) 42 (-0.3) Iran 3905 2 (1.7) 13 (0.5) India 6301 3 (1.4) 53 (-0.5) Ukraine 4680 3 (1.7) 48 (-0.5) Russian Federation 7560 4 (1.4) 17 (0.3) Morocco 5054 4 (1.6) 50 (-0.6) Iraq 7497 5 (1.2) 52 (-0.4) Sri Lanka 3792 5 (1.5) 4 (0.9) Iran 6150 6 (1.2) 1 (1.1) Türkiye 3684 6 (1.4) 34 (-0.2) Türkiye 5929 7 (1.1) 18 (0.3) Peru 3764 7 (1.4) 27 (-0.1) Algeria 5195 8 (1) 3 (0.9) Colombia 3135 8 (1.3) 8 (0.7) Peru 5419 9 (1) 8 (0.6) Russian Federation 4792 9 (1.1) 22 (0) Sudan 6889 10 (0.9) 34 (-0.2) Bulgaria 4419 10 (1.1) 2 (1) Worst 10 performance countries during 2010-19 Worst 10 performance countries during 2010-19 Uzbekistan 9451 71 (-0.7) 68 (-1.5) Nigeria 9121 71 (-0.8) 59 (-0.9) Mexico 5519 70 (-0.6) 66 (-1.3) Uzbekistan 7356 70 (-0.8) 69 (-1.6) Tunisia 6478 69 (-0.5) 45 (-0.3) Azerbaijan 6212 69 (-0.7) 47 (-0.5) Nigeria 9364 68 (-0.5) 27 (0) Benin 7355 68 (-0.6) 64 (-1) Senegal 8854 67 (-0.5) 44 (-0.3) Mexico 4114 67 (-0.6) 70 (-1.8) Burkina Faso 9944 66 (-0.5) 20 (0.3) Cameroon 7592 66 (-0.6) 55 (-0.8) Benin 8038 65 (-0.4) 43 (-0.3) Mali 8882 65 (-0.5) 19 (0.1) Niger 8924 64 (-0.4) 15 (0.4) Pakistan 6660 64 (-0.5) 26 (-0.1) 130 CHAPTER 5 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E Country Mortality rate in Performance ranking Country Mortality rate in Performance ranking 2019 (per 100,000 (performance*) in 2019 (per 100,000 (performance*) in population) population) 2010-19 2000-10 2010-19 2000-10 70-84 males (continued) 70-84 females (continued) Worst 10 performance countries during 2010-19 (continued) Worst 10 performance countries during 2010-19 (continued) Pakistan 8329 63 (-0.4) 60 (-0.7) Uganda 9373 63 (-0.5) 67 (-1.2) Afghanistan 8730 62 (-0.3) 21 (0.3) Guinea 7580 62 (-0.5) 37 (-0.3) * Performance is defined as the difference between the actual rate of decline in mortality and the rate predicted by the multilevel regression model. Countries in italics are those that encountered major crises such as hurricanes, earthquakes, flooding, tsunami, or war during 2000, 2005, 2010, 2015, or 2019, thereby showing a pseudo good or poor in performance during 2010-2019 relative to during 2000-2010. RATES OF PROGRESS IN CAUSE-SPECIFIC MORTALITY Country-level monitoring of mortality rates is important for monitoring trends over time. To reduce mortality rates, monitoring of death rates should be accompanied by deaths due to each disease cause. Globally, the pro- portion of deaths due to communicable, maternal, perinatal, and nutritional conditions (CMPs) are declining, while that from NCDs are increasing (WHO 2020). This is particularly so in LMICs, where CMPs, as a propor- tion of total deaths have declined from 39 percent in 2000 to 21 percent in 2019. Meanwhile, the contribution from NCDs have increased from 55 percent to 70 percent (WHO 2020). Here, we examine and quantify the rates of progress in cause-specific mortality for selected major causes of death, by country income levels and age, from 2000 to 2019. We measure rates of progress as average annual rates of reduction (AARR) calculated using the standard WHO methodology (UNICEF 2007). We present in Table 5.3 the country income levels 15 in upper-middle-income countries showed the and age groups with the fastest and slowest prog- fastest rate of decline, while falls at ages 70 and over ress, in terms of AARR, for males and females. in upper-middle-income countries showed an aver- Between 2000 and 2019, among CMPs, child- age annual increase of about 2.3 percent. hood-cluster diseases (based on the definition the We further estimated the number of deaths due 2019 WHO Global Health Estimates (WHO 2020): to the selected major causes of death that would whooping cough, diphtheria, measles, and tetanus) have been averted in 2019, had the rate of decline among children aged 0-14 years in upper-mid- for each of the selected causes of death been equal dle-income countries showed the fastest rate of de- to that of HIV/AIDS for ages under 49 years and cline in mortality, with an AARR of about 9 percent. tuberculosis for ages 50 years and over. These two The slowest rates of decline among CMPs in LMICs causes of deaths were selected based on the fact was observed for malaria in lower-middle-income that immense attention and efforts were invest- countries (AARR=0.4 percent for males and 1.1 ed towards these diseases during the 2000s which percent for females). Compared to CMPs, progress resulted in the remarkable drop in incidence and in NCDs are much slower: the fastest rate of decline mortality from these causes (WHO 2015a, 2015c, among NCDs was observed for respiratory diseases 2015b). If similar levels of effort were spent on all for ages 50-69 in upper-middle-income countries, causes and thus, expecting similar rates of progress, with an AARR of about 5 percent (almost half the about 13.5 million deaths would have been averted fastest rate for CMPs). The slowest progress was globally. Out of this, about 4.6 million (or 34 per- in tobacco-attributable cancers and breast cancer cent) would been premature mortality under 70 among those aged 70 years and older in low-income years. The deaths avertable by causes and age groups countries. Within injuries, drowning at ages under in Annex Table 5A.4. TABLE 5.3  Diseases and conditions showing country income level and age group with the fastest and slowest progress in mortality reduction, 2000-19 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E Male Female Fastest Slowest Fastest Slowest Disease/condition category Diseases/conditions Region Age group AARR % Region Age group AARR % Region Age group AARR % Region Age group AARR % All Causes UMICs 0-14 4.6 UMICs 70+ 0.7 UMI 0-14 4.6 HICs 70+ 0.9 Communicable, Tuberculosis UMICs 70+ 6.3 LICs 70+ 4.1 HICs 50-69 6.3 HICs 70+ 3.3 maternal, perinatal and nutritional conditions HIV/AIDS HICs 15-49 8.1 UMICs 15-49 1.9 LICs 15-49 8.0 UMICs 15-49 2.4 Diarrhoeal diseases UMICs 0-14 6.9 HICs 0-14 4.1 UMICs 0-14 7.2 HICs 0-14 4.1 Childhood-cluster diseases UMICs 0-14 9.1 LICs 0-14 5.2 UMICs 0-14 9.2 HICs 0-14 5.3 Malaria LICs 0-14 6.3 LMICs 15-49 0.4 LICs 0-14 6.4 LMICs 15-49 1.1 Respiratory infections UMICs 0-14 6.5 HICs 0-14 4.5 UMICs 0-14 6.5 HICs 40-14 4.5 Maternal conditions LMI 15-49 5.8 HICs 15-49 0.6 Neonatal conditions UMICs 0-28 days 4.8 HICs 0-28 days 2.2 UMICs 0-28 days 4.8 HICs 0-28 days 2.1 Non-communicable All cancers HICs 50-69 1.8 LICs 70+ -0.2 UMICs 50-69 1.2 LICs 70+ -0.2 diseases Tobacco-attributable cancersa HICs 50-69 2.1 LICs 70+ -0.4 UMICs 50-69 1.3 LICs 70+ -0.7 Infection attributable cancersb UMICs 50-69 4.1 HICs 70+ 0.3 UMICs 50-69 2.2 HICs 70+ 0.4 Stomach cancer HICs 50-69 3.9 LICs 70+ 0.5 UMICs 50-69 3.4 LICs 70+ 0.8 Breast cancer HICs 50-69 1.5 LICs 70+ -1.2 Cardiovascular diseases (not stroke) HICs 70+ 2.3 UMICs 70+ -0.2 UMICs 50-69 2.5 LICs 70+ 0.2 Stroke HICs 50-69 3.4 LICs 70+ 0.9 HICs 50-69 3.9 LICs 70+ 0.7 Respiratory diseases UMICs 50-69 4.8 HICs 50-69 0.7 UMICs 50-69 5.3 HICs 50-69 -0.3 Injuries Road injury HICs 15-49 3.5 LICs 70+ -1.2 HICs 70+ 3.7 LICs 70+ -1.0 Falls LICs 50-69 1.1 UMICs 70+ -2.3 LICs 50-69 1.3 UMICs 70+ -2.3 Drowning UMICs 0-14 6.2 LICs 0-14 2.7 UMICs 0-14 7.6 LICs 0-14 2.8 Suicide UMICs 50-69 3.7 HICs 15-49 0.5 UMICs 15-49 6.2 HICs 15-49 -0.3 CHAPTER 5 AARR = Average annual rate of reduction; LICs = low-income countries; LMICs = lower-middle-income countries; UMICs = upper-middle-income countries; HICs = high-income countries; aMouth and oropharynx, Oesophagus, Trachea, bronchus, lung, and Larynx cancers; bLiver, Cervix uteri, and Corpus uteri cancers. 131 132 CHAPTER 5 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E COST OF SAVING A LIFE OVER TIME In 1975, Samuel Preston, in a classic paper, showed that life expectancy is related to national income (Preston 1975). He found that life expectancy increases with national income per capita in poorer countries, but plateaus at higher income levels. He also noted an upward and lateral shift in the curve over time, indicating that for the same level of income, life expectancy increases over time (Figure 5.2). This relationship was mathematically studied by Hum and colleagues in 2012 using the Michaelis-Menten enzyme kinetics (Hum and others 2012). Treating income as the substrate that is catalyzed to increase survival, Hum and colleagues investigated the change in the level of income that is needed to achieve half of the period-specific maximum survival (“critical income”). The study found that between 1970 and 2007, the critical income fell by 50 percent for children but doubled for men aged 15-59 years. Here, we extend the 2012 analysis by Hum and colleagues to assess the trend in the critical income for ages under 15, 15-49, and 50-69 from 1990 to 2019. We redefine critical income here as the income needed to achieve 80 percent of the global maximum life expectancy. We used five-year average to calculate the average critical income at a mid-point year (starting from 1992.5 for 1990-95 and 2017 for 2015- 19). The analysis includes 71 LMICs with over 7 million population in 2019 (United Nations, Department of Economic and Social Affairs, Population Division 2022). Details of the model can be found in Annex 5.3. Table 5.4 shows the coefficients in three-age specif- remained relatively constant over the last thirty years. ic mixed effects models. Expectedly, the maximum Regrettably, that for ages 50-69 increased by about 26 survival rates for children approximates a complete percent from US$914 to US$1,180. This signifies that survival from birth to age 15 and was unchanged be- the technological gains in HICs to reduce mortality tween 1992 and 2017. While for ages 15-49 and 50-69, at middle and older ages lacks research on improving the maximum survival rates increased by 0.6 percent efficiency and cost-effectiveness. and 4.4 percent, respectively. This change represents Table 5.5 ranks the top five LMICs with the low- the technological advances for this age-group among est critical income in the three age groups. Rwanda, HICs. In terms of critical income, the pattern of prog- Madagascar, Mozambique, Niger, and the Dem- ress is reversed. The critical income for child survival ocratic Republic of Congo are among the top five fell from US$1,452 to US$800 (2017 PPP), represent- countries with the lowest critical income across all ing a decline of 46 percent. For adults aged 15-49, this three age-groups. FIGURE 5.2  Preston curve showing the relationship between life expectancy at birth and income (Preston 1975) COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 5 133 TABLE 5.4  Global maximum survival rates and critical income levels, 1990 to 2019 Under 15 years 15-49 years 50-69 years Median years 1990-95 2015-19 1990-95 2015-19 1990-95 2015-19 Maximum survival 1,000.96 1,000.81 985.40 990.98 868.68 907.34 rates (per 1000) (2.89) (0.83) (1.68) (1.16) (6.78) (5.96) Critical income 1,452.56 800.16 400.40 412.52 914.36 1,179.56 (2017 PPP) AIC 778.16 629.19 650.38 621.15 810.04 845.44 Standard errors are given in parenthesis. TABLE 5.5  Top five LMICs with the lowest critical income for ages under 15 years, 15-49 years, and 50-69 years Countries with the lowest critical income Rank Under 15 years 15-49 years 50-69 years 1 Malawi Niger Niger 2 Rwanda Congo, Dem. Rep. Congo, Dem. Rep. 3 Madagascar Madagascar Madagascar 4 Mozambique Bangladesh Ethiopia 5 Uganda Rwanda Cambodia REFERENCES 1. Cha, S. 2017. “The impact of the worldwide Millennium De- D.C. : World Bank Group. http://documents.worldbank.org/ velopment Goals campaign on maternal and under-five child curated/en/920581467995896846/Delivering-the-millenni- mortality reduction: ‘Where did the worldwide campaign work um-development-goals-to-reduce-maternal-and-child-mor- most effectively?’” Global Health Action 10 (1): 1267961. tality-a-systematic-review-of-impact-evaluation-evidence. 2. Hum, R. J., P. Jha, A. M. McGahan, and Y. –L. Cheng. 2012. “Global 8. UNICEF. 2007. “Technical Note: How to calculate average an- divergence in critical income for adult and childhood survival: nual rate of reduction (AARR) of underweight prevalence.” Analyses of mortality using Michaelis–Menten.” ELife 1: e00051. https://data.unicef.org/resources/technical-note-calculate-av- 3. Lozano, R., H. Wang, K. J. Foreman, J. K. Rajaratnam, M. Naghavi, erage-annual-rate-reduction-aarr-underweight-prevalence/. and others. 2011. “Progress towards Millennium Development 9. United Nations General Assembly. “Transforming our world : the Goals 4 and 5 on maternal and child mortality: An updated sys- 2030 Agenda for Sustainable Development”. 21 October 2015, tematic analysis.” The Lancet 378 (9797), 1139–165. A/RES/70/1. https://www.refworld.org/docid/57b6e3e44.html. 4. NCD Countdown 2030 collaborators. 2018. “NCD Countdown 10. United Nations, Department of Economic and Social Affairs, 2030: Worldwide trends in non-communicable disease mortal- Population Division. 2022. World Population Prospects 2022. ity and progress towards Sustainable Development Goal target New York: UN Department of Economic and Social Affairs, Pop- 3.4.” The Lancet 392 (10152): 1072–88. ulation Division. 5. Norheim, O. F., P. Jha, K. Admasu, T. Godal, R. J. Hum, and others. 11. United Nations, Department of Economic and Social Affairs, 2015. “Avoiding 40% of the premature deaths in each country, Sustainable Development. 2023. Goal 3 Ensure healthy lives 2010–30: Review of national mortality trends to help quantify and promote well being for all at all ages. https://sdgs.un.org/ the UN Sustainable Development Goal for health.” The Lancet goals/goal3. 385 (9964), 239–252. 12. Verguet, S., and D. T. Jamison. 2014. “Estimates of performance 6. Preston, S. H. 1975. “The changing relation between mortality in the rate of decline of under-five mortality for 113 low- and and level of economic development.” Population Studies 29 (2), middle-income countries, 1970-2010.” Health Policy and Plan- 231–48. ning 29(2), 151–63. 7. Tanner, J., A. M. Aguilar Rivera, T. L. Candland, V. Galdo, F. E. 13. Wang, H., A. A. Abajobir, K. H. Abate, C. Abbafati, K. M. Abbas, Manang, and others. 2016. “Delivering the millennium develop- and others. 2017. “Global, regional, and national under-5 mor- ment goals to reduce maternal and child mortality: A systemat- tality, adult mortality, age-specific mortality, and life expectan- ic review of impact evaluation evidence (English).” Washington, cy, 1970–2016: A systematic analysis for the Global Burden of 134 CHAPTER 5 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E Disease Study 2016.” The Lancet 390 (10100), 1084–150. progress report. World Health Organization. https://apps.who. 14. World Bank. 2021. World Bank Country and Lending Groups. int/iris/handle/10665/198065. https://datahelpdesk.worldbank.org/knowledgebase/arti- 17. World Health Organization (WHO). 2015c. Global tuberculosis cles/906519-world-bank-country-and-lending-groups. report 2015, 20th ed. World Health Organization. https://apps. 15. World Health Organization (WHO). 2015a. Accelerating prog- who.int/iris/handle/10665/191102. ress on HIV, tuberculosis, malaria, hepatitis and neglected 18. World Health Organization (WHO). 2020. Global Health Es- tropical diseases: A new agenda for 2016-2030. World Health timates 2020: Deaths by Cause, Age, Sex, by Country and by Organization. https://apps.who.int/iris/handle/10665/204419. Region, 2000-2019. https://www.who.int/data/gho/data/ 16. World Health Organization (WHO). 2015b. Global health sector themes/mortality-and-global-health-estimates/ghe-leading- response to HIV, 2000– 2015: Focus on innovations in Africa: causes-of-death. ANNEX 5.1 Interrupted time-series (ITS) analysis methodology An ITS analysis provides a robust quasi-experimental design that can be used to evaluate changes in an outcome of interest after a particular time-point (Bernal and others 2017; Kontopantelis and others 2015). In an ITS anal- ysis, a continuous series of observations on an outcome of interest for a population over time is used to establish a trend using a regression model; this trend is then interrupted (usually by an intervention) at a known time-point (Bernal and others 2017; Kontopantelis and others 2015). The difference in the trend before and after the time- point quantifies the effect of the intervention. Figure 5A.1 provides a schematic representation of an ITS analysis. FIGURE 5A.1  Schematic representation of an interrupted time series analysis COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 5 135 For our current analysis, in the absence of inter- compared to the first, using the logic that trends vention, we define the interruption point as 2010 to which had been established from 2000-2009 should examine if there are notable difference in the rate accelerate after 2010 if more attention was given to of mortality decline in the second decade of 2000 health and action on health during the second. FIGURE 5A.2  Mortality rate, global and by country income level, by age and sex, between 2000 and 2019 ANNEX 5.2 Methods used to estimate country performance during 2000 to 2019 We examine country performance in the rate of de- to that predicted a model using a set of contextual cline in mortality for ages under 15, 15-49, 50-69, factors. We include 71 LMICs (based on the World and 70-84 years between 2000 and 2019. We used Bank 2020 country classification (World Bank the framework of Verguet and Jamison (2014) to 2021)) that have a population of over 7 million, estimate performance in the rate of decline in un- which covers about 83 percent of the world popu- der-five mortality in LMICs (Verguet & Jamison, lation in 2019. For population and actual mortali- 2014). We define performance as the difference in ty, we used data from the World Population Pros- the rate of decline in actual mortality compared pects (WPP) 2022 (United Nations, Department of 136 CHAPTER 5 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E Economic and Social Affairs, Population Division ln mit = γ00 + γ10Time + β2 ln(GDPPC) + β3FeEd 2022). For predicted mortality, we used a hierarchi- + β4Coast + β5Urban + β6Electricity + cal linear model comprised of the contextual vari- β7Malaria + β8HIV + β9Skillbirth-2000 + ables: time (year), income (gross domestic product β10DTP3-2000 + µ0i + µ1iTime + εit per capita (GDPPC)), female education level (mean years of schooling by age 25), geography (percent where, γ are the country-invariant intercept and population living within 100km off the coast), ur- slope (time) components, µ are the country-specific banization, electricity access, malaria incidence, terms, and εit is a random i.i.d. normally distributed and HIV prevalence. We allow for between-coun- disturbance (for details, see Jamison, Murphy, and try variability by using a random-intercept term at Sandbu 2016 on the use of hierarchical model to ac- the country level, and between-country variability count for technical progress to predict mortality). in their uptake of knowledge and technologies (or Using the performances estimated over five “technical progress”) using a country-specific slope years, we calculate 10-year performance as the aver- on time (Jamison and others 2016). To examine the age of two consecutive 5-year performance. Finally, potential impact of public health provision on tech- we calculate the change in performance in the rate nical progress, we include in the model the percent- of decline in mortality in the second decade of 2000 age of births attended by skilled birth attendant and relative to the first decade as the difference between coverage of three doses of diphtheria, tetanus, and the two consecutive 10-year performance. pertussis (DTP3) immunization among one-year- Annex table 5A.1 shows the coefficients in the olds in 2000 as indicators of early adopters of mor- hierarchical model used to predict the mortality tality reduction technologies. Thus, the model can rates for ages under 15, 15-49, 50-69, and 70-84 be mathematically written as: years, disaggregated by sex. Table 5A.1  Results of the hierarchical model predicting levels of mortality, by age and sex COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E Coefficient Under 15 years 15-49 years 50-69 years 70-84 years Males Females Males Females Males Females Males Females Estimate SE Estimate SE Estimate SE Estimate SE Estimate SE Estimate SE Estimate SE Estimate SE Constant inter- 9.510 0.361 9.588 0.387 7.608 0.357 7.636 0.337 8.847 0.206 8.840 0.207 9.200 0.143 9.336 0.158 cept (γ00) Time (number -0.038 0.003 -0.037 0.003 -0.019 0.003 -0.019 0.002 -0.012 0.001 -0.014 0.001 -0.006 0.001 -0.008 0.001 of years since 1990, γ10) ln(GDPPC) -0.121 0.049 -0.132 0.052 -0.117 0.048 -0.104 0.047 -0.117 0.027 -0.125 0.029 0.024 0.019 0.013 0.021 FeEd 0.005 0.013 0.000 0.014 0.010 0.013 0.004 0.013 0.021 0.008 0.008 0.008 0.005 0.005 -0.002 0.006 Coast -0.004 0.001 -0.003 0.001 -0.004 0.001 -0.002 0.001 -0.003 0.001 -0.002 0.001 -0.002 0.001 -0.001 0.001 Urban 0.000 0.003 0.001 0.003 0.002 0.003 0.001 0.002 0.001 0.002 0.000 0.002 -0.002 0.001 -0.003 0.001 Electricity -0.001 0.001 -0.002 0.001 0.000 0.001 -0.004 0.001 0.001 0.001 0.000 0.001 0.000 0.000 0.000 0.000 Malaria 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 HIV 0.057 0.012 0.056 0.012 0.059 0.012 0.088 0.010 0.030 0.007 0.048 0.006 0.003 0.005 0.007 0.005 Skillbirth-2000 -0.011 0.003 -0.012 0.003 -0.001 0.003 -0.003 0.002 -0.001 0.002 -0.003 0.001 0.000 0.001 -0.001 0.001 DTP3-2000 -0.013 0.003 -0.014 0.003 -0.006 0.003 -0.009 0.002 -0.002 0.002 -0.003 0.002 -0.002 0.001 -0.004 0.001 Variance of 0.111 0.023 0.128 0.027 0.121 0.024 0.052 0.014 0.045 0.009 0.028 0.006 0.022 0.004 0.026 0.005 country-spe- cific intercept (µ0i) Variance of 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 country-specif- ic slope (µ1i) AIC -248.81 -191.82 -247.21 -196.43 -643.37 -588.00 -917.17 -853.47 Values in bold indicate significance at 95% confidence interval. Dependent variable is logged mortality rate. CHAPTER 5 137 138 CHAPTER 5 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E TABLE 5A.2  Performance rankings of 71 LMICs in the rate of decline of mortality, by age and sex: 2000-10 and 2010-19 Under 15 years Males Females Mortality rate in Performance ranking Mortality rate in Performance ranking 2019 (per 100,000 (performance*) in 2019 (per 100,000 (performance*) in Country (ordered based on ranking during 2010-19) population) 2010-19 2000-10 population) 2010-19 2000-10 Russian Federation 46 1 (5.6) 33 (-0.4) 37 2 (5.2) 52 (-1.8) Haiti 610 2 (5.6) 71 (-8.8) 474 1 (7.6) 71 (-10.2) Azerbaijan 157 3 (5.1) 63 (-2.3) 121 3 (5.2) 64 (-2.6) Belarus 27 4 (4.6) 16 (0.7) 20 5 (4.5) 34 (-0.4) Kazakhstan 105 5 (4.6) 42 (-0.9) 74 4 (4.7) 51 (-1.8) China 71 6 (3.8) 25 (0.4) 61 6 (3.8) 24 (0.5) Ukraine 55 7 (3.8) 64 (-2.4) 43 7 (3.8) 69 (-3.9) Bulgaria 55 8 (3.5) 67 (-3) 52 17 (1.4) 62 (-2.6) Malawi 392 9 (3.2) 8 (2.1) 337 9 (2.6) 7 (2.5) Burundi 685 10 (2.8) 50 (-1.6) 576 8 (3.4) 48 (-1.5) South Africa 319 11 (2) 65 (-2.7) 271 10 (2.4) 63 (-2.6) Uzbekistan 154 12 (2) 27 (0.2) 114 14 (1.8) 25 (0.5) Türkiye 77 13 (1.9) 2 (3.7) 66 11 (2.1) 2 (3.8) India 255 14 (1.5) 20 (0.6) 266 13 (1.8) 22 (0.6) Rwanda 398 15 (1.5) 1 (4.9) 345 18 (1) 1 (5.2) Uganda 443 16 (1.4) 7 (2.2) 386 26 (0.5) 8 (2.3) Morocco 167 17 (1.3) 29 (0.1) 134 15 (1.5) 29 (0.1) Iran 122 18 (1.3) 62 (-2.2) 108 16 (1.5) 61 (-2.6) Sri Lanka 61 19 (1.3) 41 (-0.9) 51 19 (1) 41 (-0.9) Senegal 416 20 (1.2) 11 (1.4) 329 12 (2) 9 (1.9) Thailand 78 21 (1.1) 34 (-0.4) 61 22 (0.9) 35 (-0.5) Iraq 246 22 (1) 53 (-1.7) 197 30 (0.4) 46 (-1.4) Egypt 167 23 (1) 37 (-0.6) 143 20 (0.9) 32 (-0.3) Argentina 87 24 (0.8) 38 (-0.7) 74 23 (0.7) 40 (-0.8) Ecuador 123 25 (0.8) 46 (-1.5) 97 24 (0.6) 50 (-1.7) Jordan 131 26 (0.7) 58 (-2) 106 25 (0.6) 54 (-1.8) Zimbabwe 516 27 (0.6) 70 (-4.4) 444 32 (0.3) 70 (-4.5) Sierra Leone 1023 28 (0.5) 26 (0.3) 928 21 (0.9) 26 (0.4) Mozambique 744 29 (0.3) 30 (0.1) 642 28 (0.4) 27 (0.3) Indonesia 206 30 (0.3) 36 (-0.5) 173 27 (0.5) 33 (-0.4) Tanzania 492 31 (0.3) 24 (0.5) 439 43 (-0.1) 18 (0.9) Bangladesh 247 32 (0.3) 10 (1.6) 217 35 (0.1) 10 (1.6) Niger 854 33 (0.2) 4 (3) 794 29 (0.4) 5 (3.2) Cameroon 783 34 (0.2) 59 (-2.1) 656 31 (0.3) 56 (-2) Guatemala 212 35 (0.1) 21 (0.6) 173 40 (-0.1) 20 (0.8) Mexico 117 36 (0.1) 15 (0.8) 95 33 (0.1) 19 (0.8) Peru 128 37 (0.1) 14 (0.8) 103 36 (0.1) 17 (0.9) Brazil 128 38 (-0.1) 3 (3.2) 100 44 (-0.3) 4 (3.3) Cambodia 287 39 (-0.1) 5 (2.7) 210 39 (0) 3 (3.6) COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 5 139 Under 15 years (continued) Males Females Mortality rate in Performance ranking Mortality rate in Performance ranking 2019 (per 100,000 (performance*) in 2019 (per 100,000 (performance*) in Country (ordered based on ranking during 2010-19) population) 2010-19 2000-10 population) 2010-19 2000-10 Nepal 269 40 (-0.2) 6 (2.5) 220 38 (0) 6 (3.1) Afghanistan 594 41 (-0.3) 32 (-0.3) 526 47 (-0.4) 39 (-0.7) Mali 1006 42 (-0.3) 44 (-1.1) 924 45 (-0.3) 42 (-1) Sudan 662 43 (-0.4) 47 (-1.5) 525 42 (-0.1) 47 (-1.4) Pakistan 599 44 (-0.4) 52 (-1.6) 504 41 (-0.1) 43 (-1) Laos 421 45 (-0.5) 35 (-0.4) 331 49 (-0.5) 30 (-0.1) Tajikistan 306 46 (-0.5) 17 (0.7) 235 48 (-0.5) 16 (0.9) Algeria 211 47 (-0.5) 69 (-3.5) 189 50 (-0.5) 68 (-3.8) Serbia 49 48 (-0.5) 9 (2) 39 57 (-1.1) 11 (1.4) Madagascar 582 49 (-0.5) 18 (0.7) 466 34 (0.1) 12 (1.4) Chad 1440 50 (-0.5) 66 (-2.8) 1300 46 (-0.3) 66 (-2.8) Honduras 153 51 (-0.6) 12 (1.4) 125 51 (-0.5) 13 (1.3) Kenya 384 52 (-0.7) 22 (0.5) 347 63 (-1.5) 23 (0.5) Ethiopia 603 53 (-0.8) 23 (0.5) 446 37 (0) 14 (1.3) Burkina Faso 913 54 (-0.9) 39 (-0.8) 821 52 (-0.7) 37 (-0.7) Colombia 124 55 (-1) 28 (0.2) 98 56 (-1.1) 28 (0.1) Bolivia 341 56 (-1.1) 19 (0.7) 294 53 (-0.7) 21 (0.7) Zambia 647 57 (-1.1) 13 (0.9) 539 59 (-1.1) 15 (1) Congo, Dem. Rep. 1023 58 (-1.2) 45 (-1.2) 869 55 (-1) 45 (-1.3) Papua New Guinea 420 59 (-1.2) 49 (-1.5) 365 58 (-1.1) 49 (-1.5) Togo 711 60 (-1.3) 56 (-1.7) 672 61 (-1.3) 60 (-2.4) Ghana 555 61 (-1.4) 60 (-2.1) 449 54 (-0.9) 58 (-2) Tunisia 152 62 (-1.6) 57 (-2) 125 65 (-1.7) 59 (-2.1) Myanmar 438 63 (-1.6) 48 (-1.5) 325 60 (-1.2) 44 (-1.3) Benin 1005 64 (-1.7) 55 (-1.7) 889 62 (-1.5) 53 (-1.8) Cote d’Ivoire 822 65 (-1.8) 40 (-0.9) 659 64 (-1.7) 38 (-0.7) Nigeria 1377 66 (-1.8) 54 (-1.7) 1351 67 (-2.1) 57 (-2) Philippines 247 67 (-2) 51 (-1.6) 194 66 (-1.8) 55 (-1.9) Guinea 1054 68 (-2.2) 43 (-1) 954 68 (-2.3) 36 (-0.6) Dominican Republic 272 69 (-2.7) 61 (-2.1) 227 69 (-2.9) 65 (-2.8) Viet Nam 245 70 (-3.1) 68 (-3.1) 144 70 (-3.4) 67 (-3.2) Malaysia 72 71 (-3.6) 31 (-0.3) 61 71 (-4.4) 31 (-0.2) 15-49 years Males Females Mortality rate in Performance ranking Mortality rate in Performance ranking 2019 (per 100,000 (performance*) in 2019 (per 100,000 (performance*) in Country (ordered based on ranking during 2010-19) population) 2010-19 2000-10 population) 2010-19 2000-10 Haiti 423 1 (7) 71 (-9.9) 246 1 (11.4) 71 (-13.3) Iraq 193 2 (4.6) 70 (-3.2) 132 12 (2.1) 62 (-2.6) South Africa 575 3 (4.5) 13 (1.2) 396 2 (6.9) 19 (0.6) 140 CHAPTER 5 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E 15-49 years (continued) Males Females Mortality rate in Performance ranking Mortality rate in Performance ranking 2019 (per 100,000 (performance*) in 2019 (per 100,000 (performance*) in Country (ordered based on ranking during 2010-19) population) 2010-19 2000-10 population) 2010-19 2000-10 Tanzania 320 4 (3.6) 45 (-0.9) 239 4 (3.1) 29 (0) Mozambique 522 5 (2.2) 62 (-2) 311 3 (4.7) 65 (-2.8) Zimbabwe 632 6 (2.2) 65 (-2.2) 478 7 (2.6) 70 (-4.6) Serbia 121 7 (2.1) 4 (1.6) 66 11 (2.1) 5 (1.8) Belarus 301 8 (2.1) 27 (0.1) 84 14 (1.9) 24 (0.2) Russian Federation 384 9 (2) 9 (1.4) 104 13 (2) 6 (1.6) Sierra Leone 367 10 (1.7) 1 (2.8) 318 10 (2.2) 2 (2) Uganda 438 11 (1.7) 8 (1.4) 315 16 (1.6) 21 (0.4) Tajikistan 186 12 (1.6) 37 (-0.4) 123 28 (0.6) 35 (-0.4) Sri Lanka 198 13 (1.5) 14 (1.1) 89 41 (-0.2) 38 (-0.6) Morocco 137 14 (1.5) 5 (1.6) 103 15 (1.6) 15 (0.9) Laos 255 15 (1.5) 16 (0.8) 183 29 (0.5) 10 (1.2) Jordan 105 16 (1.4) 33 (-0.3) 76 23 (1) 30 (-0.1) Ukraine 315 17 (1.3) 11 (1.2) 85 5 (2.9) 13 (1.1) Guatemala 304 18 (1.3) 34 (-0.3) 138 53 (-0.9) 27 (0) Brazil 234 19 (1.2) 25 (0.3) 101 21 (1.1) 17 (0.7) Zambia 488 20 (1.2) 10 (1.3) 310 6 (2.7) 7 (1.5) Kazakhstan 337 21 (1.1) 18 (0.7) 134 27 (0.7) 42 (-0.9) Türkiye 107 22 (1.1) 26 (0.2) 52 22 (1) 22 (0.3) Iran 190 23 (1) 47 (-1.1) 86 9 (2.3) 51 (-1.4) Egypt 161 24 (0.8) 39 (-0.6) 87 20 (1.1) 16 (0.8) Algeria 128 25 (0.8) 7 (1.5) 93 18 (1.2) 14 (1) Rwanda 358 26 (0.7) 2 (2.7) 250 31 (0.2) 1 (3.8) Ecuador 236 27 (0.5) 58 (-1.7) 105 36 (0) 49 (-1.4) Thailand 284 28 (0.4) 35 (-0.3) 105 25 (0.9) 33 (-0.4) India 245 29 (0.4) 56 (-1.5) 163 50 (-0.6) 28 (0) Malaysia 153 30 (0.4) 24 (0.3) 71 32 (0.2) 11 (1.1) Nepal 242 31 (0.4) 23 (0.4) 177 40 (-0.2) 26 (0.1) Azerbaijan 225 32 (0.3) 51 (-1.4) 113 44 (-0.4) 61 (-2.4) Bangladesh 179 33 (0.3) 67 (-2.5) 135 39 (-0.2) 43 (-0.9) Burundi 530 34 (0.2) 6 (1.6) 395 19 (1.2) 45 (-1) Sudan 400 35 (0.2) 21 (0.5) 255 33 (0) 34 (-0.4) Colombia 255 36 (0.1) 3 (2.3) 101 61 (-1.4) 23 (0.3) Cameroon 470 37 (-0.1) 44 (-0.8) 394 43 (-0.4) 55 (-1.9) Ethiopia 428 38 (-0.2) 32 (-0.1) 258 26 (0.9) 9 (1.3) Argentina 156 39 (-0.2) 28 (0) 80 24 (1) 25 (0.1) Niger 335 40 (-0.2) 20 (0.5) 314 35 (0) 20 (0.5) Papua New Guinea 383 41 (-0.4) 50 (-1.4) 237 34 (0) 56 (-1.9) Madagascar 413 42 (-0.4) 29 (-0.1) 277 30 (0.4) 8 (1.4) Peru 214 43 (-0.4) 49 (-1.3) 112 38 (-0.1) 44 (-0.9) COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 5 141 15-49 years (continued) Males Females Mortality rate in Performance ranking Mortality rate in Performance ranking 2019 (per 100,000 (performance*) in 2019 (per 100,000 (performance*) in Country (ordered based on ranking during 2010-19) population) 2010-19 2000-10 population) 2010-19 2000-10 Pakistan 250 44 (-0.4) 66 (-2.4) 163 49 (-0.6) 4 (1.8) China 144 45 (-0.4) 53 (-1.4) 80 42 (-0.2) 40 (-0.7) Senegal 282 46 (-0.5) 17 (0.7) 180 17 (1.4) 12 (1.1) Malawi 553 47 (-0.5) 19 (0.5) 313 8 (2.3) 36 (-0.5) Cambodia 323 48 (-0.6) 15 (0.8) 195 68 (-1.9) 3 (1.9) Myanmar 400 49 (-0.7) 59 (-1.8) 215 37 (-0.1) 54 (-1.8) Guinea 476 50 (-0.8) 48 (-1.3) 424 59 (-1.4) 37 (-0.6) Congo, Dem. Rep. 536 51 (-0.9) 38 (-0.5) 405 46 (-0.5) 52 (-1.7) Uzbekistan 222 52 (-0.9) 31 (-0.1) 130 48 (-0.5) 41 (-0.8) Nigeria 720 53 (-1) 52 (-1.4) 723 67 (-1.8) 63 (-2.7) Honduras 188 54 (-1) 30 (-0.1) 124 51 (-0.6) 46 (-1.1) Dominican Republic 279 55 (-1) 22 (0.4) 136 62 (-1.5) 18 (0.7) Chad 776 56 (-1) 69 (-3.1) 643 47 (-0.5) 67 (-3.3) Cote d’Ivoire 542 57 (-1) 55 (-1.5) 525 66 (-1.7) 68 (-3.6) Togo 456 58 (-1) 46 (-1.1) 470 58 (-1.4) 69 (-3.6) Philippines 178 59 (-1) 42 (-0.7) 132 54 (-1) 47 (-1.3) Bolivia 304 60 (-1) 36 (-0.4) 185 63 (-1.5) 31 (-0.1) Benin 493 61 (-1.2) 54 (-1.4) 415 64 (-1.5) 66 (-3.2) Indonesia 253 62 (-1.3) 57 (-1.6) 199 52 (-0.7) 58 (-1.9) Mali 450 63 (-1.4) 43 (-0.7) 407 65 (-1.7) 32 (-0.3) Ghana 471 64 (-1.7) 64 (-2.1) 342 56 (-1.2) 64 (-2.7) Tunisia 175 65 (-1.7) 63 (-2) 82 55 (-1) 59 (-2) Burkina Faso 489 66 (-1.7) 40 (-0.6) 385 60 (-1.4) 48 (-1.3) Bulgaria 241 67 (-1.9) 41 (-0.7) 109 57 (-1.4) 50 (-1.4) Viet Nam 314 68 (-2) 61 (-1.9) 112 70 (-2.3) 60 (-2.1) Kenya 592 69 (-2) 60 (-1.9) 443 71 (-3.3) 53 (-1.7) Mexico 340 70 (-2.9) 68 (-2.7) 121 69 (-2.1) 57 (-1.9) Afghanistan 530 71 (-3.5) 12 (1.2) 256 45 (-0.5) 39 (-0.6) 50-69 years Males Females Mortality rate in Performance ranking Mortality rate in Performance ranking 2019 (per 100,000 (performance*) in 2019 (per 100,000 (performance*) in Country (ordered based on ranking during 2010-19) population) 2010-19 2000-10 population) 2010-19 2000-10 Haiti 2271 1 (4) 71 (-4.9) 1473 1 (5.8) 71 (-6.6) Iraq 1503 2 (2.7) 64 (-0.7) 1043 13 (1.3) 63 (-1.4) South Africa 2453 3 (2.3) 43 (0.4) 1554 10 (1.7) 32 (0.1) Tanzania 1972 4 (1.8) 66 (-0.9) 1463 8 (1.8) 55 (-1) Jordan 1076 5 (1.6) 19 (0.9) 608 3 (2.1) 18 (0.7) Zambia 2218 6 (1.3) 10 (1.5) 1581 4 (2) 8 (1.2) Mozambique 2626 7 (1.3) 70 (-1.7) 1803 2 (2.2) 70 (-2.5) 142 CHAPTER 5 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E 50-69 years (continued) Males Females Mortality rate in Performance ranking Mortality rate in Performance ranking 2019 (per 100,000 (performance*) in 2019 (per 100,000 (performance*) in Country (ordered based on ranking during 2010-19) population) 2010-19 2000-10 population) 2010-19 2000-10 Bangladesh 1396 8 (1.1) 68 (-1.3) 814 7 (1.9) 26 (0.3) Algeria 861 9 (1.1) 1 (3.1) 590 6 (1.9) 1 (2) Peru 1013 10 (0.8) 23 (0.8) 690 15 (0.9) 46 (-0.4) Russian Federation 2437 11 (0.8) 8 (1.5) 962 28 (0.2) 20 (0.6) Azerbaijan 1478 12 (0.8) 2 (2.8) 815 27 (0.2) 5 (1.6) Türkiye 1222 13 (0.8) 9 (1.5) 550 20 (0.6) 13 (1) Afghanistan 2120 14 (0.7) 25 (0.7) 1479 23 (0.4) 48 (-0.5) Iran 1092 15 (0.6) 3 (2.4) 633 12 (1.3) 3 (1.7) Sierra Leone 2122 16 (0.5) 16 (1.1) 1708 16 (0.8) 34 (0) Laos 1829 17 (0.4) 15 (1.1) 1251 30 (0) 22 (0.5) Zimbabwe 3044 18 (0.4) 33 (0.6) 2194 14 (1) 64 (-1.5) Belarus 2545 19 (0.2) 38 (0.5) 855 29 (0) 40 (-0.1) Ethiopia 1757 20 (0.2) 49 (0) 1222 19 (0.7) 16 (0.7) Ukraine 2416 21 (0.2) 20 (0.9) 861 9 (1.8) 50 (-0.5) Sri Lanka 1590 22 (0.2) 45 (0.3) 687 55 (-0.8) 24 (0.4) Burundi 2043 23 (0.2) 12 (1.4) 1615 21 (0.6) 38 (-0.1) Ecuador 957 24 (0.1) 21 (0.8) 669 38 (-0.2) 43 (-0.3) Tajikistan 1625 25 (0.1) 14 (1.3) 1067 43 (-0.3) 2 (1.7) Cameroon 2209 26 (0.1) 59 (-0.4) 1862 41 (-0.3) 62 (-1.4) Papua New Guinea 2338 27 (0.1) 40 (0.4) 1349 17 (0.8) 53 (-0.7) Morocco 1402 28 (0) 5 (1.9) 830 11 (1.5) 15 (0.8) India 1685 29 (0) 29 (0.6) 1291 70 (-1.8) 25 (0.3) Uganda 2535 30 (0) 22 (0.8) 1811 24 (0.4) 35 (0) Brazil 1338 31 (0) 42 (0.4) 686 22 (0.6) 17 (0.7) Kazakhstan 2308 32 (-0.1) 11 (1.4) 1099 35 (-0.2) 41 (-0.3) Serbia 1735 33 (-0.1) 24 (0.7) 836 26 (0.2) 23 (0.4) Senegal 1748 34 (-0.1) 26 (0.6) 1077 18 (0.7) 12 (1) Egypt 2031 35 (-0.2) 67 (-0.9) 1247 56 (-0.8) 54 (-0.9) Guinea 1919 36 (-0.2) 58 (-0.4) 1667 54 (-0.7) 42 (-0.3) Nepal 1902 37 (-0.2) 39 (0.5) 1331 44 (-0.5) 44 (-0.3) Congo, Dem. Rep. 2026 38 (-0.2) 48 (0) 1604 42 (-0.3) 60 (-1.3) Mali 2160 39 (-0.2) 52 (-0.2) 1717 46 (-0.5) 45 (-0.3) Nigeria 2462 40 (-0.3) 62 (-0.6) 2289 63 (-0.9) 67 (-1.9) Malawi 2821 41 (-0.3) 57 (-0.4) 1677 5 (1.9) 58 (-1.1) Chad 2516 42 (-0.3) 69 (-1.5) 2047 33 (-0.1) 66 (-1.9) Argentina 1403 43 (-0.3) 27 (0.6) 716 25 (0.4) 57 (-1) Cambodia 1376 44 (-0.4) 4 (2.3) 1051 67 (-1.3) 6 (1.4) Sudan 1781 45 (-0.4) 41 (0.4) 1251 60 (-0.9) 28 (0.2) Bolivia 2118 46 (-0.5) 55 (-0.3) 1306 51 (-0.7) 39 (-0.1) Cote d’Ivoire 2525 47 (-0.5) 61 (-0.6) 2297 66 (-1.3) 65 (-1.5) COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 5 143 50-69 years (continued) Males Females Mortality rate in Performance ranking Mortality rate in Performance ranking 2019 (per 100,000 (performance*) in 2019 (per 100,000 (performance*) in Country (ordered based on ranking during 2010-19) population) 2010-19 2000-10 population) 2010-19 2000-10 Honduras 1643 48 (-0.5) 35 (0.5) 969 40 (-0.2) 29 (0.2) Togo 2140 49 (-0.5) 56 (-0.4) 1788 45 (-0.5) 68 (-1.9) Guatemala 1467 50 (-0.6) 46 (0.2) 1069 68 (-1.4) 59 (-1.1) Madagascar 1714 51 (-0.6) 17 (1.1) 1274 31 (-0.1) 10 (1.1) Indonesia 1802 52 (-0.6) 34 (0.6) 1183 32 (-0.1) 51 (-0.6) Colombia 953 53 (-0.6) 6 (1.9) 593 49 (-0.6) 4 (1.6) Benin 1903 54 (-0.6) 60 (-0.4) 1587 47 (-0.6) 69 (-1.9) Dominican Republic 1416 55 (-0.6) 50 (0) 816 61 (-0.9) 9 (1.1) Pakistan 1866 56 (-0.7) 63 (-0.7) 1237 59 (-0.9) 30 (0.2) Burkina Faso 2184 57 (-0.7) 53 (-0.2) 1609 52 (-0.7) 56 (-1) Niger 1873 58 (-0.7) 37 (0.5) 1451 34 (-0.1) 37 (-0.1) China 1129 59 (-0.9) 36 (0.5) 557 48 (-0.6) 14 (1) Myanmar 2149 60 (-0.9) 51 (-0.2) 1349 57 (-0.8) 61 (-1.3) Ghana 1849 61 (-1) 54 (-0.2) 1401 53 (-0.7) 47 (-0.4) Uzbekistan 1956 62 (-1) 13 (1.4) 1172 64 (-1) 19 (0.7) Philippines 1536 63 (-1.2) 31 (0.6) 1042 58 (-0.8) 49 (-0.5) Tunisia 1239 64 (-1.2) 28 (0.6) 590 36 (-0.2) 33 (0.1) Kenya 2733 65 (-1.2) 65 (-0.8) 1942 62 (-0.9) 52 (-0.7) Rwanda 2205 66 (-1.2) 7 (1.8) 1551 39 (-0.2) 7 (1.2) Malaysia 1496 67 (-1.3) 47 (0.1) 841 65 (-1.1) 21 (0.5) Bulgaria 2047 68 (-1.4) 32 (0.6) 886 69 (-1.5) 31 (0.2) Mexico 1423 69 (-1.5) 44 (0.4) 886 50 (-0.6) 36 (0) Thailand 1343 70 (-1.7) 30 (0.6) 596 37 (-0.2) 11 (1.1) Viet Nam 1440 71 (-2.2) 18 (0.9) 635 71 (-2.1) 27 (0.3) 70-84 years Males Females Mortality rate in Performance ranking Mortality rate in Performance ranking 2019 (per 100,000 (performance*) in 2019 (per 100,000 (performance*) in Country (ordered based on ranking during 2010-19) population) 2010-19 2000-10 population) 2010-19 2000-10 Haiti 9175 1 (1.7) 71 (-2.8) 7390 1 (2.6) 71 (-3.7) Sri Lanka 6941 2 (1.6) 42 (-0.3) 3792 5 (1.5) 4 (0.9) India 6301 3 (1.4) 53 (-0.5) 5458 21 (0.6) 38 (-0.4) Russian Federation 7560 4 (1.4) 17 (0.3) 4792 9 (1.1) 22 (0) Iraq 7497 5 (1.2) 52 (-0.4) 6176 23 (0.6) 53 (-0.7) Iran 6150 6 (1.2) 1 (1.1) 3905 2 (1.7) 13 (0.5) Türkiye 5929 7 (1.1) 18 (0.3) 3684 6 (1.4) 34 (-0.2) Algeria 5195 8 (1) 3 (0.9) 4537 13 (0.9) 24 (0) Peru 5419 9 (1) 8 (0.6) 3764 7 (1.4) 27 (-0.1) Sudan 6889 10 (0.9) 34 (-0.2) 5809 12 (1) 60 (-0.9) Colombia 4649 11 (0.9) 11 (0.6) 3135 8 (1.3) 8 (0.7) 144 CHAPTER 5 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E 70-84 years (continued) Males Females Mortality rate in Performance ranking Mortality rate in Performance ranking 2019 (per 100,000 (performance*) in 2019 (per 100,000 (performance*) in Country (ordered based on ranking during 2010-19) population) 2010-19 2000-10 population) 2010-19 2000-10 Morocco 6895 12 (0.8) 57 (-0.6) 5054 4 (1.6) 50 (-0.6) Ecuador 4340 13 (0.6) 14 (0.4) 3149 27 (0.4) 20 (0.1) Bulgaria 6815 14 (0.5) 12 (0.5) 4419 10 (1.1) 2 (1) China 5556 15 (0.5) 23 (0.2) 3601 17 (0.7) 9 (0.7) Ethiopia 7449 16 (0.5) 6 (0.7) 6100 16 (0.8) 1 (1.5) Malaysia 5877 17 (0.5) 10 (0.6) 4581 28 (0.4) 11 (0.5) Laos 8399 18 (0.4) 39 (-0.3) 6984 39 (0) 33 (-0.2) Zimbabwe 8061 19 (0.4) 28 (0) 6106 14 (0.9) 40 (-0.4) Tajikistan 8373 20 (0.4) 65 (-1.2) 6866 37 (0) 49 (-0.6) Dominican Republic 5481 21 (0.3) 61 (-0.7) 3841 50 (-0.2) 7 (0.7) Burundi 8116 22 (0.3) 2 (1) 7176 29 (0.4) 14 (0.5) Honduras 7239 23 (0.3) 22 (0.2) 5214 25 (0.5) 23 (0) Egypt 8712 24 (0.3) 67 (-1.4) 6846 59 (-0.5) 63 (-1) Serbia 7387 25 (0.3) 37 (-0.2) 4778 18 (0.7) 46 (-0.5) Belarus 9247 26 (0.3) 69 (-1.7) 4933 24 (0.6) 68 (-1.2) Bangladesh 6926 27 (0.2) 70 (-2.2) 5012 11 (1) 62 (-1) Brazil 5931 28 (0.2) 16 (0.4) 4111 32 (0.2) 18 (0.2) Thailand 4113 29 (0.1) 4 (0.9) 2602 15 (0.8) 5 (0.9) Tanzania 7678 30 (0.1) 30 (-0.1) 5560 19 (0.7) 21 (0.1) South Africa 6871 31 (0.1) 50 (-0.4) 4961 33 (0.2) 44 (-0.5) Argentina 6105 32 (0.1) 25 (0) 3585 30 (0.2) 45 (-0.5) Rwanda 8410 33 (0) 9 (0.6) 7071 26 (0.4) 6 (0.8) Jordan 6432 34 (0) 19 (0.3) 4391 22 (0.6) 12 (0.5) Cambodia 6636 35 (0) 5 (0.8) 5628 55 (-0.3) 3 (0.9) Ukraine 8323 36 (0) 35 (-0.2) 4680 3 (1.7) 48 (-0.5) Kenya 6980 37 (0) 33 (-0.2) 5818 60 (-0.5) 29 (-0.1) Bolivia 9084 38 (0) 54 (-0.5) 7063 44 (-0.1) 41 (-0.4) Malawi 8664 39 (0) 31 (-0.1) 5616 20 (0.7) 17 (0.3) Myanmar 8898 40 (0) 38 (-0.3) 7072 36 (0) 57 (-0.8) Madagascar 7396 41 (0) 48 (-0.4) 6311 41 (-0.1) 31 (-0.2) Azerbaijan 8282 42 (-0.1) 47 (-0.4) 6212 69 (-0.7) 47 (-0.5) Nepal 8060 43 (-0.1) 51 (-0.4) 6770 49 (-0.2) 42 (-0.4) Togo 9527 44 (-0.1) 46 (-0.4) 9072 35 (0.1) 65 (-1.2) Congo, Dem. Rep. 8110 45 (-0.1) 24 (0) 7248 54 (-0.3) 51 (-0.7) Viet Nam 6640 46 (-0.1) 58 (-0.6) 3867 56 (-0.3) 52 (-0.7) Sierra Leone 9466 47 (-0.2) 26 (0) 8389 43 (-0.1) 30 (-0.2) Chad 9794 48 (-0.2) 59 (-0.7) 8742 38 (0) 58 (-0.8) Guinea 8213 49 (-0.2) 62 (-0.7) 7580 62 (-0.5) 37 (-0.3) Guatemala 5312 50 (-0.2) 7 (0.6) 4238 47 (-0.1) 43 (-0.5) Papua New Guinea 9407 51 (-0.2) 55 (-0.5) 7004 42 (-0.1) 61 (-0.9) COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 5 145 70-84 years (continued) Males Females Mortality rate in Performance ranking Mortality rate in Performance ranking 2019 (per 100,000 (performance*) in 2019 (per 100,000 (performance*) in Country (ordered based on ranking during 2010-19) population) 2010-19 2000-10 population) 2010-19 2000-10 Cameroon 7953 52 (-0.2) 49 (-0.4) 7592 66 (-0.6) 55 (-0.8) Philippines 7483 53 (-0.2) 32 (-0.1) 6167 48 (-0.2) 39 (-0.4) Zambia 9090 54 (-0.2) 29 (-0.1) 7412 46 (-0.1) 25 (0) Kazakhstan 9169 55 (-0.2) 13 (0.5) 6547 40 (0) 56 (-0.8) Mozambique 10186 56 (-0.2) 56 (-0.6) 8465 31 (0.2) 54 (-0.8) Cote d’Ivoire 8452 57 (-0.2) 41 (-0.3) 6860 45 (-0.1) 36 (-0.3) Mali 9678 58 (-0.3) 36 (-0.2) 8882 65 (-0.5) 19 (0.1) Uganda 10496 59 (-0.3) 63 (-0.7) 9373 63 (-0.5) 67 (-1.2) Ghana 7825 60 (-0.3) 64 (-0.8) 7010 61 (-0.5) 66 (-1.2) Indonesia 8283 61 (-0.3) 40 (-0.3) 6260 53 (-0.3) 35 (-0.3) Afghanistan 8730 62 (-0.3) 21 (0.3) 7194 52 (-0.3) 28 (-0.1) Pakistan 8329 63 (-0.4) 60 (-0.7) 6660 64 (-0.5) 26 (-0.1) Niger 8924 64 (-0.4) 15 (0.4) 8051 51 (-0.2) 15 (0.4) Benin 8038 65 (-0.4) 43 (-0.3) 7355 68 (-0.6) 64 (-1) Burkina Faso 9944 66 (-0.5) 20 (0.3) 8720 58 (-0.5) 32 (-0.2) Senegal 8854 67 (-0.5) 44 (-0.3) 6985 34 (0.1) 16 (0.4) Nigeria 9364 68 (-0.5) 27 (0) 9121 71 (-0.8) 59 (-0.9) Tunisia 6478 69 (-0.5) 45 (-0.3) 3817 57 (-0.4) 10 (0.6) Mexico 5519 70 (-0.6) 66 (-1.3) 4114 67 (-0.6) 70 (-1.8) Uzbekistan 9451 71 (-0.7) 68 (-1.5) 7356 70 (-0.8) 69 (-1.6) 146 CHAPTER 5 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E FIGURE 5A.3 Correlation between performance in mortality decline during 2010-19 among males and females TABLE 5A.3  Change in performance in the annual rate of decline in mortality under 15 years, 15-49 years, 50-69 years, and 70-84 years, by sex from 2000-10 decade to 2010-2019 decade a. Under 15 years Change in performance in annual rate of decline of mortality under 15 years (% per year) Males (change in performance per year) Females (change in performance per year) 7.5 and above (best transition) Haiti (14.4) Haiti (17.7) Azerbaijan (7.8) Ukraine (7.7) 5.0 to 7.4 Azerbaijan (7.4) Russian Federation (7.0) Bulgaria (6.5) Kazakhstan (6.5) Ukraine (6.2) South Africa (5.0) Russian Federation (5.9) Kazakhstan (5.5) Zimbabwe (5.0) 2.5 to 4.9 South Africa (4.7) Belarus (4.9) Burundi (4.4) Burundi (4.8) Belarus (3.9) Zimbabwe (4.8) Iran (3.5) Iran (4.0) China (3.4) Bulgaria (4.0) Algeria (3.0) Algeria (3.3) Iraq (2.7) China (3.3) Jordan (2.7) Chad (2.5) COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 5 147 a. Under 15 years (continued) Change in performance in annual rate of decline of mortality under 15 years (% per year) Males (change in performance per year) Females (change in performance per year) >0 to 2.4 Chad (2.3) Jordan (2.4) Cameroon (2.3) Cameroon (2.3) Ecuador (2.2) Ecuador (2.3) Sri Lanka (2.2) Sri Lanka (1.8) Uzbekistan (1.7) Iraq (1.7) Egypt (1.6) Morocco (1.5) Thailand (1.5) Argentina (1.5) Argentina (1.4) Thailand (1.3) Pakistan (1.2) Uzbekistan (1.3) Morocco (1.2) Sudan (1.3) Sudan (1.1) India (1.2) Malawi (1.1) Ghana (1.2) India (0.9) Egypt (1.2) Indonesia (0.8) Togo (1.1) Mali (0.8) Pakistan (0.9) Ghana (0.7) Indonesia (0.9) Togo (0.4) Mali (0.6) Tunisia (0.4) Sierra Leone (0.4) Papua New Guinea (0.3) Tunisia (0.4) Sierra Leone (0.2) Papua New Guinea (0.4) Mozambique (0.2) Congo, Dem. Rep. (0.3) Benin (0.1) Afghanistan (0.3) Viet Nam (0.0) Benin (0.3) Afghanistan (0.0) Mozambique (0.2) Philippines (0.2) Malawi (0.1) Senegal (0.1) Myanmar (0.1) -2.4 to <0 Laos (0.0) Nigeria (0.0) Congo, Dem. Rep. (-0.1) Burkina Faso (0.0) Nigeria (-0.1) Viet Nam (-0.2) Myanmar (-0.1) Dominican Republic (-0.2) Burkina Faso (-0.2) Laos (-0.4) Mexico (-0.7) Tanzania (-0.2) Peru (-0.8) Senegal (-0.2) Guatemala (-0.8) Philippines (-0.4) Tanzania (-1.0) Guatemala (-0.4) Cote d’Ivoire (-1.0) Dominican Republic (-0.6) Colombia (-1.2) Mexico (-0.7) Ethiopia (-1.3) Peru (-0.8) Madagascar (-1.3) Uganda (-0.8) Tajikistan (-1.4) Cote d’Ivoire (-0.9) Bolivia (-1.4) Colombia (-1.1) Bangladesh (-1.5) Tajikistan (-1.2) Guinea (-1.6) Madagascar (-1.2) Türkiye (-1.7) Kenya (-1.2) Uganda (-1.8) Ethiopia (-1.2) Honduras (-1.8) Guinea (-1.3) Kenya (-2.1) Bangladesh (-1.3) Zambia (-2.2) Türkiye (-1.8) Bolivia (-1.8) Honduras (-1.9) Zambia (-2.0) -4.9 to -2.5 Serbia (-2.5) Serbia (-2.6) Nepal (-2.7) Niger (-2.8) Niger (-2.8) Nepal (-3.1) Cambodia (-2.8) Brazil (-3.5) Brazil (-3.3) Cambodia (-3.6) Rwanda (-3.4) Malaysia (-4.2) Malaysia (-3.4) Rwanda (-4.2) 148 CHAPTER 5 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E Change in performance in annual rate of decline of mortality under 15 years (% per year) Males (change in performance per year) Females (change in performance per year) -7.4 to -5.0 -7.5 and below (worst transition) Distribution (%) Median= 0.20 Median= 0.31 40 45 Mean = 0.88 Mean = 0.87 35 40 35 30 Distribution (%) 30 Distribution (%) 25 25 20 20 15 15 10 10 5 5 0 0 -7.5 and -7.5 to -5.0 -4.9 to -2.5 -2.4 to <0 >0 to 2.4 2.5 to 4.9 5.0 to 7.4 7.5+ -7.5 and -7.5 to -5.0 -4.9 to -2.5 -2.4 to <0 >0 to 2.4 2.5 to 4.9 5.0 to 7.4 7.5+ -5 below -5 below Change in performance between 2000-10 and 2010-19 Change in performance between 2000-10 and 2010-19 Countries in italics are those that encountered major crises such as hurricanes, earthquakes, flooding, tsunami, or war during 2000, 2005, 2010, 2015, or 2019, thereby showing a pseudo good or poor in performance during 2010-2019 relative to during 2000-2010. b. 15-49 years Change in performance in annual Males (change in performance per year) Females (change in performance per year) rate of decline of mortality between 15 and 49 years (% per year) 7.5 and above (best transition) Haiti (16.9) Haiti (24.7) Iraq (7.8) Mozambique (7.5) 5.0 to 7.4 Zimbabwe (7.2) South Africa (6.3) 2.5 to 4.9 Tanzania (4.5) Iraq (4.7) Zimbabwe (4.4) Iran (3.7) Mozambique (4.2) Tanzania (3.2) South Africa (3.3) Malawi (2.8) Bangladesh (2.8) Chad (2.7) >0 to 2.4 Ecuador (2.2) Togo (2.3) Iran (2.1) Burundi (2.2) Chad (2.1) Azerbaijan (2.1) Pakistan (2.0) Papua New Guinea (2) Tajikistan (2.0) Ukraine (1.9) Belarus (2.0) Cote d’Ivoire (1.8) India (1.9) Myanmar (1.8) Azerbaijan (1.7) Belarus (1.7) Jordan (1.6) Benin (1.6) Guatemala (1.6) Kazakhstan (1.6) Egypt (1.4) Ghana (1.5) Myanmar (1.0) Cameroon (1.5) Papua New Guinea (1.0) Ecuador (1.4) China (1.0) Thailand (1.4) Peru (0.9) Zambia (1.3) Brazil (0.9) Indonesia (1.2) Türkiye (0.8) Uganda (1.2) Cameroon (0.8) Congo, Dem. Rep. (1.2) Thailand (0.8) Jordan (1.1) Laos (0.7) Tajikistan (1.1) Russian Federation (0.6) Tunisia (1.0) Sri Lanka (0.5) Nigeria (0.9) Cote d’Ivoire (0.5) Serbia Argentina (0.8) (0.5) Peru (0.8) Guinea (0.5) Morocco (0.8) Nigeria (0.4) Kazakhstan Türkiye (0.7) (0.4) Bangladesh (0.7) COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 5 149 b. 15-49 years (continued) Change in performance in annual Males (change in performance per year) Females (change in performance per year) rate of decline of mortality between 15 and 49 years (% per year) >0 to 2.4 (continued) Ghana (0.4) Sudan (0.5) Tunisia (0.3) China (0.5) Uganda (0.3) Honduras (0.5) Indonesia (0.3) Sri Lanka (0.4) Benin (0.2) Russian Federation (0.4) Ukraine (0.1) Senegal (0.3) Togo (0.1) Egypt (0.3) Malaysia (0.1) Brazil (0.3) Philippines (0.3) Serbia (0.3) Uzbekistan (0.2) Algeria (0.2) Sierra Leone (0.2) Afghanistan (0.1) Bulgaria (0.1) -2.4 to <0 Ethiopia (0.0) Burkina Faso (-0.1) Nepal (0.0) Mexico (-0.2) Viet Nam (-0.1) Viet Nam (-0.2) Kenya (-0.1) Nepal (-0.3) Morocco (-0.1) Ethiopia (-0.4) Zambia (-0.2) Niger (-0.5) Mexico (-0.2) India (-0.6) Argentina (-0.2) Laos (-0.7) Sudan (-0.2) Guinea (-0.8) Madagascar (-0.3) Malaysia (-0.9) Philippines (-0.4) Guatemala (-0.9) Congo, Dem. Rep. (-0.4) Madagascar (-1.0) Mali (-0.7) Bolivia (-1.4) Bolivia (-0.7) Mali (-1.5) Algeria (-0.7) Kenya (-1.6) Niger (-0.7) Colombia (-1.7) Uzbekistan (-0.8) Dominican Republic (-2.2) Honduras (-0.9) Pakistan (-2.4) Malawi (-1.0) Sierra Leone (-1.1) Burkina Faso (-1.1) Senegal (-1.2) Bulgaria (-1.2) Burundi (-1.3) Dominican Republic (-1.4) Cambodia (-1.5) Rwanda (-2.0) Colombia (-2.2) -4.9 to -2.5 Afghanistan (-4.7) Rwanda (-3.6) Cambodia (-3.9) -7.4 to -5.0 -7.5 and below (worst transition) Distribution (%) Median= 0.34 Mean = 0.73 65 Median= 0.71 55 55 Mean = 1.13 45 45 35 Distribution (%) 35 Distribution (%) 25 25 15 15 5 5 -5 -5 -7.5 and -7.5 to -5.0 -4.9 to -2.5 -2.4 to <0 >0 to 2.4 2.5 -7.5 and -7.5 to -5.0 -4.9 to -2.5 -2.4 to <0 >0 to 2.4 2.5 Change in performance between 2000-10 and 2010-19 Change in performance between 2000-10 and 2010-19 Countries in italics are those that encountered major crises such as hurricanes, earthquakes, flooding, tsunami, or war during 2000, 2005, 2010, 2015, or 2019, thereby showing a pseudo good or poor in performance during 2010-2019 relative to during 2000-2010. 150 CHAPTER 5 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E c. 50-69 years Change in performance in annual Males (change in performance per year) Females (change in performance per year) rate of decline of mortality between 50 and 69 years (% per year) 7.5 and above (best transition) Haiti (8.9) Haiti (12.4) 5.0 to 7.4 2.5 to 4.9 Iraq (3.4) Mozambique (4.7) Mozambique (2.9) Malawi (3) Tanzania (2.7) Tanzania (2.8) Iraq (2.7) >0 to 2.4 Bangladesh (2.4) Zimbabwe (2.4) South Africa (1.9) Ukraine (2.4) Chad (1.2) Chad (1.8) Egypt (0.8) South Africa (1.6) Jordan (0.7) Bangladesh (1.5) Cameroon (0.5) Papua New Guinea (1.4) Nigeria (0.3) Jordan (1.4) Ethiopia (0.2) Togo (1.4) Guinea (0.2) Benin (1.4) Malawi (0.1) Argentina (1.4) Peru (0.1) Peru (1.3) Cote d’Ivoire (0.1) Cameroon (1.1) Pakistan (0.0) Congo, Dem. Rep. (1.0) Afghanistan (0.0) Nigeria (1.0) Afghanistan (0.9) Morocco (0.8) Sierra Leone (0.8) Zambia (0.7) Burundi (0.7) Uganda (0.5) Indonesia (0.5) Myanmar (0.5) Burkina Faso (0.3) Cote d’Ivoire (0.2) Egypt (0.2) Belarus (0.1) Ecuador (0.1) Kazakhstan (0.1) Niger (0.0) -2.4 to <0 Mali (0.0) Ethiopia (-0.1) Bolivia (-0.1) Algeria (-0.1) Sri Lanka (-0.1) Brazil (-0.1) Togo (-0.2) Nepal (-0.1) Benin (-0.2) Serbia (-0.2) Zambia (-0.2) Mali (-0.2) Zimbabwe (-0.2) Kenya (-0.2) Congo, Dem. Rep. (-0.3) Tunisia (-0.3) Belarus (-0.3) Ghana (-0.3) Papua New Guinea (-0.4) Senegal (-0.3) Kenya (-0.4) Philippines (-0.3) Brazil (-0.4) Türkiye (-0.4) Burkina Faso (-0.5) Guatemala (-0.4) Sierra Leone (-0.5) Russian Federation (-0.4) Dominican Republic (-0.6) Iran (-0.4) India (-0.6) Honduras (-0.4) Ecuador (-0.6) Guinea (-0.5) Russian Federation (-0.7) Laos (-0.5) Laos (-0.7) Bolivia (-0.6) Myanmar (-0.7) Mexico (-0.6) Nepal (-0.7) Pakistan (-1.0) Türkiye (-0.7) Sudan (-1.1) Guatemala (-0.7) Sri Lanka (-1.1) Ghana (-0.7) Madagascar (-1.1) COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 5 151 c. 50-69 years (continued) Change in performance in annual Males (change in performance per year) Females (change in performance per year) rate of decline of mortality between 50 and 69 years (% per year) -2.4 to <0 (continued) Ukraine (-0.7) Thailand (-1.3) Senegal (-0.8) Azerbaijan (-1.4) Uganda (-0.8) Rwanda (-1.5) Serbia (-0.8) China (-1.6) Sudan (-0.8) Malaysia (-1.6) Argentina (-0.9) Bulgaria (-1.7) Honduras (-1.1) Uzbekistan (-1.7) Indonesia (-1.1) Dominican Republic (-2) Tajikistan (-1.2) Tajikistan (-2.1) Niger (-1.3) India (-2.1) Burundi (-1.3) Colombia (-2.2) Malaysia (-1.3) Viet Nam (-2.4) China (-1.4) Kazakhstan (-1.5) Madagascar (-1.6) Philippines (-1.7) Tunisia (-1.8) Mexico (-1.8) Iran (-1.8) Morocco (-1.9) Bulgaria (-2.0) Azerbaijan (-2.0) Algeria (-2.0) Thailand (-2.3) Uzbekistan (-2.4) Colombia (-2.5) -4.9 to -2.5 Cambodia (-2.7) Cambodia (-2.7) Rwanda (-3.0) Viet Nam (-3.1) -7.4 to -5.0 -7.5 and below (worst tran- sition) Distribution (%) 75 Median= -0.68 Mean = -0.45 65 Median= -0.07 Mean = 0.26 65 55 55 45 45 Distribution (%) 35 Distribution (%) 35 25 25 15 15 5 5 -5 -5 -7.5 and -7.5 to -5.0 -4.9 to -2.5 -2.4 to <0 >0 to 2.4 2.5 -7.5 and -7.5 to -5.0 -4.9 to -2.5 -2.4 to <0 >0 to 2.4 2.5 Change in performance between 2000-10 and 2010-19 Change in performance between 2000-10 and 2010-19 Countries in italics are those that encountered major crises such as hurricanes, earthquakes, flooding, tsunami, or war during 2000, 2005, 2010, 2015, or 2019, thereby showing a pseudo good or poor in performance during 2010-2019 relative to during 2000-2010. d. 70-84 years Change in performance in annual Males (change in performance per year) Females (change in performance per year) rate of decline of mortality between 70 and 84 years (% per year) 7.5 and above (best transition) 5.0 to 7.4 Haiti (6.3) 2.5 to 4.9 Haiti (4.5) >0 to 2.4 Bangladesh (2.4) Morocco (2.3) Belarus (1.9) Ukraine (2.2) 152 CHAPTER 5 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E d. 70-84 years (continued) Change in performance in annual Males (change in performance per year) Females (change in performance per year) rate of decline of mortality between 70 and 84 years (% per year) >0 to 2.4 (continued) Sri Lanka (1.9) Bangladesh (2) India (1.9) Sudan (1.9) Egypt (1.7) Belarus (1.8) Iraq (1.7) Türkiye (1.6) Tajikistan (1.6) Peru (1.4) Morocco (1.4) Zimbabwe (1.3) Sudan (1.1) Iraq (1.3) Russian Federation (1.1) Togo (1.3) Dominican Republic (1) Serbia (1.2) Türkiye (0.8) Mexico (1.2) Uzbekistan (0.8) Iran (1.2) Laos (0.7) Russian Federation (1.0) Mexico (0.6) India (1.0) South Africa (0.5) Mozambique (1.0) Serbia (0.5) Algeria (0.9) Ghana (0.5) Papua New Guinea (0.8) Guinea (0.5) Chad (0.8) Viet Nam (0.5) Myanmar (0.8) Chad (0.5) Uzbekistan (0.8) Bolivia (0.5) Kazakhstan (0.7) Zimbabwe (0.5) Argentina (0.7) Uganda (0.4) Ghana (0.7) Nepal (0.4) Uganda (0.7) Papua New Guinea (0.4) South Africa (0.7) Peru (0.4) Tanzania (0.6) Madagascar (0.4) Tajikistan (0.6) Mozambique (0.4) Colombia (0.6) Pakistan (0.3) Sri Lanka (0.6) China (0.3) Egypt (0.5) Colombia (0.3) Honduras (0.5) Azerbaijan (0.3) Viet Nam (0.4) Togo (0.3) Benin (0.4) Myanmar (0.2) Congo, Dem. Rep. (0.4) Ukraine (0.2) Malawi (0.4) Cameroon (0.2) Ecuador (0.3) Tanzania (0.2) Bolivia (0.3) Kenya (0.2) Guatemala (0.3) Ecuador (0.1) Nepal (0.3) Honduras (0.1) Philippines (0.2) Iran (0.1) Laos (0.2) Algeria (0.1) Cote d’Ivoire (0.2) Malawi (0.1) Cameroon (0.2) Cote d’Ivoire (0.1) Madagascar (0.1) Argentina (0.0) Sierra Leone (0.1) Indonesia (0.0) China (0.1) Bulgaria (0.1) Jordan (0.1) Nigeria (0.0) Indonesia (0.0) Brazil (0.0) -2.4 to <0 Bulgaria (0.0) Thailand (-0.1) Mali (0.0) Guinea (-0.1) Benin (-0.1) Burundi (-0.1) Philippines (-0.1) Zambia (-0.1) Malaysia (-0.1) Malaysia (-0.1) Zambia (-0.1) Afghanistan (-0.1) Congo, Dem. Rep. (-0.1) Azerbaijan (-0.2) Senegal (-0.2) Burkina Faso (-0.2) Tunisia (-0.2) Senegal (-0.3) Sierra Leone (-0.2) Kenya (-0.3) Brazil (-0.2) Rwanda (-0.3) Jordan (-0.2) Pakistan (-0.4) COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 5 153 d. 70-84 years (continued) Change in performance in annual Males (change in performance per year) Females (change in performance per year) rate of decline of mortality between 70 and 84 years (% per year) -2.4 to <0 (continued) Ethiopia (-0.3) Niger (-0.6) Nigeria (-0.5) Mali (-0.6) Afghanistan (-0.6) Ethiopia (-0.7) Rwanda (-0.6) Dominican Republic (-0.9) Burundi (-0.7) Tunisia (-1.0) Burkina Faso (-0.7) Cambodia (-1.2) Kazakhstan (-0.8) Niger (-0.8) Guatemala (-0.8) Cambodia (-0.8) Thailand (-0.8) -4.9 to -2.5 -7.4 to -5.0 -7.5 and below (worst transition) Distribution (%) 75 Median= 0.15 Mean = 0.27 85 Median= 0.30 Mean = 0.46 65 75 55 65 45 55 Distribution (%) Distribution (%) 45 35 35 25 25 15 15 5 5 -5 -5 -7.5 and -7.5 to -5.0 -4.9 to -2.5 -2.4 to <0 >0 to 2.4 2.5 -7.5 and -7.5 to -5.0 -4.9 to -2.5 -2.4 to <0 >0 to 2.4 2.5 Change in performance between 2000-10 and 2010-19 Change in performance between 2000-10 and 2010-19 Countries in italics are those that encountered major crises such as hurricanes, earthquakes, flooding, tsunami, or war during 2000, 2005, 2010, 2015, or 2019, thereby showing a pseudo good or poor in performance during 2010-2019 relative to during 2000-2010. ANNEX 5.3 Model using to estimate critical income We used the methods of Hum and others (2012) for (kinc), and maximal survival rates, such that: estimating critical income. We used country specific population by age groups (both sexes), and country specific deaths by age-groups, from World Population Prospects 2022 (United Nations, Department of Eco- nomic and Social Affairs, Population Division 2022), to derive a survival rate for children aged 0 to 14, adults 15 to 49, and seniors 50 to 69. For this analysis, we included only countries with a population of over 7 million (which covers 99.9 percent of the world pop- ulation in 2019). We also calculated an average mid- We used a mixed effect model to calculate the global, point year (1992.5 and 2017, used a five- and four-year as well as country-level, critical incomes for all coun- average). GDP per capita (PPP, constant $2017) was tries in the analysis. We also calculated the income sourced from the World Bank (World Bank 2023). required to achieve 80 percent of the maximal health The five- and four-year averages were used to limit the in high income countries—which is mathematically, influence of sudden, dramatic, changes in health or four times the critical income. Using the country-lev- economic development in that country. el critical incomes derived from the mixed-effects We adapted the Michaelis-Menten model model, we ranked the top countries with the lowest for age-specific global critical income estimates critical income values for each age-grouping. 154 CHAPTER 5 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E TABLE 5A.4  Deaths avertable globally in 2019 from specific causes of deaths and age groups if rate of decline was same as that from HIV/AIDS (ages 0-14, 15-49) and tuberculosis (ages 50-69, 70+) Male Female Reference Actual Deaths Reference Actual Deaths Cause of Death (CoD) Target Age group Reference CoD + Age group AARR (%) AARR (%) avertable AARR (%) AARR (%) avertable Malaria 15-49 years Malaria, 0-14 years* 4.2 -0.3 17,995 4.4 0.4 10,144 Tobacco attribut- 50-69 years Tuberculosis, 50-69 years 4.8 1.3 472,818 4.6 0.6 186,517 able cancers 70+ years Tuberculosis, 70+ years 5.2 0.6 493,762 4.6 0.1 246,112 Infection attribut- 50-69 years Tuberculosis, 50-69 years 4.8 2.7 70,734 4.6 1.3 127,536 able cancers 70+ years Tuberculosis, 70+ years 5.2 1.6 67,814 4.6 1.1 99,394 Stomach cancer 50-69 years Tuberculosis, 50-69 years 4.8 2.5 90,323 4.6 2.6 36,181 70+ years Tuberculosis, 70+ years 5.2 1.9 119,737 4.6 2.4 56,648 Breast cancer 50-69 years Tuberculosis, 50-69 years 4.6 0.5 154,974 70+ years Tuberculosis, 70+ years 4.6 0.3 122,670 CVD (not Stroke) 50-69 years Tuberculosis, 50-69 years 4.8 1.4 1,053,736 4.6 1.6 535,396 70+ years Tuberculosis, 70+ years 5.2 0.7 1,994,677 4.6 1.0 2,051,883 Stroke 50-69 years Tuberculosis, 50-69 years 4.8 2.0 462,694 4.6 2.5 261,105 70+ years Tuberculosis, 70+ years 5.2 1.6 940,435 4.6 2.0 883,607 Respiratory diseases 50-69 years Tuberculosis, 50-69 years 4.8 2.8 208,510 4.6 2.6 129,185 (COPD, Asthma, 70+ years Tuberculosis, 70+ years 5.2 2.0 761,899 4.6 1.9 581,277 other) Road Injury 15-49 years HIV/AIDS, 15-49 years 4.1 1.2 235,125 5.4 2.1 58,207 50-69 years Tuberculosis, 50-69 years 4.8 0.3 142,057 4.6 0.7 49,126 70+ years Tuberculosis, 70+ years 5.2 0.6 58,194 4.6 0.8 28,550 Falls 50-69 years Tuberculosis, 50-69 years 4.8 0.4 58,962 4.6 0.6 28,721 70+ years Tuberculosis, 70+ years 5.2 -1.0 124,481 4.6 -0.8 145,487 Drowning 0-14 years HIV/AIDS, 0-14 years 6.2 4.4 16,255 6.1 4.9 7,687 Suicide 15-49 years HIV/AIDS, 15-49 years 4.1 2.0 99,838 5.4 3.6 39,369 50-69 years Tuberculosis, 50-69 years 4.8 2.4 50,435 4.6 2.1 22,584 70+ years Tuberculosis, 70+ years 5.2 1.9 32,751 4.6 2.5 13,146 Deaths avertable in 2019, 0-69 years 2,979,482 1,646,732 Deaths avertable in 2019, 70+ years 4,593,750 4,228,774 Total deaths avertable in 2019 7,573,232 5,875,506 *The reference group of malaria at ages 0-14 years was used for malaria deaths at ages 15-49 years, based on relevancy. ANNEX REFERENCES 1. Bernal, J. L., S. Cummins, and A. Gasparrini. 2017. “Interrupted time Reeves. 2015. “Regression based quasi-experimental approach series regression for the evaluation of public health interventions: when randomisation is not an option: Interrupted time series A tutorial.” International Journal of Epidemiology 46 (1), 348–55. analysis.” BMJ 350, h2750. 2. Hum, R. J., P. Jha, A. M. McGahan, and Y. –L. Cheng. 2012. “Global 5. United Nations, Department of Economic and Social Affairs, divergence in critical income for adult and childhood survival: Population Division. 2022. World Population Prospects 2022. Analyses of mortality using Michaelis–Menten.” ELife 1: e00051. New York: UN Department of Economic and Social Affairs, Pop- 3. Jamison, D. T., S. M. Murphy, and M. E. Sandbu. 2016. “Why has ulation Division. under-5 mortality decreased at such different rates in different 6. Verguet, S., and D. T. Jamison. 2014. “Estimates of performance countries?” Journal of Health Economics 48, 16–25. in the rate of decline of under-five mortality for 113 low- and 4. Kontopantelis, E., T. Doran, D. A. Springate, I. Buchan, and D. middle-income countries, 1970-2010.” Health Policy and Plan- COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 5 155 ning 29(2), 151–163. bank-country-and-lending-groups. 7. World Bank. 2021. “World Bank Country and Lending Groups.” 8. World Bank. 2023. “World Development Indicators.” Washing- Washington, D.C., The World Bank Group. https://datahelp- ton, D.C., The World Bank Group https://databank.worldbank. desk.worldbank.org/knowledgebase/articles/906519-world- org/source/world-development-indicators. 156 CHAPTER 6 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E Behavior Change and Healthy Longevity Chapter 6 Behavioral Aspects of Healthy Longevity Chapter 7 Taxation of Harmful Products, including Tobacco, Alcohol, and Sugar-Sweetened Beverages (SSBs), and Related Topics 6 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 6 157 Behavioral Aspects of Healthy Longevity Ana Maria Rojas a,b, Ana Maria Munoz Boudet a,b, Ellen Moscoe a,b, Julian Jamison c, and Carlos Riumallo Herl a,b a World Bank, Poverty and Equity Global Practice b Erasmus University, Rotterdam c University of Exeter 158 CHAPTER 6 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E INTRODUCTION Poor health in old age comes at a high cost for individuals, households, and societies. Increases in life expectancy across countries have not always added years of healthy life. Accordingly, health promotion and disease prevention, particularly as they relate to preventable non-communicable diseases (NCDs) whose onset occurs later in life, have become a central focus of health policy. This agenda is challenging, as it relates to people’s lifestyles and habits, which in turn influence NCD incidence and progression. In 2019, the three leading causes of death around duce the prevalence of NCD risk factors (Blaga et the world – cardiovascular diseases, neoplasms, and al. 2018), and that this can be achieved via behavior chronic respiratory diseases – were largely attribut- change. According to the World Health Organiza- able to behavioral risk factors or lifestyle habits (Vos tion (WHO), 80 percent of premature heart disease et al 2020). Habits such as unhealthy eating, smok- and stroke are preventable with either changes in ing, alcohol consumption, and lack of exercise (in individual behavior or with adherence to appropri- addition to delayed and inconsistent engagement ate treatment guidelines (World Health Organiza- with health care) have been linked across many tion 2014). Interventions that address health behav- studies to the development and persistence of risk iors as a means to reduce the burden of NCDs have factors such as hypertension, diabetes, high choles- been identified as among the most cost-effective terol, and high body-mass index (Forouzanfar et al solutions for low- and middle-income countries 2016; Peel et al 2005). These are, in turn, among the (LMICs) (Watkins et al. 2022). Thus, if behaviors top-ranking contributors to unhealthy aging and and in turn their outcomes are modifiable, there is deaths from NCDs (Vos et al 2020). While many an opportunity for policies to address health habits NCDs are not associated with aging, and similarly in the context of promoting healthy longevity. many NCDs are unrelated to health habits and “life- Even in the face of existing knowledge regard- style” factors, this chapter will focus on the group of ing the magnitude of the link between behavioral NCDs that typically develop later in life and can be habits, NCD risk factors, and disease prevention, it prevented or delayed through changes in habits that remains the case that a large share of deaths can be address underlying risk factors. attributed to behavioral risks. One explanation for There is, however, substantial complexity in this is the set of barriers that individuals face when the link between behavioral habits and subsequent attempting to either modify their habits, access care, NCD risk factors. Behavioral habits can lead to or manage their risk factors adequately. multiple risk factors, and specific risk factors can be Traditional neoclassical economics suggests caused by a combination of behavioral habits, and possible explanations such as a lack of relevant in- both are impacted by structural factors linked to formation or large countervailing costs. In turn, unhealthy aging. Dissecting each habit and/or risk policy makers and researchers have often focused factor independently highlights this complexity. on the health system, user fees, or information, in For example, unhealthy eating is associated with a seeking to remove external obstacles and encourage high body-mass index (Rosenheck 2008), hyperten- healthy behaviors. But, while effective, these actions sion (Ibrahim and Damasceno 2012), and diabetes alone have been insufficient to address the full com- (Malik et al. 2010) – each of which is itself a risk plexity of health behavior change and habit forma- factor for cardiovascular diseases. At the same time, tion. Relative ignorance or limited knowledge of hypertension is also influenced by smoking, alco- these issues might once have been widespread and hol consumption, and lack of exercise. Unhealthy undoubtedly still exist among certain population eating is not only dependent on individuals’ taste groups. However, in most countries, the majority preferences or food choices, but also on foods’ avail- of adults are now aware of the health risks of, for ability in their living environments and food costs example, tobacco and alcohol use (e.g., see WHO (Ruhm 2012; Lu and Goldman 2010). Global Adult and Youth Tobacco Surveys). Hence, Despite the complex interactions linking be- lack of knowledge is unlikely to be a driving expla- havioral habits and NCD risk factors, evidence nation of unhealthy behaviors. Similarly, although shows that changes in underlying habits can re- for some people in some situations the actual cost of COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 6 159 a behavior change (e.g., giving up smoking) might add to the understanding on how to modify habits be high enough that the “rational” decision is not that lead to NCDs and support a policy and inter- to do so, this is not the case across all such behav- vention agenda to inform long-term investments in iors, nor for most people. A modest reduction in healthy longevity. Long-run flourishing depends as unhealthy behaviors such as high calorie consump- much on what individuals do or do not do as it does tion, or a modest increase in healthy behaviors such on their interactions with health systems—from as walking more, generally has both a low cost and seeking and obtaining preventive care to treatment often also some direct near-term benefits such as adherence and follow-up. Both domains display the improved energy and sleep, entirely independent of same characteristic of delayed and uncertain con- long-term health outcomes. Hence this is again un- sequences that leave people vulnerable to inaction likely to be a driving explanation, and indeed many and/or repeated deferral. people report that they wish they acted differently Furthermore, by applying a behavioral lens to in these respects (Faries 2016), strongly suggesting groups as well as individuals, elements such as peer that they are not in fact optimizing their utility ac- effects and social norms can be considered. Mental cording to the traditional framework. health is directly affected by, among other factors, Policy efforts towards behavioral and psycho- a sense of community connectedness and of agen- logical obstacles that influence unhealthy lifestyles cy and confidence about the future (not to mention and habit formation are needed. Unlike many other the emotional disposition of the members of one’s causes of morbidity and mortality, NCDs are typi- network). Physical health is at minimum indirectly cally delayed—in the sense that negative outcomes affected by local information flows and self-efficacy: occur years or even decades after the precipitat- only if someone believes that their actions can make ing behaviors—and are often only weakly linked a difference to their well-being, and ideally observes to identifiable individual moments in time. Why this narrative in others, is there an impetus to make would I then change my behavior today for some- healthier decisions. In some sense the behavioral per- thing that might (or might not) happen in the fu- spective centers on how people see themselves in rela- ture? On the other hand, both accidents and com- tion to their worlds, which is crucial for robust aging. municable diseases tend to impact one’s life in the Focusing on NCDs, this study develops a frame- short-run (sometimes immediately) and are, for the work of the relation between factors behind indi- most part, possible to link to a single causal event viduals’ behaviors and the achievement of healthy -which is therefore both salient and clearly amena- longevity at the individual, community, and health ble to prevention. This feedback can and often does systems levels. It then uses the framework to discuss trigger a suite of self-initiated behavioral adjust- policy lessons from the behavioral science literature ments to reduce recurrence. that can aid efforts to reduce the incidence of NCDs Behavioral biases (Tversky and Kahneman, and improve the effectiveness of treatment. These 1974) such as impatience, time inconsistency, and policy lessons are not seen as replacing traditional over-optimism are more congruent with weighting policy tools, such as taxation, but as complements long-term consequences in a mistakenly low man- to increase their efficacy and impact. ner, evaluated by the lights of the individuals them- The chapter is organized as follows. The next selves. Similarly, other relevant individual biases section presents the framework. The following three predict heightened difficulty to assess very small sections apply it to different stages of NCD trajec- probabilities, as well as a tendency to focus dispro- tories (formation, screening, and management) portionately on concrete, identifiable consequences through individual, community-level, and health at the expense of diffuse accumulated outcomes. system perspectives. The last section presents policy Behavioral science insights into psychological recommendations stemming from behaviorally in- barriers, mental models, biases, and other factors at formed intervention studies across different fields. play in individual decision-making and action can 160 CHAPTER 6 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E A BEHAVIORAL SCIENCE FRAMEWORK FOR NCD POLICIES We can identify three main stages in people’s NCD trajectories: (1) formation - when and how the disease starts occurring; (2) detection - when and how it is identified, both by the individual (i.e., symptoms) and medical professionals; and (3) management - the many actions that are required once the disease is identified to slow its progression, manage symptoms, minimize disability, and support the patient’s quality of life. These stages refer to the individual or patient journey (Devi et al. 2020). At each of these stages, we can identify behavioral factors (biases, mental models, beliefs, norms, and others) that are at play at different levels: individual, community, and health system. Drawing on the behavioral science literature, we propose a framework that categorizes the behavioral factors operating at each level and stage (Figure 6.1). FIGURE 6.1 Conceptual model of traditional and behavioral determinants of NCDs and healthy aging Individual factors refer to characteristics includ- not only one of the main elements of policies, and a ing a person’s capabilities, opportunities, and mo- basic requirement for behavior change, but also the tivation (adapting the COM-B approach from main connector between individuals, social contexts, Michie et al. 2011). Our framework assumes that and health systems. Social and behavioral change for (desired) behaviors to take place, a mix of ele- campaigns (SBCC) and health communications ments is required that can be classified under these more broadly are their own category of intervention; three headings. Capabilities primarily encompass therefore we do not focus on SBCC interventions the knowledge and skills needed to engage in the specifically in this study; rather, we include informa- desired behavior. Opportunity includes factors tion-related interventions when they leverage a spe- such as physical and social resources that are gen- cific behavioral insight of interest. The type of infor- erally not controlled by the individual but facilitate mation—including its source, framing, format, and the behavior: for example, having access to health the timing of exposure—will influence the knowl- services or to a market where fruits and vegetables edge it generates in an individual, and thus whether are sold. Motivation is not only the intention to act, information turns into effective communication. but the reasons that direct the individual to act, in- Socio-demographic determinants of health, such cluding goals, the ability to make decisions, emo- as socio-economic status and education, are also tions, habits, identity, and other related mental pro- included among individual factors. While the pro- cesses. These three components are interconnected posed framework, and this chapter, do not focus on and reinforce each other. socio-demographic determinants in detail, they are While information is part of the last set of factors, strong moderators of behaviors and habits. Behavior- given the cross-cutting role it plays, we distinguish it al biases, understood as behavior patterns that do not as a standalone factor. This is because information is conform with expected rational behavior (Gilovich et COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 6 161 al. 2002), for the most part occur at this level. Some of es, social norms and expectations, as well as mental the most salient include cognitive biases such as loss models (how the world is seen) can, instead of con- aversion (weighing a loss more heavily than an equiv- tributing to healthy aging outcomes, inhibit or pre- alent gain) and present bias (impatience, weighing vent them. One channel for such effects is conformity the immediate future disproportionately over more traps that can reinforce health-detrimental actions.2 distant events). Individual factors related to limited For example, some cultures strongly encourage peo- attention are also included at this level. These are fac- ple, often men or youth, to conform to a lifestyle tors which describe the inability to focus on a large centered around drinking, smoking, and other risky set of important decisions and information, including behaviors. In other cases, norms about women’s mo- cognitive overload (Mullainathan and Shafir 2013). bility might restrict their physical exercise or limit Community-level factors refer to social norms, other behaviors that promote health. roles in society, and other factors that structure Health system factors include the structure choices and behavior.1 Individuals are embedded in and characteristics of the health system that in- community and social contexts that provide con- fluence healthy behaviors and the detection and stant flows of information about different behaviors. management of NCDs. People’s interactions with The context and how the community relates to or the health system and health workers are affected views a specific behavior will affect the individual’s by their prior experiences with the system and the capabilities, motivation, and opportunities. We un- behaviors of those who work in it. Relevant factors derstand the community as the network of relevant include health workers’ perceptions about the need social cues for individual decision-making stem- for and potential benefits of a given intervention, ming from reference and influence groups associat- their confidence in their ability to deliver a solu- ed with the individual’s identity and belonging (i.e., tion, and their views about their patients (Abry et their neighborhood, family, and friends, but also al. 2013; Durlak and DuPre 2008). those who share certain core identity elements, for Pricing and provider incentives play a role here, example gender, role in the family, profession, eth- as does choice architecture, or how the system is nicity, and other factors). According to Akerlof and structured to motivate decision-making. Issues of Kranton (2000), people hold multiple identities and bias also affect health care providers. Health work- conforming to them creates intrinsic utility that re- ers, like all people, are subject to the biases, heuris- sults in people’s following a subgroup norm. Norms tics, and mental shortcuts described above, and their evolve as shared practices and identities shaping and decision-making around patient care is impacted by reinforcing perceptions and behaviors. Community these factors. Similarly, norms and mental models influences can stem directly from peers or be more associated with their identity as health care providers diffuse. For the most part, they are either normative are at play as part of the health system environment. or provide a framework for action or decision. Figure 6.2 provides an example of how the Regarding NCD risk and healthy aging, social framework can be applied to the formation, detec- norms and mental models of choice are the two most tion, and management of hypertension. salient channels of community impact. In many cas- 1   Social norms refer to the unwritten rules of behavior that guide social interactions and are critical for policy makers who attempt to shift behavior, as many health behaviors are adopted to conform to social norms. Choice architecture refers to the ways in which structures influence decision making, including health behaviors and other behaviors impacting healthy longevity. 2   Andreoni et al. (2017) define conformity traps as those that keep groups and individuals in a bad equilibrium despite knowledge of inefficiency and preferences for a different status quo. The trap is due to the pressure to conform to the behavior of the majority and the resistance to being the first deviant from the established norm. Hence a norm will persist because of the strength of the social payoff to compliance. 162 CHAPTER 6 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E FIGURE 6.2 Example of specific biases affecting hypertension formation, detection, and management A behavioral focus is not foreign to the health sector. (Cawley and Ruhm 2011), it seems insufficient to see However, our proposed framework expands on what such behaviors as an outcome of trade-offs between has typically been the approach to behavior change the individual’s “stock of health capital” and short-run in health policy and program implementation, increases in welfare or future discounting. In the clas- largely related to information and communications. sic model, if the individual perceives that the benefit Traditional behavior-change interventions in health of smoking is greater than the risk of suffering a heart have centered on making information available to attack or lung cancer in the future, then he or she will encourage appropriate action by individuals, com- be more likely to engage in this risky/unhealthy activi- munities, and health system actors. Our framework ty. Seen in this way, the engagement in risky behaviors enriches this model using a behavioral science lens. should depend solely on the weight that the individual As such, it serves two purposes. First, it allows us to gives to the discounted value of his/her life expectan- unpack and organize what is known about interven- cy or future well-being (Grossman 1972; Cawley and tions that address behavioral barriers related to NCD Ruhm 2011; Lundberg and Shapira 2014). However, formation, detection, and management (and also to this is not the case. Risk perceptions are left to the fu- identify knowledge gaps). Second, it facilitates an as- ture (time inconsistent preferences leading to present sessment of the presence and strength of behavioral bias), while group pressures, social norms, and role elements in health policies. This could enable policy models (what others do) take precedence and moti- makers and program designers to identify interven- vate behavior. Classic tobacco control policies—pro- tions that target different challenges, as well as differ- viding information about risks, imposing smoking ent kinds of potential beneficiaries. For example, in- bans, raising excise taxes, and other measures—in- terventions could be crafted to focus on motivation crease frictions to the behavior by making it more dif- for those groups that already possess capability and ficult and costly. These established policy approaches opportunity factors, while investing in comprehen- can be complemented and made more effective by in- sive packages for those that do not. corporating insights from behavioral science. This is Smoking provides one example of an important the perspective that the next sections take. In them, NCD risk behavior. Many factors that could predict we use concepts from behavioral economic theory the decision to engage in such behaviors have been that describe the biases, mental models, and heuristics extensively studied across different disciplines (Gross- affecting decision-making that are most relevant to man 1972; Becker and Murphy 1988; Orphanides and healthy aging and NCDs.3 The review is not exhaustive Zervos 1995; Cawley and Ruhm 2011). If we under- and does not cover all the possible behavioral factors stand such behaviors as actions, or deliberate inac- at play. Rather, it prioritizes the more prevalent and tions, that affect the individual’s health or that of others those best analyzed in the literature. 3   Annex Table 6A.1 summarizes the relevant biases and provides examples related to aging and NCDs. COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 6 163 NCD FORMATION: HEALTHY LIFESTYLE AND HABIT FORMATION This section examines how behavioral biases influ- affect their health outcomes). The formation of pri- ence health behaviors which affect the formation ors can be explained through two behavioral con- of NCDs. We focus on some of the main risky be- cepts. The first is the “law of small numbers” bias haviors, including physical inactivity, tobacco use, (Rabin 1998, 2002): people’s tendency to assume and alcohol use, but the discussions extend beyond that information derived from observing a small these. Following the framework presented above, number of events reflects the experience of the larg- we disaggregate the policy-relevant factors at work er “population” from which those events are drawn. into health system, community, and individual. The second is the availability heuristic, the reliance on more vivid or memorable evidence to construct Health system factors that affect a prior about the probability of a particular char- NCD formation acteristic or outcome: for example, that a physician is high quality, will exert adequate effort, or would More and more health systems, at least in middle- detect a worrisome health condition (Tversky and and high-income countries, have started focusing Kahneman 1973). Reliance on reports from family on NCD prevention, as well as treatment. Howev- and friends may thus create distortions. If the sys- er, many of these efforts appear to assume that it is tem does not address these issues (and/or dismisses enough to create prevention programs or formally them), they are reinforced. include preventative health in primary health care The health system, in its “external” presentation, models, in order for doctors to deliver such services is largely seen as focused on diseases. This means, effectively and for individuals to use them. This has from an individual point of view, a signal that you not proven to be the case. As discussed with regard only seek healthcare, even if preventative, when you to individual-level factors, providing information have a disease (or symptoms of it). The difficulty of (i.e., generating knowledge) may not by itself be suf- making appointments, the unpleasantness of clin- ficient, because individuals may not act on that in- ics or hospitals where those around you are ill, and formation. People might lack the motivation to ob- many other factors can act as disincentives to seek tain care for a variety of attitudinal reasons such as preventative health care. Countries have started ad- perceived need, distrust in physicians or medical in- dressing some of these priors and biases by creating stitutions, dissatisfaction with previous medical care, “healthy living centers” or similar support structures and fear of being diagnosed with something (Bairey (whether online or in person), that remove the dis- Merz et al. 2017; Friedman 1994; Hall et al. 2001). ease label from preventative health. These obstacles can be exacerbated by system The health system is largely associated with a features (design, focus on disease versus prevention, specific set of individuals: medical professionals. complicated processes) that can increase people’s Physicians, like other persons, fall prey to behavioral “hassle perception” about engaging with the health biases. Partly based on training, they are more prone system, health insurance, and related institutions to mental shortcuts that associate symptoms with and processes (Liebman and Zeckhauser 2008). diseases than to connecting current habits with fu- Markets for medical care are characterized by un- ture events, including NCDs. This reflects a common certainty and by asymmetric information between inclination toward the status quo and inertia, refer- physician and patient (Frank 2007). How patients ring to people’s tendency to maintain the current sta- obtain information about the health system and tus of things. This status is used as a reference point, doctors prior to an initial contact, how referrals and any change regarding this point is seen as a loss. are obtained, how easy or difficult it is to deal with Physicians are habitual in their processes when health insurance: all these factors play central roles it comes to a patient, and new processes or actions in the formation of “priors” or pre-existing beliefs might not be salient to them. Reminders and check- and the decision to engage with the health system. lists can work to disrupt automaticity. The literature Disease prevention efforts must confront indi- suggests that more patients receive preventive care viduals’ priors on the health system and physicians, when doctors are reminded to order, deliver, or pre- which may stem from personal experiences and/or scribe appropriate preventive care measures such as reports from family and friends (whether or not re- vaccinations (Austin et al. 1994; Davis et al. 1995; lated to the dimensions of care that most directly Hunt et al. 1998). For example, a recent systematic 164 CHAPTER 6 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E review, covering 3,502 individual studies investigat- Examples include aspirational smoking as a sign of ing behavior change interventions among primary emancipation or freedom among women or youth health care (PHC) physicians, found that reminders (Amos and Haglund 2000; Toll and Ling 2005) or improve the rate of screening and vaccination while increased rates of obesity among the poor as a re- reducing unnecessary imaging for lower back pain sult of aspirational food consumption (Schneider (Chauhan et al. 2017). Other systematic reviews et al. 2020). Community-level signals of what is also found a positive relationship between physi- an accepted or desired behavior do not come only cian reminders and uptake of screening services from individuals. Physical cues such as ashtrays in (Dexheimer et al. 2008). public places, “smoking-friendly” labels, unhealthy The health sector’s performance when it comes to or highly caloric foods at social gatherings, and the added task of preventing future diseases can others are all cues in social settings that normalize benefit from more motivated workers. A study unhealthy behaviors, more so when they are also as- investigating major factors influencing job motivation sociated with identity elements. among public health staff in Viet Nam found that Social references and cues are particularly recognition by managers, colleagues, and the strong for the initiation of behaviors at early ages, community was the most important motivating factor such as adolescence and early youth. In consider- to improve job performance (Dielaman et al. 2003). ing these issues, behavioral science distinguishes Similarly, peer comparison can operate as additional between injunctive norms (people’s perceptions of motivation for the system to increase delivery of what behaviors are approved or disapproved) and preventive care services such as vaccinations (Barton descriptive norms (people’s observations of what and Schoenbaum 1990; Briss et al. 2000). Peer behaviors are performed in reality). A systematic comparison feedback can also boost referrals to other review examining evidence of the longitudinal as- preventative services. For example, in a study on the sociations between different types of social norms likelihood that physicians will refer tobacco users to a towards smoking and youth smoking uptake (ini- local smoking quit line, benchmarking doctors against tiation and escalation) finds that descriptive norms the top 10 percent of performers increased their relating to parents’ and close friends’ smoking likelihood of refer-ring patients (Bentz et al. 2007). behavior appeared to be consistent predictors of youth smoking initiation, more so than the de- Community-level factors that affect scriptive norms of more distal social networks and NCD formation injunctive norms (East et al. 2021). A study com- paring adolescent men in the capital cities of Arme- People follow reference and influence groups and nia and India found that, in Armenia, adolescent conform to identities and norms because it creates men were more likely to have a smoking father and intrinsic utility (Akerlof and Kranton 2000). More were 20 times more likely to accept and smoke a dominant identities and strong social norms have cigarette from pressuring peers than male adoles- been found to promote or discourage certain cents in India, where smoking prevalence among behaviors. While relevant research on NCD the adult population is lower (Pandey et al. 2011). formation has been limited, evidence suggests that Among youth in England, Canada, and the United reference groups and peers play an important role States, East et al. (2019) find positive social norms (see for example Serrano Fuentes et al. (2019) for around vaping associated with use, with perceived a review on obesity and social ties). For example, peer approval of vaping being double that of smok- research has consistently shown that men are ing. The same authors explored whether current more likely than women to engage in certain risky smokers’ social norms towards smoking and elec- behaviors, such as smoking and drinking, and that tronic cigarettes varied across a wider set of Euro- such behaviors are linked to masculine identity pean countries, including Romania, Spain, Hunga- and peer pressure (see for example Keenan et al. ry, Poland, Greece, Germany, and England (East et 2015). al. 2018). At the time of the study, England had the Social norms can also trigger aspirational lowest smoking prevalence among these countries. consumption, i.e., the purchase of goods or foods Accordingly, smokers from England had the least prompted by relative deprivation and inequality pro-smoking norms. Peer smoking broadly aligned (Bellet and Colson-Sihra 2018; Elgar et al. 2005). with country-level smoking rates, while the approv- COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 6 165 al of e-cigarette use in public similarly aligned with because the decision was made without nutrition as country-level e-cigarette use.4 a salient factor. The same is true for lack of knowl- Similarly, community-level factors that affect edge about how to interpret nutritional information alcohol use, such as parents’ alcohol-related be- or to connect information (e.g., about calories and havior, frequently spending time with friends who protein) with foods and serving amounts using the drink, and the drinking norms observed in the wid- information provided on food labels. Interventions er social environment have been found to increase that have changed food labeling by simplifying it the risk of excessive drinking (Kuntsche et al. 2017). and increasing salience suggest that information Halim et al. (2012) studied norms influencing alco- presentation is as important as its provision. For ex- hol use among college students, considering both ample, a regulatory change in Chile mandated that descriptive norms (i.e., the individual’s perception a simple warning about high calories and high sugar of the prevalence of alcohol consumption) and in- be placed in front of products in store displays. This junctive norms (i.e., the individual’s perception change led to a reduction in the consumption of of whether drinking is approved by their peers). sugary cereals by consumers adjusting their beliefs Further distinguishing between distal (socially dis- regarding how healthy certain cereals were (Bara- tant peers) and proximal (socially close peers), the hona et al. 2020; Araya et al. 2020). researchers found that descriptive norms, proxi- Information, however, needs to be aligned with mal injunctive norms, and social motives all inde- individual motivational factors that can affect NCD pendently predicted alcohol consumption. Kremer formation. These factors are related to the internal, and Levy (2008) find that peers (roommates) in- both reflective and automatic, processes influenc- fluence college students’ alcohol use and academic ing decision making and human behavior. Present performance, hypothesizing that peers exert influ- bias, our tendency to value instant gratification ence via their impact on preferences and habits. more than future benefits, and time inconsistency, The adverse effects tend to be particularly large for our tendency to be impatient when choosing be- students who drank heavily in high school and are tween receiving benefits today and receiving them paired with a roommate who drank heavily. in the future, are two examples of these automatic Christakis and Fowler (2007) find that social processes. The emphasis on the short term at the ties and obesity prevalence are connected. They expense of the long term occurs because people tracked over 12,000 individuals over the course perceive the present as tangible, while the future of more than 30 years, finding clear evidence that is abstract and hypothetical. Present bias creates obesity spreads through social connections. For ex- self-control problems that affect many health be- ample, the risk of an individual’s becoming obese haviors, from exercising, eating healthy foods, and increased by 57 percent if a friend became obese in getting a mammogram to other preventive or pro- a given timeframe. tective health behaviors (defined by effort today and benefits tomorrow). Benefits today versus costs to- Individual decision-making that affects morrow factor in when people smoke, overeat, skip NCD habit formation health visits, and more (Luoto and Carman 2014). A short-term gratification focus is at odds with When it comes to caring for one’s health, individu- healthy habit formation. Reflexive thinking that als are asked to make frequent decisions, often with- continuously prioritizes instant gratification (watch out adequate information or the ability to interpret television over exercising) or impulses to smoke, and use information. Behaviors may not be viewed drink, or eat unhealthy foods can end in unhealthy as health risks, or may be driven by a motivation habits being formed. completely unrelated to health. For example, lack of Another psychological barrier informing NCD knowledge on the nutritional content of a meal can formation is overconfidence; both overestimating lead to unhealthy eating, either due to a mispercep- one’s performance and over-ranking one’s perfor- tion of the nutritional content of a given meal, or mance relative to others. Overconfidence explains 4   Regulations and laws, such as gradually banning smoking in public places, may also alter social norms (Acemoglu and Jackson 2014). Laws change material payoffs and thereby affect the factual behavior of the group understood as an equilibrium object, and thus the descrip- tive social norm associated with it. Laws may also provide information about societal values (Benabou and Tirole 2011). 166 CHAPTER 6 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E risk-taking behavior, including risky health behav- Other biases to consider in disease formation iors. Some people understand heart attack and can- at the individual level are loss aversion — the idea cer risks from unhealthy behaviors such as overeat- that a loss causes greater distress than the happi- ing and smoking but tend to believe themselves less ness caused from a gain of similar magnitude; the vulnerable to negative outcomes than their peers, endowment effect — the overvaluation of a good even if they adopt risky behaviors (Khwaja et al. when we possess it; and framing effects — the ten- 2007, 2009). Della Vigna and Malmendier (2006) dency to draw different conclusions depending on show overconfidence in habit formation related to how the information is presented. Cawley (2011) exercise: many gym members overpay, as they ex- found that subsidizing healthy food options was ercise too little, and propose overconfidence about less effective in promoting healthy eating than was future self-control as one possible explanation for taxing the unhealthy food options. People viewed such observed behavior. the discount on healthy foods as a gain, which was Arni et al. (2021) measured health perception subsequently valued less than the loss of having to biases -- their label for overconfidence -- using ob- pay more for the unhealthy food. Another example jectively measured cholesterol and blood pressure from Kahneman (2011) is that cold cuts described in Germany and found that 30 percent of the pop- as “90% fat-free” sound a lot better than cold cuts ulation have biased perceptions about their choles- that have “10% fat.” terol levels. The researchers also found that about Many of the aspects previously cited also influ- 30 percent of respondents overestimated their rank ence the information and motivation dimension. in the population health distribution by at least 30 However, individual biases can also impact oppor- ranks relative to a reference group. Furthermore, tunity factors that lead to NCD formation. These are greater overestimation of own health was associat- related to external factors that make execution of a ed with more engagement in unhealthy behaviors, behavior possible. Physical opportunity, opportuni- such as consuming unhealthy food, not exercising, ties provided by the environment, and social oppor- and sleeping fewer hours. This is akin to saying: “Be- tunity, for instance to perform regular exercise, can cause I believe I am very healthy and can afford it, promote or diminish healthy behaviors. For exam- I eat more fast food and exercise less.” Notably, this ple, Zimmerman (2011) argues that obesity is not a study finds that smoking is not correlated with bi- rational choice, and that weight gain is not the result ased health perceptions. of people maximizing their utility given fixed pref- The finding concerning smoking is consistent erences, but rather it relates primarily to the influ- with Darden (2017), who finds that cardiovascular ence of food-producing companies, which attempt biomarker information provided at repeated health to change people’s tastes toward fatty and high-calo- exams does not significantly alter smoking behavior. rie foods. This takes several forms, including direct Cognitive dissonance, our tendency to often hold advertising, product placement, and advertising in contradictory opinions, may explain these results. schools. It is important to note that this does not ex- People smoke, for example, even when knowing plain increasing obesity in low- and middle-income of its danger. This self-contradiction is unpleasant countries. In such contexts, the growth of income (causes dissonance), so people develop cognitive per capita as well as the increasing affordability of tools that justify the contradiction. The most obvi- fast food relative to healthy food have been im- ous way to address the contradiction, of course, is to portant drivers in the growth of unhealthy eating stop smoking, but that is difficult to do (Rice 2009). habits. Consequently, behavioral factors such as the Instead, people use other strategies, such as exagger- one described for obesity are more likely to explain ating how much pleasure they get out of smoking larger variations in high-income countries than in (that is, overweighting the present) or underesti- low- and middle-income countries. mating the future costs (e.g., “Aunt Anna smoked her Finally, individual capabilities, skills, and abil- whole life and lived to age 89 without health prob- ities also affect NCD formation. Decision fatigue lems”). Another possible explanation is that signals and cognitive overload, for instance, make the de- and information about own health, be they objective cision-maker prone to poorer decisions. In market- as in Darden (2017) or perceived as in Arni (2021), ing, sweets or other foods that people might not spe- are not powerful drivers of smoking behavior, possi- cifically search for in a supermarket are placed near bly because its addictive nature prevents the proper the checkout aisles: customers having already made evaluation of the health consequences of smoking. dozens of decisions on what to buy have reduced COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 6 167 mental resources available, thus becoming more ple to focus on immediate needs over longer-term prone to impulse buying (known as ego depletion) goals (Mullainathan and Shafir 2013). In these sit- (Cohen 2012a, 2012b; Baumeister 2002). Examples uations, health decisions, if not obviously critical, of cognitive overload and food choices have been might take second place to many other, more ur- tested experimentally, for example by asking study gent and immediate needs. Similarly, consumption participants to choose between cake and fruit sal- of what is available now (whether healthy or not) ad while having to remember a number. Compared and impulsive behavior might be a simpler choice. with a two-digit number, those with a seven-digit This section has so far discussed how different number were 50 percent more likely to choose cake behavioral biases can impact the formation of NCDs over fruit (Shiv and Fedorikhin 1999). along the pathway to healthy ageing. However, it is Decision fatigue does not have to emerge from important to highlight that behavioral biases—and a single domain of decision-making. While some therefore the toolkits discussed in later sections— individuals might face decision overload in a given may be different when dealing with positive habit moment, others may confront a constant demand to formation as opposed to eliminating negative hab- make decisions due to limited opportunities. Poor its. Different behavioral biases may become dom- people are particularly subject to decision fatigue inant in certain circumstances, and therefore the because so many of their decisions involve signifi- tools required to address them need to adapt to the cant trade-offs that even the smallest decision can setting. Most positive habits, such as beginning to have far-reaching monetary and welfare implica- exercise regularly, require fairly prolonged expo- tions (Spears 2011; Mani et al. 2013). As suggest- sure and often convenient access to an appropriate ed by Bertrand et al. (2004) and Shah et al. (2012), contextual environment, so overconfidence (in time among others, poor individuals may exhibit the management and/or self-efficacy) is a common is- same basic weaknesses and biases as do people from sue. While stopping some negative behaviors (such other walks of life, except that, in poverty, with its as unhealthy eating) may follow a similar dynamic, narrow margins for error, the same behaviors often many behaviors in this category are instead more manifest themselves in more pronounced ways and like smoking or other addictions, where loss aver- can lead to worse outcomes. Material scarcity, for sion and present bias, for example, can make it diffi- example, has been found to change the allocation of cult to overcome the initial hurdle. attention, promoting tunnel vision and forcing peo- NCD DETECTION: SCREENING The second main phase of the individual journey to healthy aging has to do with screening and identification of illnesses. As in the previous section, our analysis follows health system, community, and individual-level factors and discusses, within each, behavioral barriers and biases influencing health-seeking behavior and screening and detection practices. Health system factors that affect NCD screening The role of health system factors in the uptake of NCD sion-making process is also prone to potential behav- screenings involves both the doctor-patient interaction ioral biases that may influence the physician’s choices. and broader system characteristics. In each of these ar- Several studies have highlighted how physician eas, we find behavioral components that influence the risk preferences impact treatment and screening appropriate uptake and use of NCD screening. decisions. Some show, for example, that physician The doctor-patient relationship is often present- risk aversion is strongly associated with patient ed as an agency relationship: the physician makes in- charges and expenditures over time (Allison et al. formed decisions on behalf of the patient (McGuire 1998; Fiscella et al. 2000). This highlights how, in- 2000). Under this model, the physician makes deci- dependent of the patient’s preferences, physician sions in line with the patient’s preferences- including risk preferences and biases could impact people’s with regard to risk and time. However, such a deci- health care-seeking behaviors. Even though a recent 168 CHAPTER 6 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E study found that risk preferences between patients access and personal biases that further reduces the and physicians were generally aligned (Galizzi et al. likelihood of people’s accessing screening services. 2016), they may be misaligned in certain settings or A large share of biases also falls within the motiva- when patients inaccurately estimate their own risks, tion dimension from the individual perspective. For and this possible misalignment may help explain instance, higher access or opportunity costs have why NCD screening remains low. routinely been linked with a greater likelihood of In contrast to the risk preference scenario, a delaying screening (Lagarde et al. 2007). One expla- field experiment in Greece found a significant dif- nation for this is the simple demand and supply the- ference in time preferences between patients and ory, whereby higher prices discourage people from their physicians (Galizzi et al. 2016). This study seeking health care, but another behavioral explana- highlights that, while improving communication tion is that higher costs counterbalance discounted between physicians and patients could improve the future benefits. This issue links back to individual uptake of screening, it is of even greater importance motivation, but it is important to note that health to integrate a wide set of characteristics in the com- system factors such as prices or distance to health munication process, including the potential mis- care establishments can further strengthen behav- match in time preferences. ioral biases that hinder access to screening services. Physicians are important drivers of patients’ health care-seeking choices insofar as they are the Community-level factors that affect choice architects for patients within the health care NCD screening realm. A mixed-method study in the United States highlighted how physicians determine patients’ As social and community forces sometimes support choice possibilities, even though doctors proved unhealthy or risky behaviors that can, over time, not to be fully competent in their roles as choice ar- lead to an NCD, they might also play a role in when chitects (Hart et al. 2021). This reaffirms the physi- and for what reasons people seek health care. While cian-patient interaction as a key stage in which phy- preventative care is not always offered when individ- sician biases can influence the appropriate uptake of uals show no signs of illness, screening might not screenings and other procedures. take place even when these signs are present. Despite Physicians themselves are, as part of the system, widespread evidence of the benefits of screening, affected by coordination failures and social influence. many cultural norms may deter individuals from A systematic review of health care provider barriers pursuing it. For instance, men seem to be less likely to hypertension awareness, treatment, and follow up to seek mental health services, as doing so clashes found that a common barrier was social influence with traditional masculine identity around strength in the form of a lack of care coordination with col- and self-sufficiency (Pattyn et al. 2015). Similarly, as leagues, along with social pressures and conflicting discussed earlier, if social networks and communi- roles within a medical practice. Professional identity, ties are formed by individuals with similar charac- expressed as a lack of trust in the evidence on which teristics (e.g., people who smoke or are overweight), guidelines were based, was also found to influence these networks are less likely to see the behavior and/ hypertension treatment (Khatib et al. 2014). or its consequences as warning signs to seek help. Health care accessibility issues, particularly di- The role of norms has been tested when mak- rect costs and opportunity costs are a primary rea- ing them salient in, for example, reminders to at- son why individuals forego screenings. In a recent tend screenings or other health messages that refer review of the cost-effectiveness of interventions to to peer behaviors (Camilloni et al. 2013). A recent reduce NCDs in LMICs, Watkins et al. (2022) find study in Armenia found that messages appealing that increased screenings for diabetes and cervical to peer behaviors increased screening attendance cancer are among the most cost-effective. Reducing about 15 percentage points for hypertension and barriers within the health system itself will be crucial diabetes (De Walque et al. 2020). However, the per- to increasing screening for these and other condi- sonalized invitation referencing peer behaviors did tions. While health system accessibility is mainly in- not have an add-on impact compared to standard fluenced by policies, there is an interaction between personalized invitation.5 A further study with the 5   The intervention tested 1) personalized invitations from a physician, 2) personalized invitations with information about peer screening behavior, 3) personalized invitations with a labeled but unconditional financial incentive, and 4) personal invitations with a conditional financial incentive. COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 6 169 Individual decision-making that affects same population found that participants described NCD screening and identification learning that neighbors and peers attended screen- ings as a potential motivator for service use, more In the case of ageing-related diseases, two behavior- so if the screening helped uncover a disease or ob- al factors are largely responsible for the low uptake tain treatment (Gong et al. 2020). Social norms may of prevention. The fi rst is th at, in contrast to ot her not only influence initial access or use of screening illnesses, most NCDs remain asymptomatic for a services, but are probably also relevant to the sub- long period of time (Bovet et al. 2015) and lack early sequent interactions that occur within the health saliency. The second is inaccurate individual beliefs system, which can then influence the full uptake and overconfidence, which lead people to underval- of screening options. Cultural aspects like personal ue the benefits of screening (Baicker et al. 2015). health beliefs and qualities linked to interpersonal Even with a clear understanding of which risk relationships, such as distance, trust, and directness, factors lead to NCDs, individuals tend to under- influenced patients’ judgment of a physicianpatient estimate their likelihood of suffering from an NCD. interaction (Gao et al. 2009). This issue relates to the risk-taking behaviors Stigma and fear are related to social norms and described earlier concerning the formation of NCDs, community assessment of health-seeking behaviors whereby individuals believe themselves to be less and can operate as enforcement mechanisms of pre- likely than their peers to suffer such poor fortunes vailing norms. Early work highlighted how structur- (Khwaja et al. 2007, 2009). Evidence demonstrates al stigma- a social phenomenon with roots in social that actual risk and perceived risk for cardiovascular structures- exerted substantial influence in health diseases were weakly correlated, with most high-risk care-seeking behavior for many causes (Hatzenbue- individuals being incorrectly optimistic (Van der hler 2014). Even though evidence on stigma remains Weijden et al. 2007), and that people under-estimate scarce in the literature, a growing number of studies their own risk for high cholesterol while over- now highlight how social perceptions of a disease, its estimating the risk of others (Arni et al. 2021). As causes, and its consequences can influence individ- individuals underestimate their own risk, they are ual health care-seeking behavior. A recent system- then less likely to value the benefits of screening and atic review found that blame, shame, and fear were to consider that these benefits are larger than the the main factors behind stigma for diseases such as costs of attending a screening session. cancer and that this had social, behavioral, psycho- This highlights the role of present bias in the logical, and medical consequences (Rai et al. 2020). low uptake of screening. People generally tend to be Fear of stigma as a community judgement can time-inconsistent and delay screenings, and the stem from different conditions. Conventional body reduced value that people attribute to screenings image norms and failure to meet these may deter reinforces the impact that present bias can have in people, especially women, from seeking screening postponing appropriate preventive care. A theoretical services (Kilpela, et al. 2015; Fingeret et al. 2014). study using US data found that time-inconsistent In the United States, as in many other countries, the health investment behaviors, including with regard to uptake of disability programs became stigmatized screenings, could explain a loss of about four years in to the extent that it discouraged participation in the life expectancy (Strulik and Wener 2021). Further screening process for individuals to be declared dis- studies have also shown that high discount rates are abled (Hansen et al. 2014). Stigma can also influence associated with lower screening rates for cancer and the uptake of cancer screening. For example, Mc- measurements of blood cholesterol (Bradford 2010). Caffery et al. (2006) report that, in the UK, screening It is important to emphasize the relevance of motiva- and testing positive for the human papillomavirus tion in terms of healthy ageing, since preventive (HPV) — a risk-factor for certain forms of cancer screenings are not a onetime process. Usually, — was accompanied by feeling stigmatized. Stigma screening for NCDs needs to be conducted annually associated with mental diseases is among the leading after certain ages. Therefore, these biases can have reasons for the underuse of mental health services long-term consequences if motivation for screening (Evans-Lacko et al. 2012). Similarly, the stigma as- uptake remains low. sociated with degenerative diseases like Alzheimer Loss aversion also influences the motivation or other forms of dementia is likely to discourage in- to attend health screenings in a subtle way. Even dividuals from following a screening procedure, for people with accurate perceptions may not take up fear of the consequences that might follow it. screening to avoid the realization of a foreseeable 170 CHAPTER 6 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E bad outcome. A qualitative study in the United health management programs (Beatton et al. 2021). States found that fatalism and negative outcome Overall, these studies emphasize that loss aversion perspectives were important barriers to participa- is an important driver in declining motivation and tion in screening for breast cancer (Peek et al. 2008). procrastination around screening. Similar findings were obtained across multiple Finally, decision fatigue and cognitive overload countries, linking fatalism to lower attendance at can also lead to poor decision-making processes cancer screening programs (De Los Monteros and with regard to screening. A potential diagnosis of Gallo 2011). Another literature review based on the a chronic condition that would entail regular ex- UK population found fear of a negative outcome penditures may lead to unmanageable situations as one of the reasons individuals did not attend in poorer households. As in the case of habit for- the regular health checks proposed by the Nation- mation, households unable to afford possible treat- al Health Service (NHS) (Harte et al. 2018). While ment in the future may prefer to avoid attending one might think that loss aversion would only affect a screening consultation (Spears 2011). Cognitive voluntary attendance at screenings, other studies overload can also have an impact on an individu- suggest that this fear also leads people to reject ex- al’s capability to follow up on the different stages plicit invitations to screening programs (Witte and of a screening process. Even though invitation let- Allen 2000). More importantly, loss aversion does ters increase the uptake of preventive screening not only have a personal impact. A field experiment programs, often they contain an overwhelming in Australia showed that loss aversion among rel- amount of information, limiting the potential of the atives can influence an individual’s enrollment in invitation to increase screening uptake. NCD MANAGEMENT This section discusses how behavioral factors influence the long-term management of NCDs via sustained lifestyle changes, medication adherence, continued medical care, and other measures necessary for managing chronic conditions. Health system factors that affect NCD management therapies seem to be lower than appropriate (McCa- rthy 2019; Greene et al. 2018; Salami et. al. 2017d) Traditional behavioral change approaches to better or why physicians treating cardiovascular disease align clinical practice with guidelines have mainly under-prescribe statins and fail to intensify treat- focused on improving access to information and ment when indicated (Asch et al. 2015). Due to the guidelines. These methods are based on conven- cognitive biases impacting their decision-making in tional economic theory and presume that physi- the context of NCD treatment, doctors may judge cians are perfectly rational decision-makers whose risk to be low without available personal experienc- main barrier is lack of information. es (for example, knowing someone or experiencing Physicians, however, make numerous complex themselves the consequences of under- or over-pre- decisions on diagnostic and treatment plans dai- scription), may be less careful than expected when ly, often with limited information and under time not observed, and may falter without an injunction pressure (Blumenthal 2015). When making a high from authority (Doctor et al. 2018).  Such biases volume of decisions under conditions of uncertain- may explain why physicians sometimes continue to ty, the decision-making process may be guided by deliver care that robust evidence has shown to be of environmental cues and heuristics (Blumenthal low value (Scott et al. 2017). 2015; Scott et al. 2017). These make physicians especially vulnerable to systematic cognitive bias- Community-level factors that affect es (Saposnik et. al. 2016), for example, distorting NCD management probability estimation and impairing information synthesis (Wang and Groene 2020). Some of the community-level factors mentioned in Behavioral science can help explain why, for prior sections — social norms, peer pressure, roles, example, prescription rates of guideline-directed identities, and stigma — also affect NCD manage- COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 6 171 Individual decision-making that affects ment, both through patient and provider behavior. NCD management On the providers’ side, norms about treating “someone else’s patient,” identity related to health Management of NCDs in most cases requires adher- providers’ scope of work (i.e., nurse vs. doctor), ence to medications, ongoing follow-up medical care, and shared beliefs about treatments, medications, and sustained changes to lifestyle and health behaviors, clinical guidelines, and actual capabilities to help such as physical activity or smoking, which are difficult patients manage and control NCDs can hinder how to achieve. Adherence to one regime (medicines) might much providers engage with patients in their on- displace cognitive resources from these additional sets going care. Khatib et al. (2014) find evidence that of changes and new habits, potentially enabling other bi- providers doubt the efficacy of certain hypertension ases related to decision-making to take hold. Cognitive medications, are reluctant to initiate aggressive an- biases particularly relevant to NCD management are ti-hypertensive drug treatment due to possible side those that affect how people perceive losses (Kahneman effects, and question whether following clinical and Tversky 1979), value the present vs. the future guidelines would improve outcomes. (O’Donoghue and  Rabin 1999), and understand am- On the patients’ side, shared beliefs, myths and biguity (Fox and Tversky 1995; Fox and Weber 2002). misinformation, social norms, and community prac- In cases where habit change is difficult, the immediate tices can also hinder appropriate treatment and med- negative side effects of giving up certain unhealthy be- ication adherence. For example, in the United States, haviors (tobacco or specific foods, for example) may African American and Latino patients’ beliefs about be perceived as worse than the potential long-term the effectiveness and side-effects of medications have consequences of non-adherence to the recommended been found to lower adherence (Wexler et al. 2009; behaviors, even though those consequences are in fact Peters et al. 2008). In Zimbabwe, poor awareness greater, even when accounting for their occurrence in from both patients and the community has been the distant future (Luoto and Carman 2014). found to result in the use of alternative, and poten- Even when people are motivated, limited atten- tially harmful, remedies (Kamvura et al. 2022). Com- tion can interfere with medication adherence, care munity suspicion of government has also been found seeking, and lifestyle changes. The limited ability to to hinder NCD management (Heller et al. 2019). notice, store, and process information can help ex- Lack of social support, mainly from family and plain the difficulty many individuals have in adher- friends, is an important barrier to treatment and ing to a prescribed course of treatment (Kessler and medication adherence and healthy lifestyle chang- Zhang 2014). The formation of habits (or changing es (Devassy et al. 2020; Khatib et al. 2014). The existing ones) requires self-efficacy, effort, and in- ability to make lifestyle changes to control NCDs vestment in the creation of an opportunity “event” in many cases involves changing social practices. that links the needed healthy behavior to an envi- Improving eating patterns, for instance, may entail ronmental cue or trigger, such as placing medicines new cooking patterns within a household or having in a visible location in the morning, even when the to cook a healthier meal for oneself (Khatib et al. motivation (feeling sick or fear of future conse- 2014).  Social interactions that involve smoking quences) becomes less salient. or drinking exert peer pressure on the person As people also rely upon knowledge that is salient undergoing treatment. The opposite holds, as well: and easy to retrieve from memory when judging prob- a positive social environment and good social sup- abilities and making decisions (Tversky and Kahneman port have been shown to have a significant impact 1973; Bruine de Bruin et al. 2016), the salience of a giv- on compliance with the medical management of en piece of information, or the degree to which that chronic disease (Basu and Millett 2013). piece of information stands out relative to other infor- Finally, apprehension about stigma and stereo- mation, can also affect decisions related to medication types, often stemming from past negative interac- adherence and habit change. If a piece of information tions with the health system, can lead patients to fear is not particularly salient or becomes less salient over discrimination based on factors including race/eth- time, it might be overlooked or more easily forgotten. nicity, age, gender, and weight. This has been associ- The salience of symptoms and of the treatment it- ated with higher distrust of physicians and dissatis- self can influence medication adherence (Kessler and faction with health care, poorer mental and physical Zhang 2014). Over time, if a medication’s beneficial health (including self-rated health, hypertension, effects become less noticeable, adherence can fade, as and depressive symptoms), and lower odds of re- in the case of low adherence to tuberculosis treatment ceiving the influenza vaccine (Abdou et al. 2016). regimes (Munro et al. 2007; Pradipta et al. 2020). 172 CHAPTER 6 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E RECOMMENDATIONS FOR A POLICY TOOLKIT Behaviorally informed interventions for NCD formation, detection, and management Traditional policy tools based on the rational subject model have been moderately effective in changing un- healthy behaviors. In the case of tobacco use, for instance, raising prices through taxes on cigarettes, providing information about the perils of smoking, and introducing barriers to the purchase and use of tobacco have curbed the behavior, but large shares of the population continue to smoke. Similarly, information about the risks of overeating, drinking alcohol, sedentary lifestyles, and other unhealthy habits does not seem to be enough to inflect these behaviors. Insights from the behavioral sciences can complement these traditional tools, increasing their efficacy and impact. Behavioral interventions that aim to address the barriers across the whole journey towards healthy factors behind risky behaviors appear to support the aging. Nor should these be considered stand-alone reduction of NCD lifestyle risk factors. Systematic interventions that can operate independently of reviews have assessed the evidence around these other services, health system resources, and oth- interventions. Blaga et al. (2018) reviewed the effec- er structural factors. It is in the combination of a tiveness of strategies used to promote tobacco cessa- strong structure of opportunities to adjust behav- tion, reduce alcohol consumption, promote health- iors and habits, supported by motivation and a ful diets, and increase physical activity, finding sense of ability to act, that behavior change can take promising results with potential impact on NCDs. place and be sustained over time. The policy toolkit Möllenkamp et al. (2019) studied the effectiveness outlined here is both ample and complimentary to of behavioral interventions in improving self-man- existing policy efforts. These efforts include classical agement among adults with chronic diseases and interventions such as taxation, subsidies, or restric- reported positive impacts. These reviews and oth- tions. The purpose of this review is not to compare er studies find that the most commonly tested (and the effects of behavioral and classical interventions; successful) behavioral interventions are reminders, however, we emphasize the need for future studies planning prompts, nudges (i.e., indirect influencers to evaluate how different combinations of behav- inspired by cognitive biases and boundaries), and ioral and classical tools can maximize the effect to feedback focused on self-management. Most finan- achieve healthy ageing. cial subsidies, fines, and other attempts to directly In terms of interpretation, the tools described change the cost-benefit calculation of a decision do below can be linked to the original framework by not count as behavioral, but small financial incen- way of identifying the “Behavioral Phenotype” of a tives constitute a nudge if they work primarily via given policy’s intended target population. This means salience or signaling of norms. Other reviews have identifying and segmenting the target population not focused on behavioral tools such as choice architec- only by risk characteristics or sociodemographic ture (Hollands et al. 2013; Skov et al. 2013; Thapa traits but based on a larger set of profile elements, and Lyford 2014), finding promising but under-re- including behavioral and psychological variables, searched impacts. Moderate impacts have been motivation, ability, and opportunity factors. While found on time discounting when it comes to obesity for some people reminders are all that is needed to (Barlow et al. 2016), while altering default settings close gaps in treatment adherence, for others more and providing social reference points have been complex interventions to inform, motivate, and cre- found to have positive effects on changing physician ate adequate social support might be required. behavior (Wang and Groene 2020). Ongoing man- The tools and literature presented in this section agement of NCDs, mainly via medication adher- are a selection of the relevant evidence that speaks ence, is among the most cost-effective strategies for to the three stages of NCD formation, detection, NCD control (Watkins et al. 2022); the policy tools and management. We leveraged randomized field discussed here can support adherence by address- interventions and systematic reviews to showcase ing barriers that currently prevent individuals from the behavioral toolkit available to policymakers to entering and remaining in care, thereby improving tackle NCDs. However, the selection is not the result the effectiveness of such interventions. of a systematic review. The literature search strat- The behavioral science toolkit is large. No single egy comprised English-language, peer-reviewed policy tool can address the complexity of behavioral electronic records from Google Scholar published COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 6 173 between 2000 and 2022. Reference lists of relevant short term, but effects disappear when the incentives studies were also hand searched for eligible papers. are removed. Combinations with other behavioral The approach prioritized studies from low- or mid- insights such as peer and framing effects tend to ren- dle-income countries and was supplemented with der better lasting results. One study focused on US evidence from high-income countries. The selected school children found that incentives can increase field studies, mega-studies, and systematic reviews rates of fruit and vegetable consumption, particu- conducted or assessed randomized control trials larly among lower-income children, who are most (RCT) to measure the interventions’ effectiveness. likely to benefit (Just and Price 2011). We gave precedence to mega-studies and sys- Financial incentives to patients and physicians tematic reviews to ensure the external validity or are increasingly used to manage NCDs. According generalizability of the tools. However, there are gaps to a meta-analysis by Haff et al. (2015) and a map- where evidence is mixed or very limited, and this ping review by McGill et al. (2018), contemporary is reflected in the toolkit. There is more robust evi- incentive designs that leverage people’s “decision dence base related to financial incentives on a wide biases” may improve the efficacy of incentive inter- range of behaviors, as compared with salience or ventions. A meta-analysis of financial incentives in- plan-making. We also have little information about cluding 23 RCTs to promote physical activity found the persistence of the reported effects for even the that modest incentives (US$1.40/day) increased most successful interventions, as studies of this mean daily step counts during the intervention pe- kind rarely follow participants for more than a few riods for interventions of short and long durations months after the intervention. This limitation is and after incentives were removed (Mitchell 2020). most important for interventions that aim to build An additional body of evidence supports the habits, for example around exercise, as opposed to notion that financial incentives could motivate interventions focused on one-time (or infrequent) weight loss (Finkelstein et al. 2007, 2017; Volpp et behaviors such as screenings. al. 2008b; John et al. 2011; Kullgren et al. 2013; Ad- The review does not include interventions fo- ams et al. 2017). Finkelstein et al. (2016) found that cused on leveraging peer effects (e.g., role models) participants who received a FitBit and cash-based or community mechanisms (e.g., support groups): financial incentives had significantly more min- not because these don’t exist, but because the strat- utes of moderate to vigorous physical activity com- egies presented below (incentives, goal setting and pared with a control group (29 additional minutes planning, reminders, and others) can all benefit weekly), but participants who received a FitBit and from a community focus to increase support for charity-based financial incentives, or a FitBit alone, healthy behaviors and reduce the potential negative did not (Finkelstein et al. 2016)—providing evi- impacts of social networks. Continuing efforts to dence that financial incentives may be more effec- communicate and inform for behavior change, and tive when received directly. Kullgren et al. (2013) to update people’s perceptions of beliefs or behav- examined individual versus group-based financial iors among their peers, can complement strategies incentives for promoting weight loss among obese more focused on the individual. employees. A group-based financial incentive was more effective than an individual incentive and Small financial incentives monthly weigh-ins at promoting weight loss among obese employees at 24 weeks. Financial incentives have been tested as means to re- Financial incentives have also been tested ward healthy behaviors across a variety of contexts, for treatment management. Volpp et al. (2008b) and as a strategy that works to promote healthy be- conducted a randomized trial of the use of finan- haviors, particularly among less well-off populations cial incentives to encourage greater adherence to (Vlaev et al. 2019, Haff et al. 2015). Interventions warfarin, a blood thinning medication. This study have been successful to promote exercise (Charness found that a variable rewards mechanism — in- and Gneezy 2009; Cawley and Price 2013), weight centivizing people with a daily lottery that offered loss (Sen et al. 2017; Cawley and Price 2013; John, a small chance of a large payout — resulted in a sig- et al. 2011; Volpp et al. 2008a), smoking cessation nificant reduction in the fraction of missed doses. (Volpp et al. 2009), and adherence to chronic disease However, after the variable rewards were removed medications (Volpp et al. 2008b). Overall, financial 28 weeks later, adherence returned to baseline lev- incentives seem to promote healthy behaviors in the els. John et al. conducted an extended version of 174 CHAPTER 6 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E the same intervention in 2011 and replicated the both social norms and individual accountability to pattern that incentivized participants lost signifi- influence physical activity. Similarly, but leveraging cantly more weight by 32 weeks but that this effect gamification, Patel et al. (2017) found that families was not significantly sustained at 68 weeks (John who received a daily intervention that assigned et al. 2011). In 2020, Barankay et al. conducted a points and levels based on step counts achieved step trial where participants were given lottery-framed goals on a significantly higher proportion of days financial incentives when adherent to statins. Each compared with the control group (53 percent vs. intervention group had a significantly higher rate of 32 percent) and had significantly higher mean step statin adherence compared to the control group at counts compared with their baseline (1661 vs. 636). six months but mean reductions in low-density li- Relatedly, Patel et al. (2016b) examined peer com- poprotein cholesterol (LDL-C) levels were non-sig- parisons in a trial where participants received team- nificantly different between the intervention groups based step count feedback compared with either the and control group at 12 months.6 50th or 75th percentile, with or without financial These interventions’ designs included many incentives. The peer comparison to the 50th per- insights from behavioral economics, including peo- centile and financial incentives were most effective ple’s tendency to overestimate small probabilities at increasing physical activity. and be particularly emotionally attracted to small Incentive framing can also impact NCD man- probabilities of large rewards, therefore often pre- agement. Patel et al. (2016) conducted a subsequent ferring a lottery-based incentive over a known, con- study to test the efficacy of “gain-framed,” “loss- sistent incentive amount. Also, the daily lottery in- framed,” and lottery-based financial incentives (Pa- centive provided regular feedback, which addressed tel et al. 2016). Only participants in the loss-framed present bias, as even small rewards and punish- arm achieved a significantly greater mean pro- ments can have large effects on behavior when they portion of days meeting their step goal compared occur immediately. This feedback also addressed with the control group, revealing the importance people’s tendency to wish to avoid regret, because of framing when delivering. In 2019, Chokshi et al. people found out each day if they would have won replicated the finding that loss-framed financial in- had they been compliant (Luoto and Carman 2014). centives are the most effective at increasing physical In an intervention tying physician financial activity but in patients with ischemic heart disease incentives to clinical goals to increase attention to (Choksi et al. 2019). A pilot study by Riegel et al. patient lipid management, primary care physicians (2020) examined the impact of loss-framed finan- were randomly assigned to control, physician incen- cial incentives on patients’ adherence to aspirin tives, patient incentives, or shared physician-patient following admission for acute coronary syndrome. incentives. In primary care practices, shared finan- At 90  days, the incentive-receiving patients had cial incentives for physicians and patients, but not a nonsignificant higher rate of aspirin adherence incentives to physicians or patients alone, resulted compared with the usual care group (90 percent vs. in a statistically significant difference in reduction 81 percent) and a nonsignificant lower rate of re- of LDL-C levels at 12 months (Asch et al. 2015). hospitalization (13 percent vs. 24 percent). To examine the effect of combining financial Contingency management is another tool of- incentives with a team-based approach, Patel et fered to provide incentives, for instance, to addicts al. (2016a) conducted a randomized clinical trial to abstain from consuming the addictive good (Caw- where financial incentives were delivered to partic- ley and Ruhm 2011; Higgins et al. 2002). It reinforc- ipants when they met their step counts as individu- es or rewards individuals with vouchers exchange- als, teams, or a combination of both. The combined able for retail goods and services or the opportunity financial incentive arm had a significantly greater to win prizes to promote positive behavioral change mean proportion of participants achieving their (Petry 2011). A meta-analysis of voucher-based re- step goal as compared with the control group (35 inforcement therapy found overwhelming evidence percent vs. 18 percent), highlighting the power of of increased abstinence; the vouchers raised compli- 6   The discrepancy in statistical significance between the intervention’s impact on statin adherence and the more clinically meaningful metric of LDL-C levels highlights the continued need for similar studies to directly measure health outcomes in addition to the health behaviors which lead to such outcomes (Hare 2021). COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 6 175 ance by an average of 30 percent, with larger effect 12-month urine test than those in the control group, sizes for rewards that were more valuable or were who were only informed about the consequences of delivered immediately (Lussier et al. 2006). How- continuing to smoke. This result was (marginally) ever, there is some evidence that contingency man- significant. However, those who used CARES were agement may be more effective at treating the use 31 to 53 percentage points more likely to pass the of opiates and cocaine than tobacco (Prendergast 12-month urine test than the control group. et al. 2006). A striking feature of these programs is Commitment contracts have also been used the relatively high success rates obtained for small to encourage other forward-looking behaviors, vouchers – as little as US$2.50 for a single negative including weight loss (Volpp, et al. 2008). Obese test for cocaine (Higgins et al. 2002; Petty and participants from the United States were random- Martin 2002), suggesting that the immediacy of the ly assigned to one of three treatments: monthly re-ward is enough to overcome the urge to use. weigh-ins, a lottery, or a “deposit contract incentive One important caveat regarding the financial system.” In the deposit contract incentive system, incentives described above is the ambiguity concer- participants committed between US$0 and US$252 ning their long-term impact. Most of the studies per month of their own money towards the goal cited here examine the effect of financial incentives of losing weight. The money would be returned on NCD management or the uptake of health care to them with a 1:1 incentive match upon meeting services only in the short term. More importantly, their goal (losing 16 pounds in 16 weeks), but they even some of these short-term studies find evidence would forfeit the money if they failed to meet their that the appropriate behavior often stops once the weight loss goal. Such a contract incorporates the financial incentives end. This brings into question behavioral economic findings of overconfidence the potential long-term effects of such incentives and and optimism bias, as people are ex-ante overly emphasizes that their usefulness may be limited to optimistic about achieving their weight loss goals cases where short-term but “sticky” habit formation and therefore more likely to enter the contract than is required. This caveat should be taken as a cautio- they otherwise might be. This incentive design also nary note, since many governments in low- and takes advantage of loss aversion because, once the middle-income countries still consider financial money is committed, the motivation to lose weight incentives - whether conditional or unconditional - is heightened by the motivation to avoid the loss of as a silver bullet for many problems. The studies the funds. Finally, the contract can potentially ap- above highlight that the long-term effects of these peal to those who are sophisticated about their pres- tools are not certain and that such approaches might ent-biased preferences: their “planner” selves today therefore only offer temporary solutions. will sign up their future “doer” selves for the hard work of following through on the binding commit- Commitment contracts – devices ment. Although the study was based on very small sample sizes, it revealed that both incentive systems Another tool for curbing unhealthy behaviors is worked. Average weight loss was 13 to 14 pounds commitment contracts, for example encouraging across both types of incentive programs, compared people to sign contracts to quit smoking and pro- with just four pounds in the control group that un- viding rewards for those who do. In the Philippines, derwent monthly weigh-ins. Furthermore, nearly Gine et al. (2010) randomly assigned 2,000 regular 90 percent of participants chose to participate in smokers into two groups - one that was offered the the deposit contract. However, in both treatments, opportunity to join a commitment program and participants were found to regain the weight upon one that was not. The program (CARES) offered withdrawal of the incentives (John et al. 2012). individuals a savings account into which they de- In South Africa, Schwartz et al. (2014) test- posited funds for six months, after which they took ed a voluntary self-control commitment device to a urine test for nicotine and cotinine. If these sub- help grocery shoppers make healthier food pur- stances were not detected, the money was returned chases. Participants, who were already enrolled to them; otherwise, their money was confiscated in a large-scale incentive program that discounts and donated to a pre-chosen charity. The group the price of eligible groceries by 25 percent, were that received CARES (includes both populations offered the chance to put their discount on the who accepted and did not accept the program) were line. Agreeing households pledged that they would 3.5 to 5.7 percentage points more likely to pass the increase their purchases of healthy food by 5 per- 176 CHAPTER 6 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E centage points above their household baseline for ence to self-monitoring and increased weight loss each of six months. If they reached that goal, their (Burke et al. 2012). However, the overall effects on discount was awarded as usual; otherwise, their weight loss were small, around 2 percent of body- discount was forfeited for that month. Thirty-six weight, on average. Barnes et al. (2012) found that percent of households that were offered the bind- using symbols such as stars, rather than presenting ing commitment agreed; they subsequently showed numbers, was easier for most people to comprehend. an average 3.5-percentage-point increase in healthy In addition to this, applications or processes that rely grocery items purchased in each of the six months; on sending SMS can also suffer from user fatigue. households that declined the commitment and This means that such an approach can have a strong control-group households that were given a hypo- effect in the short term, due in part to its novelty, but thetical option to precommit did not show such an that the intervention’s effectiveness can decrease and increase (Schwartz et al. 2014). even disappear as the novelty wears off, if users un- The basic principle explaining why commit- subscribe or withdraw from the messaging system. ment contracts are hypothesized to work is that Reminders and saliency applied to health work- committing oneself today to a behavior tomorrow ers can also promote medication adherence. Using can overcome hyperbolic discounting and present a randomized controlled trial format, Doctor  et bias (Luoto and Carman 2014). That is, it is easy to al. (2018) monitored the effect of notifying physi- plan to start dieting or exercising tomorrow. If your cians who had a patient die of opioid overdose within present self can precommit your future self to follow 12 months of a prescription. The physicians received through on such behaviors, intentions and actions an injunction to prescribe safely from their county’s can align where they otherwise might deviate. medical examiner. The hypothesis behind this inter- vention was that, as people rely upon knowledge that Saliency and vividness is impactful, recent, and easy to retrieve from mem- ory when judging probabilities and making decisions Saliency and vividness have also been found effec- (Tversky and Kahneman 1973; Bruine de Bruin et al. tive in highlighting the risks of unhealthy behav- 2016), decisions to avoid harms could occur more iors and inflecting those behaviors. For example, frequently after receipt of the letter, because the ef- the government of Uruguay implemented a major fects of opioid harms would be more readily avail- non-price tobacco campaign with measures includ- able to memory. Clinicians are also disproportion- ing programs to treat nicotine dependence at health ately exposed to patients who return to their clinics centers, banning tobacco advertising nationwide, uneventfully for an opioid refill. The letter alerted rotating warnings with pictograms on each pack of clinicians to those patients who do not return, owing cigarettes, restricting brands to a single presenta- to death via an overdose. Clinicians would then pre- tion, and increasing the size of pictograms to 80 per- scribe with greater care, if they perceived that they cent of the front and back of each pack. Results from are being watched, particularly by figures of authority the program evaluation indicate that, from 2007 to (Cialdini 2003). To receive a message communicat- 2012, the proportion of pregnant women who had ing a patient’s overdose death from the medical ex- quit smoking by their third trimester increased aminer would have particular weight. This interven- markedly, from 15 to 42 percent. Quitting smoking tion led to reductions in high-intensity prescribing, by the third trimester increased birth weight by an reductions in the likelihood that an opioid-naïve pa- estimated 163 grams (Harris et al. 2015). Although tient would receive a prescription, and a reduction in the Uruguayan campaign has elements that oper- overall cumulative opioid intake (Doctor et al. 2018). ate at both the individual level and the system level (e.g., access to treatment programs), overall salien- Plan-making cy is about the decision-making context and can therefore be considered most closely linked to other Plan-making can also make adherence more salient. community-level interventions. Milkman et al. (2011) designed a field experiment Another randomized trial tested the role of to test whether suggesting that individuals write feedback in the form of mHealth SMS messages in down when they intended to get an influenza vac- encouraging greater adherence to self-monitoring cine (thus making it more salient) increased rates in a weight loss regimen. The two-year study found of vaccination. This approach aligns two behavior- that daily SMS feedback messages enhanced adher- al tools frequently used together to promote better COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 6 177 behaviors regarding NCD management, namely health care providers. Some studies have compared mental contrasting with implementation intentions the effects of different technologies for sending shift (MCII). During mental contrasting, people first reminders. Results indicate that text messages, tele- are prompted to indicate a wish (what they want phone calls, and even reminders made in person to achieve) and then to imagine what fulfilling that present similar rates of reduction in absenteeism wish would feel like. This is followed by an identi- (Perron et al. 2013; Mugo et al. 2016). Vervloet et fication of the obstacle(s) to fulfilling the wish in al. (2012) reviewed 13 randomized controlled trials present reality, and by the identification of a plan that test the effectiveness of different electronic re- to address such obstacles (e.g., a wish to be health- minders for increasing adherence to various types of ier to be able to play with a grandchild, obstructed chronic medication. They found evidence suggest- by a lack of regular exercise today, that would be ing that electronic reminders encourage medication tackled by a daily walk after lunch). By focusing on adherence in the short run (less than six months) the motivation factors - wish and outcome of the but are less effective over the long run. Furthermore, wish - and identifying an obstacle and plan to ad- effectiveness can vary across medications for a giv- dress it that is under the individual’s control, effort en type of electronic reminder. While text messages toward the desired future is strengthened (Oettin- are an effective tool for increasing adherence among gen et al. 2001; Oettingen 2015). The technique has adult patients with HIV (Hardy et al. 2011; Pop-El- been shown to reduce drinking (Wittleder et al. eches et al. 2011) or children requiring influenza 2019), improve healthy food consumption (Loy et vaccinations (Stockwell et al. 2012), they have mixed al. 2016), increase physical activity and weight loss effects on adherence for women taking oral contra- (Marquardt et al. 2017), and reduce smoking (Mut- ceptives (Hou et al. 2010; Castano et al. 2012). ter et al. 2020). Stadler, Oettingen, and Gollwitzer (2009) found that women who used MCII doubled Choice architecture their amount of weekly moderate-to-vigorous phys- ical exercise (about one hour more per week) and A favorite approach for behavioral economists to that the effect was sustained for four months. mitigate present bias is simply to remove the temp- tation from one’s choice set, or to make the healthy Reminders and unhealthy options equally convenient. Choice architecture refers to the practice of influencing To tackle limited memory and forgetfulness, elec- choice by “organizing the context in which people tronic reminders are increasingly being used by make decisions” (Thaler et al. 2013, p. 428). A sys- the health system to address poor medication ad- tematic review by Blaga et al. (2018) found 10 stud- herence and appointment attendance rates. These ies that employed choice architecture to improve nu- reminders alert individuals by either providing an trition by making changes to serving lines, vending auditory or visual reminder or by sending an au- machines, or by indicating healthful and unhealthful tomatic electronic message such as a text message foods using traffic-light-like labels. Like salience, all (Kessler and Zhang 2014). A randomized evalua- these adjustments change the environment in which tion in Cordoba, Argentina, found that sending this people make health-relevant choices and should be kind of reminder significantly improved attendance considered as community-level interventions within at appointments (by 29.4 percent), an increase from the framework presented, in comparison to many of 50.6 percent (control group) to 65.5 percent (exper- the individual-level ideas described above. imental group) (p<.05) (Bonino et al. 2020). A study Another study looked at the use of portion con- in Australia showed that a scheduled reminder in- trol plates for dispensing meals to obese patients in tervention through telephone calls improved atten- Canada. Anchoring patients on smaller plate sizes dance by 16.3 percent (Ritchie et al. 2000). Another resulted in a significantly greater weight loss and investigation evidenced a 17.6 percent increase in less need for diabetes medications after six months the rate of attendance at scheduled controls among among those assigned to the intervention group vs. outpatients of an endoscopy clinic who received re- a control group that was given the usual care, or di- minders (Lee and McCormick 2003). etary teaching (Pederson, Kang, and Kline 2007). Alternative reminders such as personal phone Anchoring bias has also been used to encour- calls can increase adherence but often require cost- age healthier shopping at grocery stores. In a study ly time investment and personal commitment from in grocery stores, researchers put a piece of duct tape 178 CHAPTER 6 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E across the top of shopping carts with a sign indicat- group was given a menu that highlighted unhealthy ing that fruits and vegetables should go in front of fare, and as a result they were 47 percent less likely to the tape and behind was reserved for everything else. purchase a low-calorie sandwich. The results suggest Fruit and vegetable sales more than doubled without that highlighting the healthy options can improve affecting the store’s profitability (Bannister 2010). people’s chances of selecting more healthful fare. Similarly, another study included the design In the field of suicide prevention, changes in of a menu featuring healthful options on the front the choice architecture have been implemented page and unhealthy options at the back of the menu to deter paracetamol overdose. In 1998, the UK (Downs et al. 2009). This treatment harnessed the changed its laws on paracetamol packaging, mak- status quo bias by making the healthy options the ing it illegal for pharmacies to sell the drugs in large implicit default, given that one had to look in the containers. This change in legislation presented back of the menu to find the unhealthy fare. Those only a small barrier, as people could still buy packs who received the menu highlighting healthy fare in other outlets. Nonetheless, friction costs led to an were 48 percent more likely to purchase a low-cal- estimated 43 percent reduction in suicides in the 11 orie sandwich relative to the standard menu that years following the introduction of the legislation had a mix of healthy and unhealthy options. A third (Hallsworth et al. 2016). CONCLUSION This chapter proposes a framework for integrating insights from behavioral science into policies aimed at reduc- ing NCDs and improving healthy longevity. As NCDs become a key focus of global health policy, a framework and toolkit for understanding the role of various behavioral biases and barriers become crucial to designing effective interventions that produce lasting impacts. Despite existing knowledge regarding the magnitude of the link between behavioral habits, NCD risk factors, and disease prevention, it remains the case that a large share of deaths can be attributed to behavioral risks. The framework presented in this study has aimed some sense, this is especially true for behaviorally to clarify the relationship between individual be- informed policies, it is still the case that the fun- haviors and the achievement of healthy longevity damental insights, and many of the corresponding through entry points at the individual, community, barriers, are universal. All humans care about what and health systems levels for the three main stag- others think of them, even if the specific reference es of the NCD journey: formation, detection, and class (peers, religious leaders, or other categories) management. The study used this framework to varies by age and culture. One implication is that unpack the main behavioral barriers identified for most of the biases discussed, as well as the impacts each of these levels and stages. It then documented of targeted programs, don’t differ greatly by gender how behaviorally informed policies—for instance or income, for example. That said, women may be involving framing, timeliness, motivation, or future especially receptive to interventions based on as- focus—can support healthy longevity. The central pirations (for themselves or their children), while insight from behavioral science lies in identifying men are especially unlikely to seek health care and the obstacles and barriers that constrain people’s may need additional support in that domain. ability to realize their aspiration toward healthy ag- Finally, it is worth facing the fact that all inter- ing, whatever the form of the subsequent optimal ventions, behavioral or otherwise, involve making policy. This chapter has outlined a framework and tradeoffs and considering opportunity costs; this roadmap for diagnosing these underlying issues. is particularly true in resource- and capacity-con- The policy toolkit then proposes key areas of in- strained settings across much of the world. For- tervention that can be used across NCDs to reduce tunately, many behavioral “nudges” can be imple- their formation, improve their detection, and ame- mented essentially for free, making them even more liorate their management, and that can be adapted relevant in such contexts. This still requires the to different population groups. capacity to take the initiative and to know what to While different groups in different contexts are do, for which there are existing sources of external likely to benefit from different interventions, and in support (such as the World Bank’s eMBeD team) COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 6 179 but also a growing number of internal units (see policy approaches can be viewed as complements OECD Interactive BI Unit Map). Behavioral orga- to, not substitutes for, traditional approaches (Bryan nizations exist at either the national or sub-nation- et al. 2017). Behavioral measures are often used to al level in Brazil, Chile, Colombia, India, Lebanon, build on or leverage existing frameworks, which can Peru, South Africa, Tanzania, Tunisia, and Zambia therefore ease their adoption from both a financial – among others. It should be reiterated that all these and political perspective. REFERENCES 1. Abdou, C. M., Fingerhut, A. W., Jackson, J. S., & Wheaton, F. 13. Barnes AJ, Hanoch Y, Wood S, Liu P-J, Rice T. 2012. One fish, two (2016). Healthcare stereotype threat in older adults in the fish, red fish, blue fish: effects of price frames, brand names, health and retirement study.  American journal of preventive and choice set size on Medicare Part D insurance plan deci- medicine, 50(2), 191-198. sions. Med. Care Res. Rev. 69: 460–73. 2. Acemoglu, D., & Jackson, M. O. (2017). Social norms and the 14. Barton, M. B., and S. C. Schoenbaum. 1990. “Improving Influ- enforcement of laws. Journal of the European Economic Associ- enza Vaccination Performance in an HMO Setting: The Use of ation, 15(2), 245-295. Computer-Generated Reminders and Peer Comparison Feed- 3. Adams MA, Hurley JC, Todd M, Bhuiyan N, Jarrett CL, Tucker back.” American Journal of Public Health 80 (5): 534–36. WJ, Hollingshead KE, Angadi SS. Adaptive goal setting and 15. Basu S, Millett C. Social epidemiology of hypertension in mid- financial incentives: a 2 × 2 factorial randomized controlled dle-income countries. Hypertension. 2013;62(1):18–26. trial to increase adults’ physical activity. BMC Public Health. 16. Baumeister RF. “Yielding to Temptation: Self-Control Failure, Im- 2017;17(1):286. pulsive Purchasing, and Consumer Behavior.” Journal of Con- 4. Allison, J. J., Kiefe, C. I., Cook, E. F., Gerrity, M. S., Orav, E. J., & sumer Research 2002, vol. 28, pp. 670-676. Centor, R. (1998). The association of physician attitudes about 17. Beatton, T., Moores, C. J., Sarkar, D., Sarkar, J., Silva Goncalves, uncertainty and risk taking with resource use in a Medicare J., & Vidgen, H. A. (2021). Do parental preferences predict HMO. Medical Decision Making, 18(3), 320-329. engagement in child health programs?. Health Economics, 5. Amos A & Haglund M. From social taboo to “torch of freedom”: 30(11), 2686-2700. the marketing of cigarettes to women. Tob Control 2000;9:3–8. 18. Bellet, C., & Colson-Sihra, E. (2018). The conspicuous consumption 6. Asch, D. A., Troxel, A. B., Stewart, W. F., Sequist, T. D., Jones, J. B., of the poor: Forgoing calories for aspirational goods. Working paper Hirsch, A. G., ... & Volpp, K. G. (2015). Effect of financial incen- 19. Benabou, R., & Tirole, J. (2011). Laws and norms (No. w17579). tives to physicians, patients, or both on lipid levels: a random- National Bureau of Economic Research. ized clinical trial. Jama, 314(18), 1926-1935. 20. Bentz, Charles, Bruce Bayley, Kerry Bonin, Lori Fleming, Jack 7. Austin, S. M., E. A. Balas, J. A. Mitchell, and B. G. Ewigman. 1994. Hollis, Jacquelyn Hunt, Benjamin LeBlanc, Tim McAfee, Nico- “Effect of Physician Reminders on Preventive Care: Meta-Anal- la Payne, and Joseph Siemienczuk. 2007. “Provider Feedback ysis of Randomized Clinical Trials.” Proceedings. Symposium on to Improve 5A’s Tobacco Cessation in Primary Care: A Cluster Computer Applications in Medical Care 121–24. Randomized Clinical Trial.” Nicotine & Tobacco Research 9 (3): 8. Baicker, K., Mullainathan, S., & Schwartzstein, J. (2015). Behav- 341–49. ioral hazard in health insurance. The Quarterly Journal of Eco- 21. Bernal, P., & Martinez, S. (2020). In-kind incentives and health nomics, 130(4), 1623-1667. worker performance: Experimental evidence from El Salvador. 9. Bairey Merz, C. Noel, Holly Andersen, Emily Sprague, Adam Journal of health economics, 70, 102267. Burns, Mark Keida, Mary Norine Walsh, et al. 2017. “Knowledge, 22. Blaga, O.M., Vasilescu, L. and Chereches, R.M., 2018. Use and ef- Attitudes, and Beliefs Regarding Cardiovascular Disease in fectiveness of behavioural economics in interventions for life- Women: The Women’s Heart Alliance.” Journal of the American style risk factors of non-communicable diseases: a systematic College of Cardiology 70 (2): 123–32. review with policy implications. Perspectives in public health, 10. Barankay I, Reese PP, Putt ME, Russell LB, Loewenstein G, Pag- 138(2), pp.100-110. notti D, Yan J, Zhu J, McGilloway R, Brennan T, Finnerty D. Effect 23. Blumenthal-Barby, J. S., & Krieger, H. (2015). Cognitive biases of Patient Financial Incentives on Statin Adherence and Lipid and heuristics in medical decision making: a critical review Control: A Randomized Clinical Trial. JAMA Network Open. using a systematic search strategy. Medical Decision Making, 2020;3(10):e2019429. 35(4), 539-557. 11. Barlow, P., McKee, M., Reeves, A., Galea, G., & Stuckler, D. (2017). 24. Bonino PM, Salvia N, Nocetti G, Michelini Y, Parker DG. Efecto Time-discounting and tobacco smoking: a systematic review del envío de recordatorios sobre la asistencia a los turnos en and network analysis. International journal of epidemiology, pacientes bajo tratamiento de consumo de sustancias en la 46(3), 860-869. ciudad de Córdoba (Argentina). Rev. Salud Pública (Córdoba) 12. Barlow, P., Reeves, A., McKee, M., Galea, G., & Stuckler, D. (2016). [Internet]. 30 de octubre de 2020 [citado 23 de noviembre de Unhealthy diets, obesity and time discounting: a systematic 2021];24(3):95-101. literature review and network analysis. Obesity reviews, 17(9), 25. Bovet, P., Chiolero, A., Paccaud, F., & Banatvala, N. (2015). 810-819. Screening for cardiovascular disease risk and subsequent man- 180 CHAPTER 6 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E agement in low and middle income countries: challenges and Factor of Obesity and Chronic Disease.”New England Journal of opportunities. Public health reviews, 36(1), 1-15. Medicine, vol. 367, pp. 1381-1383. 26. Bradford, W. D. (2010). The association between individual time 41. Cohen DA, Babey SH. (2012) “Contextual Influences on Eating preferences and health maintenance habits. Medical Decision Behaviours: Heuristic Processing and Dietary Choices.” Obesity Making, 30(1), 99-112. Reviews, vol. 13, pp. 766-779. 27. Briss, P. A., L. E. Rodewald, A. R. Hinman, A. M. Shefer, R. A. Stri- 42. Cutler D. M, Everett (2010) W. Thinking outside the pillbox— kas, R. R. Bernier, V. G. Carande-Kulis, H. R. Yusuf, S. M. Ndiaye, Medication adherence as a priority for health care reform. N and S. M. Williams. 2000. “Reviews of Evidence Regarding In- Engl J Med.; 362(17):1553-5. terventions to Improve Vaccination Coverage in Children, Ado- 43. Davis, D. A., M. A. Thomson, A. D. Oxman, and R. B. Haynes. lescents, and Adults. The Task Force on Community Preventive (1995) “Changing Physician Performance. A Systematic Review Services.” American Journal of Preventive Medicine 18 (1 Suppl): of the Effect of Continuing Medical Education Strategies.” JAMA 97–140. 274 (9): 700–705. 28. Bruine de Bruin, W., Lefevre, C. E., Taylor, A. L., Dessai, S., Fisch- 44. Dawson, N. V., & Arkes, H. R. (1987). Systematic errors in medi- hoff, B., & Kovats, S. (2016). Promoting protection against a cal decision making. Journal of General Internal Medicine, 2(3), threat that evokes positive affect: The case of heat waves in the 183-187. United Kingdom. Journal of experimental psychology: applied, 45. De Los Monteros, K. E., & Gallo, L. C. (2011). The relevance of 22(3), 261. fatalism in the study of Latinas’ cancer screening behavior: A 29. Bryan, C.J., Mazar, N., Jamison, J.C., et al. (2017). Overcoming systematic review of the literature. International journal of be- behavioral obstacles to escaping poverty. Behavioral Science havioral medicine, 18(4), 310-318. & Policy, 3(1), 80-91. 46. de Walque D, Chukwuma A, Ayivi-Guedehoussou N, Koshka- 30. Burke, L. E., Styn, M. A., Sereika, S. M., Conroy, M. B., Ye, L., karyan M. (2020) Invitations, incentives, and conditions: a ran- Glanz, K., ... & Ewing, L. J. (2012). Using mHealth technology domized evaluation of demand-side interventions for health to enhance self-monitoring for weight loss: a randomized trial. screenings in Armenia. Policy Research Working Paper. World American journal of preventive medicine, 43(1), 20-26. Bank No.: 9346:1-32. 31. Busso, M., Cristia, J., & Humpage, S. (2015). Did you get your 47. DeFulio, A., & Silverman, K. (2012). The use of incentives to rein- shots? Experimental evidence on the role of reminders. Journal force medication adherence. Preventive medicine, 55, S86-S94. of health economics, 44, 226-237. 48. Devassy, S. M., Benny, A. M., Scaria, L., Nannatt, A., Fendt- 32. Camilloni L, Ferroni E, Cendales BJ, Pezzarossi A, Furnari G, Newlin, M., Joubert, J., ... & Webber, M. (2020). Social factors Borgia P, et al. Methods to increase participation in organised associated with chronic non-communicable disease and co- screening programs: A systematic review. BMC Public Health. morbidity with mental health problems in India: a scoping BioMed Central. 2013;13:1–16. review. BMJ open, 10(6), e035590. 33. Castaño, P. M., Bynum, J. Y., Andrés, R., Lara, M.,& Westhoff, C. 49. Devi, R., Kanitkar, K., Narendhar, R., Sehmi, K., & Subramaniam, (2012). Effect of daily text messages on oral contraceptive con- K. (2020). A narrative review of the patient journey through the tinuation: a randomized controlled trial. Obstetrics & Gynecol- lens of non-communicable diseases in low-and middle-in- ogy, 119(1), 14-20. come countries. Advances in Therapy, 37(12), 4808-4830. 34. Cawley, J., & Ruhm, C. J. (2011). The economics of risky health 50. Dexheimer, Judith W., Thomas R. Talbot, David L. Sanders, S. behaviors. In Handbook of health economics (Vol. 2, pp. 95- Trent Rosenbloom, and Dominik Aronsky. (2008) “Prompting 199). Elsevier. Clinicians about Preventive Care Measures: A Systematic Re- 35. Chang, L. L., DeVore, A. D., Granger, B. B., Eapen, Z. J., Ariely, D., view of Randomized Controlled Trials.” Journal of the American & Hernandez, A. F. (2017). Leveraging behavioral economics to Medical Informatics Association : JAMIA 15 (3): 311–20. improve heart failure care and outcomes. Circulation, 136(8), 51. Dieleman, M., Cuong, P. V., Anh, L. V., & Martineau, T. (2003). 765-772. Identifying factors for job motivation of rural health workers in 36. Chauhan, Bhupendrasinh F., Maya Jeyaraman, Amrinder Singh North Vietnam. Human resources for health, 1(1), 1-10. Mann, Justin Lys, Becky Skidmore, Kathryn M. Sibley, Ahmed 52. Doctor, J. N., Nguyen, A., Lev, R., Lucas, J., Knight, T., Zhao, H., & Abou-Setta, and Ryan Zarychanksi. 2017. “Behavior Change Menchine, M. (2018). Opioid prescribing decreases after learn- Interventions and Policies Influencing Primary Healthcare Pro- ing of a patient’s fatal overdose. Science, 361(6402), 588-590. fessionals’ Practice—an Overview of Reviews.” Implementation 53. Dupas, P. (2011). Health behavior in developing countries. Science 12 (1): 3. Annu. Rev. Econ., 3(1), 425-449. 37. Chokshi, N. P., Adusumalli, S., Small, D. S., Morris, A., Feingold, 54. East, K. A., Hitchman, S. C., McDermott, M., McNeill, A., Herbeć, J., Ha, Y. P., ... & Patel, M. S. (2018). Loss‐framed financial incen- A., Tountas, Y., ... & Vardavas, C. I. (2018). Social norms towards tives and personalized goal‐setting to increase physical activity smoking and electronic cigarettes among adult smokers in among ischemic heart disease patients using wearable devic- seven European Countries: findings from the EUREST-PLUS ITC es: the ACTIVE REWARD Randomized Trial. Journal of the Amer- Europe Surveys. Tobacco Induced Diseases , 16. ican Heart Association, 7(12), e009173. 55. East, K. A., Hitchman, S. C., McNeill, A., Thrasher, J. F., & Ham- 38. Chow, C. K., Teo, K. K., Rangarajan, S., Islam, S., Gupta, R., Avezum, mond, D. (2019). Social norms towards smoking and vaping A., ... & Yusuf, S. (2013). Prevalence, awareness, treatment, and and associations with product use among youth in England, control of hypertension in rural and urban communities in high-, Canada, and the US.Drug and Alcohol Dependence,205, 107635. middle-, and low-income countries. Jama, 310(9), 959-968. 56. East, K., McNeill, A., Thrasher, J. F., & Hitchman, S. C. (2021). So- 39. Cialdini, R. B. (2003). Crafting normative messages to protect cial norms as a predictor of smoking uptake among youth: a the environment. Current directions in psychological science, systematic review, meta‐analysis and meta‐regression of pro- 12(4), 105-109. spective cohort studies. Addiction, 116(11), 2953-2967. 40. Cohen DA, Babey SH (2012). “Candy at the Register—A Risk 57. Elgar, F. J., Roberts, C., Parry-Langdon, N., & Boyce, W. (2005). COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 6 181 Income inequality and alcohol use: a multilevel analysis of in financial incentive-based approaches to changing health drinking and drunkenness in adolescents in 34 countries. The behaviors: a meta-analysis. Am J Health Promotion;29:314–23. European Journal of Public Health, 15(3), 245-250. 76. Halim, A., Hasking, P., & Allen, F. (2012). The role of social drink- 58. Emanuel, E. J., Ubel, P. A., Kessler, J. B., Meyer, G., Muller, R. W., ing motives in the relationship between social norms and alco- Navathe, A. S., ... & Volpp, K. G. (2016). Using behavioral eco- hol consumption.  Addictive behaviors, 37(12), 1335-1341. nomics to design physician incentives that deliver high-value 77. Hall, M. A., E. Dugan, B. Zheng, and A. K. Mishra. (2001). “Trust in care. Annals of internal medicine, 164(2), 114-119. Physicians and Medical Institutions: What Is It, Can It Be Mea- 59. Evans-Lacko, S., Brohan, E., Mojtabai, R., & Thornicroft, G. sured, and Does It Matter?”The Milbank Quarterly 79 (4): 613–39. (2012). Association between public views of mental illness and 78. Hallsworth M, Chadborn T, Sallis A, Sanders M, Berry D, Greaves self-stigma among individuals with mental illness in 14 Euro- F, et al. (2016) Provision of social norm feedback to high pre- pean countries. Psychological medicine, 42(8), 1741-1752. scribers of antibiotics in general practice: a pragmatic national 60. Faries, M. D. (2016). Why we don’t “just do it” understanding the randomised controlled trial. The Lancet;387: 1743–1752. intention-behavior gap in lifestyle medicine. American journal 79. Hansen, H., Bourgois, P., & Drucker, E. (2014). Pathologizing pov- of lifestyle medicine, 10(5), 322-329. erty: New forms of diagnosis, disability, and structural stigma 61. Fingeret, M. C., Nipomnick, S., Guindani, M., Baumann, D., under welfare reform. Social science & medicine, 103, 76-83. Hanasono, M., & Crosby, M. (2014). Body image screening for 80. Hardy, H., Kumar, V., Doros, G., Farmer, E., Drainoni, M. L., Ry- cancer patients undergoing reconstructive surgery. Psycho‐on- bin, D., ...& Skolnik, P. R. (2011). Randomized controlled trial of cology, 23(8), 898-905. a personalized cellular phone reminder system to enhance ad- 62. Finkelstein EA, Haaland BA, Bilger M, Sahasranaman A, Sloan herence to antiretroviral therapy. AIDS patient care and STDs, RA, Nang EEK, Evenson KR. (2016) Effectiveness of activity 25(3), 153-161. trackers with and without incentives to increase physical ac- 81. Hart, J., Yadav, K., Szymanski, S., Summer, A., Tannenbaum, A., tivity (TRIPPA): a randomised controlled trial. Lancet Diabetes Zlatev, J., ... & Halpern, S. D. (2021). Choice architecture in phy- Endocrinol.;4(12):983–95. sician–patient communication: a mixed-methods assessments 63. Finkelstein, E. A., K. W. Tham, B. A. Haaland and A. Sahasranaman of physicians’ competency. BMJ quality & safety, 30(5), 362-371. (2017), ‘Applying economic incentives to increase effectiveness 82. Harte, E., MacLure, C., Martin, A., Saunders, C. L., Meads, C., Walter, of an outpatient weight loss program (TRIO) − A randomized F. M., ... & Usher-Smith, J. A. (2018). Reasons why people do not controlled trial’, Social Science & Medicine, 185(July): 63–70. attend NHS Health Checks: a systematic review and qualitative 64. Finkelstein, E. A., L. A. Linnan, D. F. Tate and B. E. Birken (2007), synthesis. British Journal of General Practice, 68(666), e28-e35. ‘A pilot study testing the effect of different levels of financial in- 83. Hatzenbuehler, M. L., & Link, B. G. (2014). Introduction to the centives on weight loss among overweight employees’, Journal special issue on structural stigma and health. Social Science & of Occupational and Environmental Medicine, 49(9): 981–989. Medicine. 65. Fiscella, K., Franks, P., Zwanziger, J., Mooney, C., Sorbero, M., & 84. Heller, D. J., Kumar, A., Kishore, S. P., Horowitz, C. R., Joshi, R., Williams, G. C. (2000). Risk aversion and costs. Journal of Family & Vedanthan, R. (2019). Assessment of barriers and facilitators Practice, 49(1), 12-12. to the delivery of care for noncommunicable diseases by non- 66. Fox, C. R., & Tversky, A. (1995). Ambiguity aversion and comparative physician health workers in low-and middle-income countries: ignorance. The quarterly journal of economics, 110(3), 585-603. a systematic review and qualitative analysis.  JAMA Network 67. Fox, C. R., & Weber, M. (2002). Ambiguity aversion, comparative Open, 2(12), e1916545-e1916545. ignorance, and decision context. Organizational behavior and 85. Higgins, Stephen T., Sheila M. Alessi, and Robert L. Dantona. human decision processes, 88(1), 476-498. 2002. “Voucher-Based Incentives: A Substance Abuse Treat- 68. Franks, P., Williams, G. C., Zwanziger, J., Mooney, C., & Sorbero, ment Innovation.” Addictive Behaviors, 27: 887- 910. M. (2000). Why do physicians vary so widely in their referral 86. Hollands, G. J., Shemilt, I., Marteau, T. M., Jebb, S. A., Kelly, M. P., rates?. Journal of general internal medicine, 15(3), 163-168. Nakamura, R., ... & Ogilvie, D. (2013). Altering micro-environ- 69. Friedman, Emily. 1994. “Money Isn’t Everything: Nonfinancial ments to change population health behaviour: towards an ev- Barriers to Access.” JAMA 271 (19): 1535–38. idence base for choice architecture interventions. BMC public 70. Galizzi, M. M., Miraldo, M., Stavropoulou, C., & Van Der Pol, M. health, 13(1), 1-6. (2016). Doctor–patient differences in risk and time preferences: 87. Holtgrave, D. R., Lawler, F., & Spann, S. J. (1991). Physicians’ risk A field experiment. Journal of health economics, 50, 171-182. attitudes, laboratory usage, and referral decisions: the case of 71. Gao, G., Burke, N., Somkin, C. P., & Pasick, R. (2009). Considering an academic family practice center. Medical Decision Making, culture in physician—patient communication during colorectal 11(2), 125-130. cancer screening. Qualitative health research, 19(6), 778-789. 88. Hou, M. Y., Hurwitz, S., Kavanagh, E., Fortin, J., & Goldberg, A. B. 72. Gilovich, T., Griffin, D., & Kahneman, D. (Eds.). (2002). Heuristics (2010). Using daily text-message reminders to improve adher- and biases: The psychology of intuitive judgment. Cambridge ence with oral contraceptives: a randomized controlled trial. University Press. Obstetrics & Gynecology, 116(3), 633-640. 73. Gong, E., Chukwuma, A., Ghazaryan, E., & de Walque, D. (2020). 89. Hunt, D. L., R. B. Haynes, S. E. Hanna, and K. Smith. 1998. “Ef- Invitations and incentives: a qualitative study of behavioral fects of Computer-Based Clinical Decision Support Systems nudges for primary care screenings in Armenia. BMC health on Physician Performance and Patient Outcomes: A Systematic services research, 20(1), 1-14. Review.” JAMA 280 (15): 1339–46. 74. Greene SJ, Butler J, Albert NM, DeVore AD, Sharma PP, Duffy CI, 90. John, L. K., Loewenstein, G., Troxel, A. B., Norton, L., Fassbend- et al. Medical therapy for heart failure with reduced ejection er, J. E., & Volpp, K. G. (2011). Financial incentives for extended fraction. J Am Coll Cardiol. 2018;72(4):351–66. weight loss: a randomized, controlled trial. Journal of general 75. Haff N, Patel M. S., Lim R, et al. (2015) The role of behavioral internal medicine, 26(6), 621-626. economic incentive design and demographic characteristics 91. Kahneman, D. (2011). Thinking, fast and slow. Macmillan. 182 CHAPTER 6 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E 92. Kahneman, D. & Tversky, A. (1979). Prospect theory: An analysis scribing: A Randomized Clinical Trial. JAMA Intern Med.;174: 425. of decision under risk. Econometrica, 47(2), 363-391. 111. Meeker D, Linder JA, Fox CR, Friedberg MW, Persell SD, Gold- 93. Kamvura, T. T., Dambi, J. M., Chiriseri, E., Turner, J., Verhey, R., & stein NJ, et al. (2016) Effect of Behavioral Interventions on Inap- Chibanda, D. (2022). Barriers to the provision of non-communi- propriate Antibiotic Prescribing Among Primary Care Practices: cable disease care in Zimbabwe: a qualitative study of primary A Randomized Clinical Trial. JAMA;315: 562. health care nurses. BMC nursing, 21(1), 1-12. 112. Milkman, K. L., Beshears, J., Choi, J. J., Laibson, D., & Madrian, 94. Keenan, K., Saburova, L., Bobrova, N., Elbourne, D., Ashwin, S., B. C. (2011). Using implementation intentions prompts to en- & Leon, D. A. (2015). Social factors influencing Russian male hance influenza vaccination rates. Proceedings of the National alcohol use over the life course: a qualitative study investigat- Academy of Sciences, 108(26), 10415-10420. ing age based social norms, masculinity, and workplace con- 113. Mitchell, M. S., Orstad, S. L., Biswas, A., Oh, P. I., Jay, M., Pakosh, text. Plos one, 10(11), e0142993. M. T., & Faulkner, G. (2020). Financial incentives for physical 95. Kessler, J. B., & Zhang, C. Y. (2014). Behavioral Economics and activity in adults: systematic review and meta-analysis. British Health. Paper for Oxford Textbook of Public Health. journal of sports medicine, 54(21), 1259-1268. 96. Kilpela, L. S., Becker, C. B., Wesley, N., & Stewart, T. (2015). Body 114. Möllenkamp, M., Zeppernick, M.,& Schreyögg, J. (2019). The image in adult women: Moving beyond the younger years. Ad- effectiveness of nudges in improving the self-management of vances in Eating Disorders: Theory, Research and Practice, 3(2), patients with chronic diseases: a systematic literature review. 144-164. Health Policy, 123(12), 1199-1209. 97. Kullgren, J. T., Troxel, A. B., Loewenstein, G., Asch, D. A., Norton, 115. Mugo, P. M., Wahome, E. W., Gichuru, E. N., Mwashigadi, G. M., L. A., Wesby, L., ... & Volpp, K. G. (2013). Individual-versus group- Thiong’o, A. N., Prins, H. A., ... & Sanders, E. J. (2016). Effect of text based financial incentives for weight loss: a randomized, con- message, phone call, and in-person appointment reminders trolled trial. Annals of internal medicine, 158(7), 505-514. on uptake of repeat HIV testing among outpatients screened 98. Kuntsche, E., Kuntsche, S., Thrul, J., & Gmel, G. (2017). Binge for acute HIV infection in Kenya: a randomized controlled trial. drinking: Health impact, prevalence, correlates and interven- PLoS One, 11(4), e0153612. tions. Psychology & health, 32(8), 976-1017. 116. Mullainathan, S., & Shafir, E. (2013). Scarcity: Why having too 99. LaBrie, J. W., Hummer, J. F., Huchting, K. K., & Neighbors, C. little means so much. New York, NY: Times Books. (2009). A brief live interactive normative group intervention 117. Murphy, S., Moore, G., Williams, A., & Moore, L. (2012). An ex- using wireless keypads to reduce drinking and alcohol conse- ploratory cluster randomised trial of a university halls of res- quences in college student athletes. Drug and alcohol review, idence based social norms intervention in Wales, UK. BMC 28(1), 40-47. public health, 12(1), 1-8. 100. Lagarde, M., Haines, A., & Palmer, N. (2007). Conditional cash 118. O’Donoghue, T., & Rabin, M. (1999). Doing it now or later. transfers for improving uptake of health interventions in low- American economic review, 89(1), 103-124. and middle-income countries: a systematic review. Jama, 119. Oettingen, G., Kappes, H. B., Guttenberg, K. B., and Gollwitzer, 298(16), 1900-1910. P. M. (2015). Self-regulation of time management: mental con- 101. Lee, C. S., & McCormick, P. A. (2003). Telephone reminders to re- trasting with implementation intentions. Eur. J. Soc. Psychol. duce non-attendance rate for endoscopy. Journal of the Royal 45, 218–229. doi: 10.1002/ejsp.2090. Society of Medicine, 96(11), 547-548. 120. Oettingen, G., Pak, H., and Schnetter, K. (2001). Self-regulation 102. Liebman, J., & Zeckhauser, R. (2008). Simple humans, complex of goal setting: turning free fantasies about the future into insurance, subtle subsidies (No. w14330). National Bureau of binding goals. J. Personal. Soc. Psychol. 80, 736–753. Economic Research. 121. Patel, M. S., Asch, D. A., Rosin, R., Small, D. S., Bellamy, S. L., Eber- 103. Linder JA, Meeker D, Fox CR, Friedberg MW, Persell SD, Gold- bach, K., ... & Volpp, K. G. (2016a). Individual versus team-based stein NJ, et al. Effects of Behavioral Interventions on Inappro- financial incentives to increase physical activity: a randomized, priate Antibiotic Prescribing in Primary Care 12 Months After controlled trial. Journal of general internal medicine, 31, 746-754. Stopping Interventions. JAMA. 2017;318: 1391. 122. Patel, M. S., Benjamin, E. J., Volpp, K. G., Fox, C. S., Small, D. S., 104. Lu, Y., & Goldman, D. (2010). The Effects of Relative Food Prices Massaro, J. M., ... & Murabito, J. M. (2017). Effect of a game- on Obesity--Evidence from China: 1991-2006 (No. w15720). based intervention designed to enhance social incentives to National Bureau of Economic Research. increase physical activity among families: the BE FIT random- 105. Luoto, J., & Carman, K. G. (2014). Behavioral economics guide- ized clinical trial. JAMA internal medicine, 177(11), 1586-1593. lines with applications for health interventions. Washington 123. Patel, M. S., Volpp, K. G., Rosin, R., Bellamy, S. L., Small, D. S., DC: Inter-American Development Bank. Fletcher, M. A., ... & Asch, D. A. (2016b). A randomized trial of so- 106. McCaffery, K., Waller, J., Nazroo, J., & Wardle, J. (2006). Social and cial comparison feedback and financial incentives to increase psychological impact of HPV testing in cervical screening: a physical activity. American Journal of Health Promotion, 30(6), qualitative study. Sexually transmitted infections, 82(2), 169-174. 416-424. 107. McCarthy CP, Murphy S, Cohen JA, et al. Underutilization of car- 124. Peek, M. E., Sayad, J. V., & Markwardt, R. (2008). Fear, fatalism diac rehabilitation for type 2 myocardial infarction. J Am Coll and breast cancer screening in low-income African-American Cardiol. 2019;25917:2005–7. women: the role of clinicians and the health care system. Jour- 108. McGill B, O’Hara BJ, Bauman A, et al. Are financial incentives for nal of general internal medicine, 23(11), 1847-1853. lifestyle behavior change informed or inspired by behavioral 125. Peel, N. M., McClure, R. J., & Bartlett, H. P. (2005). Behavioral economics? A mapping review. Am J Health Promot 2019;33. determinants of healthy aging. American journal of preventive 109. McGuire, T. G. (2000). Physician agency. Handbook of health medicine, 28(3), 298-304. economics, 1, 461-536. 126. Perron, N. J., Dao, M. D., Righini, N. C., Humair, J. P., Broers, B., 110. Meeker D, Knight TK, Friedberg MW, Linder JA, Goldstein NJ, Fox Narring, F., ... & Gaspoz, J. M. (2013). Text-messaging versus CR, et al. (2014) Nudging Guideline-Concordant Antibiotic Pre- telephone reminders to reduce missed appointments in an ac- COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 6 183 ademic primary care clinic: a randomized controlled trial. BMC atic review. BMC medical informatics and decision making, health services research, 13(1), 1-7. 16(1), 1-14. 127. Persell SD, Doctor JN, Friedberg MW, Meeker D, Friesema E, 144. Schneider, P., Popkin, B., Shekar, M., Eberwein, J. D., Block, C., Cooper A, et al. Behavioral interventions to reduce inappropri- & Okamura, K. S. (2020). Health and economic impacts of ate antibiotic prescribing: a randomized pilot trial. BMC Infect overweight/obesity. Health and Economic Consequences of an Dis. 2016;16: 373. Impending Global Challenge, 69. 128. Peters DH, Garg A, Bloom G, Walker DG, Brieger WR, et al. (2008) 145. Scott, I. A., Soon, J., Elshaug, A. G., & Lindner, R. (2017). Coun- Poverty and access to health care in developing countries. Ann tering cognitive biases in minimising low value care. Medical N Y Acad Sci 1136: 161–71. Journal of Australia, 206(9), 407-411. 129. Petry, N. M. (2011). Contingency management: what it is and 146. Sen, A., Huffman, D., Loewenstein, G., Asch, D. A., Kullgren, J. why psychiatrists should want to use it. The psychiatrist, 35(5), T., & Volpp, K. G. (2017). Do Financial Incentives Reduce Intrin- 161-163. sic Motivation for Weight Loss? Evidence from Two Tests of 130. Petty, N.M., and B. Martin. 2002. “Low-Cost Contingency Man- Crowding Out. Retrieved from http://repository.upenn.edu/ agement for Treating Cocaineand Opiod-Abusing Metadone hcmg_papers/8. Patients.” Journal of Consulting and Clinical Psychology, 70(2): 147. Serrano Fuentes, N., Rogers, A., & Portillo, M. C. (2019). So- 398-405. cial network influences and the adoption of obesity-related 131. Pop-Eleches, C., Thirumurthy, H., Habyarimana, J. P., Zivin, J. behaviours in adults: a critical interpretative synthesis re- G., Goldstein, M. P., De Walque, D., ... & Bangsberg, D. R. (2011). view. BMC public health, 19(1), 1-20. Mobile phone technologies improve adherence to antiretro- 148. Skov, L. R., Lourenco, S., Hansen, G. L., Mikkelsen, B. E., & Schof- viral treatment in a resource-limited setting: a randomized ield, C. (2013). Choice architecture as a means to change controlled trial of text message reminders. AIDS (London, En- eating behaviour in self‐service settings: a systematic review. gland), 25(6), 825. Obesity Reviews, 14(3), 187-196. 132. Prendergast, Michael, Deborah Podus, John Finney, Lisa Green- 149. Shiv, B., & Fedorikhin, A. (1999). Heart and mind in conflict: The well, and John Roll. 2006. “Contingency management for treat- interplay of affect and cognition in consumer decision mak- ment of substance use disorders: a meta-analysis.” Addiction, ing. Journal of consumer Research, 26(3), 278-292. 101(11): 1546–1560. 150. Stockwell, M. S., Kharbanda, E. O., Martinez, R. A., Vargas, C. Y., 133. Rai, S. S., Syurina, E. V., Peters, R. M., Putri, A. I., & Zweekhorst, Vawdrey, D. K., & Camargo, S. (2012). Effect of a text messaging M. B. (2020). Non-communicable diseases-related stigma: a intervention on influenza vaccination in an urban, low-income mixed-methods systematic review. International journal of en- pediatric and adolescent population: a randomized controlled vironmental research and public health, 17(18), 6657. trial. Jama, 307(16), 1702-1708. 134. Ridout, B., & Campbell, A. (2014). Using F acebook to deliver a 151. Strulik, H., & Werner, K. (2021). Time-inconsistent health behav- social norm intervention to reduce problem drinking at univer- ior and its impact on aging and longevity. Journal of Health sity. Drug and alcohol review, 33(6), 667-673. Economics, 76, 102440. 135. Riegel, B., Stephens‐Shields, A., Jaskowiak‐Barr, A., Daus, M., & 152. Suri, G., Sheppes, G., Schwartz, C., & Gross, J. J. (2013). Patient Kimmel, S. E. (2020). A behavioral economics‐based telehealth inertia and the status quo bias: when an inferior option is pre- intervention to improve aspirin adherence following hospital- ferred. Psychological science, 24(9), 1763-1769. ization for acute coronary syndrome. Pharmacoepidemiology 153. Taylor, Michael J., Ivo Vlaev, John Maltby, Gordon DA Brown, and drug safety, 29(5), 513-517. and Alex M. Wood. “Improving social norms interventions: 136. Ritchie, P. D., Jenkins, M., & Cameron, P. A. (2000). A telephone Rank-framing increases excessive alcohol drinkers’ informa- call reminder to improve outpatient attendance in patients tion-seeking.” Health psychology 34, no. 12 (2015): 1200. referred from the emergency department: a randomised con- 154. Thaler, R. H., Sunstein, C. R., & Balz, J. P. (2013). Choice archi- trolled trial. Australian and New Zealand journal of medicine, tecture. In E. Shafir (Ed.), The behavioral foundations of public 30(5), 585-592. policy (pp. 428-439). Princeton, NJ: Princeton University Press. 137. Ritov, I., & Baron, J. (1992). Status-quo and omission biases. 155. Thapa, J. R., & Lyford, C. P. (2014). Behavioral economics in the Journal of risk and uncertainty, 5(1), 49-61. school lunchroom: can it affect food supplier decisions? A sys- 138. Ruhm, C. J. (2012). Understanding overeating and obesity. tematic review. International Food and Agribusiness Manage- Journal of Health economics, 31(6), 781-796. ment Review, 17(1030-2016-82992), 187-208. 139. Sacarny A, Barnett ML, Le J, Tetkoski F, Yokum D, Agrawal S. Ef- 156. Toll BA & Ling PM. The Virginia Slims identity crisis: an inside fect of Peer Comparison Letters for High-Volume Primary Care look at tobacco industry marketing to women.  Tob Con- Prescribers of Quetiapine in Older and Disabled Adults: A Ran- trol 2005;14:172–80. domized Clinical Trial. JAMA Psychiatry. 2018;75: 1003. 157. Torrente, F., Bustin, J., Triskier, F., Ajzenman, N., Tomio, A., Mastai, 140. Sacarny A, Yokum D, Finkelstein A, Agrawal S. Medicare Letters R., & Boo, F. L. (2020). Effect of a Social Norm Email Feedback To Curb Overprescribing Of Controlled Substances Had No De- Program on the Unnecessary Prescription of Nimodipine in tectable Effect On Providers. Health Affairs. 2016;35: 471–479. Ambulatory Care of Older Adults: A Randomized Clinical Trial. 141. Salami JA, Warraich H, Valero-Elizondo J, et al. National trends JAMA network open, 3(12), e2027082-e2027082. in statin use and expenditures in the US adult population from 158. Tversky, A; D. Kahneman, (1973). Availability: A heuristic for 2002 to 2013: insights from the medical expenditure panel sur- judging frequency and probability. Cognitive Psychology. 5, vey. JAMA Cardiol. 2017;2(1):56–65. 207–232. 142. Samuelson, W., & Zeckhauser, R. (1988). Status quo bias in deci- 159. Tversky, A. and Kahneman, D., (1974). Judgment under Uncer- sion making. Journal of risk and uncertainty, 1(1), 7-59. tainty: Heuristics and Biases: Biases in judgments reveal some 143. Saposnik, G., Redelmeier, D., Ruff, C. C., & Tobler, P. N. (2016). heuristics of thinking under uncertainty.  Science,  185(4157), Cognitive biases associated with medical decisions: a system- pp.1124-1131. 184 CHAPTER 6 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E 160. van der Weijden, T. (2007). van SB, Stoffers HE, Timmermans DR, tional journal of epidemiology, 40(3), 712-718. Grol R. Primary prevention of cardiovascular diseases in general 167. Wang, S. Y., & Groene, O. (2020). The effectiveness of behav- practice: mismatch between cardiovascular risk and patients’ ioral economics-informed interventions on physician behav- risk perceptions. Med Decis Making, 27(6), 754-761. ioral change: A systematic literature review. PloS one, 15(6), 161. Vervloet, M., Linn, A. J., van Weert, J. C., De Bakker, D. H., Bouvy, e0234149. M. L., & Van Dijk, L. (2012). The effectiveness of interventions 168. Watkins et al. (2022). “NCD Countdown 2030: efficient path- using electronic reminders to improve adherence to chronic ways and strategic investments to accelerate progress towards medication: a systematic review of the literature. Journal of the the Sustainable Development Goal target 3.4 in low-income American Medical Informatics Association, 19(5), 696-704. and middle-income countries.” The Lancet 399, no. 10331: 162. Vlaev, I., King, D., Darzi, A. and Dolan, P., 2019. Changing health 1266-1278. behaviors using financial incentives: a review from behavioral 169. Weinstein, N. D. (1980). Unrealistic optimism about future life economics. BMC public health, 19(1), pp.1-9. events. Journal of personality and social psychology, 39(5), 806. 163. Volpp, K.G., Troxel, A.B., Pauly, M.V., Glick, H.A., Puig, A., Asch, 170. Weller, D. P., Patnick, J., McIntosh, H. M., & Dietrich, A. J. (2009). D.A., Galvin, R., Zhu, J., Wan, F., DeGuzman, J. and Corbett, E., Uptake in cancer screening programmes. The lancet oncology, 2009. A randomized, controlled trial of financial incentives for 10(7), 693-699. smoking cessation. N Engl J Med, 360, pp.699-709. 171. Wexler R, Elton T, Pleister A, Feldman D (2009) Barriers to blood 164. Volpp KG, John LK, Troxel AB, Norton L, Fassbender J, Loewen- pressure control as reported by african american patients. J stein G. (2008a). Financial incentive–based approaches for Natl Med Assoc 101 (6) 597–603. weight loss: a randomized trial. JAMA. 2008;300(22):2631–7. 172. Witte, K., & Allen, M. (2000). A meta-analysis of fear appeals: 165. Volpp, K.G., Loewenstein, G., Troxel, A.B., Doshi, J., Price, M., Implications for effective public health campaigns. Health ed- Laskin, M. and Kimmel, S.E., (2008b). A test of financial incen- ucation & behavior, 27(5), 591-615. tives to improve warfarin adherence. BMC health services re- 173. Wittleder, S., Kappes, A., Oettingen, G., Gollwitzer, P. M., Jay, M., search, 8(1), pp.1-6. & Morgenstern, J. (2019). Mental contrasting with implemen- 166. Von Wagner, C., Baio, G., Raine, R., Snowball, J., Morris, S., At- tation intentions reduces drinking when drinking is hazardous: kin, W., ... & Wardle, J. (2011). Inequalities in participation in an an online self-regulation intervention. Health Education & Be- organized national colorectal cancer screening programme: havior, 46(4), 666-676. results from the first 2.6 million invitations in England. Interna- COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 6 185 ANNEX 6.1 TABLE 6A.1  Summary of Behavioral Economics Concepts Concept Definition Example Present bias Tendency to value instant gratification more than Patient decides to consume an unhealthy but delicious meal future benefits now despite knowing that it contributes to poor health outcomes in the future Time inconsistency Tendency to be impatient when choosing be- tween receiving benefits today and receiving them in the future, but patient when choosing between benefits at two different moments in the future Intention-action gap Patient understands the importance of symptom monitor- ing but does not do it Overconfidence Overestimation of own performance and over- Patient understands that exercise is beneficial, but that placement of own performance relative to others. knowledge fails to lead to increased physical activity Over-optimism Tendency to underestimate the probability of negative events and overestimate the probability of positive events. Availability heuristic Tendency to judge the probability of a future event Patient immediately following the death of an overweight occurring based on the ease with which an occur- friend from a heart attack may be more likely to eat healthier, rence of such an event comes to mind but over time, as the recency and vividness fades, such habits dissipate. Cognitive dissonance Tendency to often hold contradictory opinions People smoke even though they have read countless times that it is dangerous. This self-contradiction causes dissonance, so people need to develop cognitive tools to justify the contradic- tion. For example, by overly exaggerating how much pleasure they get out of smoking underestimating the future costs Loss aversion Loss causes distress that is greater than the happi- ness caused by a gain of the same size Endowment effect The overvaluation of a good when we possess it Framing effects Tendency to draw different conclusions depending Food options or incentives can be presented in a way that on how the information is presented highlights the positive or negative aspects of a decision, leading an option to be perceived as more or less attractive Anchoring An initial exposure to a certain number or attribute When faced with a decision under uncertainty, individuals serves as point of reference and impacts subse- attribute to much weight to the initial exposure, which, with- quent judgements out further awareness, distorts estimates and judgements Cognitive overload (choice The cognitive load is the amount of mental effort and decision fatigue) and memory used at a given moment in time. Overload is when the volume of information or choices provided exceeds an individual’s capacity to process it. Social norms The unwritten rules governing behavior within a society Peer effects The influence exerted by a peer group on its individual members to fit in with or conform to the group’s norms and expectations Stigma The negative social attitude attached to a charac- teristic of an individual that may be regarded as a mental, physical, or social deficiency and can lead unfairly to discrimination against and exclusion of the individual. Source: Original for this publication with input from Chang et al. (2017); Emanuel et al. (2016). 7 186 CHAPTER 7 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E Taxation of harmful products, including tobacco, alcohol and sugar-sweetened beverages (SSBs), and related topics Guillermo Paraje a, Prabhat Jha b, William Savedoff c, Alan Fuchs d a Escuela de Negocios, Universidad Adolfo Ibáñez b Centre for Global Health Research, Unity Health Toronto, University of Toronto c Social Insight d Poverty and Equity, Middle East and North Africa, World Bank COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 7 187 INTRODUCTION THE CONSUMPTION OF UNHEALTHY PRODUCTS AND THE HEALTHY LONGEVITY AGENDA Even as human beings are living longer and generally healthier lives, consumption of tobacco, alcohol, and pro- cessed foods with high sugar content are causing substantial death and disease. In recent decades, the magnitude of these health effects has become increasingly apparent in terms of cancers, cardiovascular disorders, diabetes, and other illnesses which are obstacles to healthy aging. Consumption of tobacco, alcohol, and sugar-sweet- ready-strained pension systems, health systems, ened beverages (SSBs) have negative consequences and family members providing informal care for on health for people of all ages. However, these neg- older people in poor health (Bloom et al. 2015). Fis- ative effects are especially evident at older ages. The cal tools (e.g., taxes) like the ones discussed in this main contributors to the disease burden in people 60 document can play a fundamental role in prevent- years of age and older at the global level are cardio- ing chronic diseases and improving people’s lives. vascular diseases (30.3 percent), cancer (15.1 per- This chapter discusses taxation as a proven cent), and chronic respiratory diseases (9.5 percent), method to discourage consumption of harmful all causes closely associated with the consumption of products, improve population health, and contrib- tobacco, alcohol, and SSBs (Prince et al. 2015). ute to healthy aging. It discusses the economic ra- In a rapidly aging world, it is crucial to improve tionale for taxing these products and documents the health of older adults. 23 percent of the total their health effects and costs to society. It presents burden of disease is attributable to illness in adults evidence on the effectiveness of taxation in raising ages 60 and older (Prince et al. 2015). Suffering prices and reducing consumption. Taking account chronic illnesses is strongly associated with func- of health benefits as well as indirect financial effects, tional dependence. In Latin America, for instance, it also shows how taxation disproportionately ben- 14 percent of the over-65 population are care de- efits poorer households. The chapter considers the pendent, while functional dependence is negatively role of industries in opposing taxation and con- associated with socioeconomic status (Aranco et al. cludes with a discussion of how this topic relates to 2022). A recent study found that labor income re- the broader agenda of healthy longevity. ceived by members of households with older adults For the purpose of this review, tobacco prod- needing chronic care was substantially lower than ucts are defined as those that contain tobacco that in those households where such care was not need- can be smoked, inhaled, or chewed. So-called “elec- ed. Household consumption was also about 12 per- tronic cigarettes” (electronic nicotine delivery sys- cent lower and catastrophic health care costs about tems, or electronic non-nicotine delivery systems) 64 percent more likely in households where older raise additional issues beyond the scope of this adults needed care (Guerchet et al. 2018). study. With respect to sugar, this review will focus Unless countries reduce smoking, alcohol use, on SSBs. Because SSBs are a significant source of and SSB consumption, people reaching old age in added sugars in people’s diets and lack other nutri- the coming decades will increasingly experience tional benefits, many governments have initiated preventable chronic illnesses, with the burden that efforts to restrict sugar intake by focusing on these this implies for health systems and caregivers, espe- beverages. Researchers have been able to generate cially women. Improving the health status of older evidence about reducing sugar consumption by adults can produce social benefits in the form of studying such policies. SSBs are defined as all types an extended productive life span (comprehensive of beverages containing free sugars. These include longevity boost) and a reduced fertility rate (Hei- carbonated and non-carbonated soft drinks, fruit/ jdra and Reijnders 2016). Action toward these vegetable juices and drinks, liquid and powder con- goals must unfold across the life course. The years centrates, flavored water, energy and sports drinks, gained through increased life expectancy need to ready-to-drink tea, ready-to-drink coffee, and fla- be as healthy as possible to avoid burdening al- vored milk drinks (WHO 2017b). 188 CHAPTER 7 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E THE CONSUMPTION OF HARMFUL PRODUCTS AND THEIR EFFECTS Tobacco, alcohol, and sugar have been around a long time. However, current consumption patterns – driven by rising incomes, falling prices, and easier access – have led to higher levels and more widespread use of these products. The effects on population health and economic activity are increasingly well documented. Consumption patterns Though the share of current drinkers among peo- Consumption of tobacco, alcohol, and SSBs is asso- ple aged 15 years and older decreased between 2000 ciated with negative health consequences. Scientif- and 2016 (from 47.6 percent to 43 percent percent), ic evidence of these impacts is consistent, signifi- the per capita volume of pure alcohol consumed has cant, and growing. increased from 5.7 liters to 6.4 liters and is project- It is estimated that tobacco kills between 5 ed to grow further to 7 liters by 2025 (WHO 2018; to 8 million people in the world each year, most- Manthey et al. 2019). If only drinkers are considered, ly in low- and middle-income countries (LICs and pure alcohol consumption per capita reached 15.1 li- MICs), along with a more uncertain but high num- ters in 2016 (up from 11.1 liters in 2000) – (WHO ber of second-hand (passive) smokers (WHO 2021; 2018). Considering global population growth in the Jha and Peto 2014). Highly toxic and addictive, to- period 2000-2016, the number of drinkers world- bacco has posed one of the main public health chal- wide increased by 16 percent, and the amount of lenges of the second half of the 20th century and of pure alcohol consumed globally increased by 58 the 21st. Although tobacco consumption remains percent (an average annual increase of 3.1 percent very high, it is falling, marking a reversal of the versus a 1.7 percent average annual increase in glob- rising trends seen before the late 1990s (Jha 1999). al constant dollar per capita GDP) – (United Nations Relying on direct surveys for some countries and 2019; World Bank 2023). Consumption of spirits model-based estimates for most others, the World contributed the most to increasing consumption of Health Organization (WHO) estimates that global pure alcohol at the global level (44.8 percent). Beer smoking prevalence in people aged 15 years and and wine followed, with increases of 34.3 percent older fell by a third between 2000 and 2020, from and 11.7 percent, respectively (WHO 2018). about 33.3 percent to 22.8 percent (WHO 2019b). Global alcohol consumption prevalence was Differences by gender are notable. Current-smoking about 70 percent higher among men ages 15 years prevalence for men fell by a quarter (50 percent to and older than among women in the same age group 37.5 percent), while for women, it dropped by half in 2016 (53.6 percent versus 32.3 percent, respective- (16.7 percent to 8 percent). In some countries, such ly), while the consumption of pure alcohol among as China, India, and Indonesia, cigarette smoking drinkers ages 15 years and older is 180 percent high- is largely concentrated among men, but in others, er among men than among women (19.4 versus 7 li- including many Latin American countries, smoking ters) – (WHO 2018). However, the gender gap in al- among women is common. Adopting United Na- cohol consumption has decreased significantly over tions (UN) population projections (United Nations the last century, and there are strong indications that 2019), these figures imply that the absolute number the gap will continue to narrow, due to faster growth of current smokers worldwide fell almost 7 percent in consumption among women (Slade et al. 2016). between 2000 and 2020, with a decrease of 35 per- Patterns of alcohol consumption have also cent for women, but an increase of 2 percent for changed, but with greater regional variation and far men. Decreases in both prevalence and number of less reliable survey data available than for cigarettes. smokers are projected for both sexes and for all age The prevalence of heavy episodic drinking (HED)1 groups through 2025 (WHO 2019b). It is estimated among drinkers 15 years and older declined by only that one smoking death occurs for every 0.8 to 1.1 about 1 percent per year from 2010 to 2016, from million cigarettes smoked (Jha 2020). Thus, the 7.4 41.9 percent, to 39.5 (WHO 2018). HED is more trillion cigarettes consumed in 2019 alone will lead, prevalent among men (50.2 percent versus 19.9 per- eventually, to at least 7 million premature deaths. cent for women in 2016), though this gap declined For alcohol, the situation is quite different. sharply, from 3.8 times higher in 2010 to 1.5 times 1   HED is defined as 60 or more grams of pure alcohol on at least one single occasion at least monthly. COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 7 189 higher in 2016 (WHO 2018; WHO 2014). product prices and consumers’ incomes determines In the case of SSBs, consumption has been in- product affordability. If consumers’ incomes increase creasing, though there are few estimates of the evo- faster than the prices of certain goods, then indi- lution of consumption at the global level. Global viduals would have to devote a lower proportion of consumption for carbonated soft drinks, excluding their income to purchase these goods. Under most bottled still or carbonated water, fruit or vegetable conditions (i.e., normal goods) this would lead to an juices, coffee, tea, or sports drinks, increased from 36 increase in demand for these products. liters per person per year in 1997 to 43.1 liters in 2010 Directly measuring the affordability of key (Basu et al. 2013). Another study estimated that per health-damaging products is difficult and costly. capita SSB and fruit juice consumption in the pop- One of the many measures of affordability often ulation aged 20 years and older was 1.22 liters per relied upon instead is the Relative Income Price week in 2010 (Singh et al. 2015). This was 30 percent (RIP). The RIP measures what proportion of per higher than the weekly consumption of milk. Con- capita GDP would be required to buy 100 packs of sumption was similar for men and women but high- cigarettes or 100 liters of beer or SSBs (Blecher and er among younger people. For example, people who van Walbeek 2009; Blecher et al. 2017, 2018). Based were 20-39 years old consumed 139 percent more on the RIP, the affordability of tobacco over the last of these drinks than those ages 60 years and older. three decades (1990-2006) has varied a great deal. Upper-middle-income countries (UMICs) had the In HICs, tobacco has generally become less afford- highest per capita consumption for persons 20 years able (though there are variations among these coun- and older at 1.7 liters per person per week, while tries), while in LICs and MICs tobacco has general- high-income countries (HICs) ranked second with ly become more affordable i.e., a smaller fraction of 1.4 liters per week. Lower-middle-income countries per capita GDP is needed to buy 100 packs of cig- (LMICs) and LICs had 1.3 and 0.7 liters per week, re- arettes (Blecher and van Walbeek 2009). Between spectively. In the case of LICs, consumption of SSBs 2008 and 2018, the RIP increased in 117 out of and fruit juices was 27 percent higher than consump- 168 countries, indicating that tobacco became less tion of milk. Recent estimates suggest that consump- affordable in those countries. In 42 of those coun- tion of SSBs in HICs will stabilize (or decrease slight- tries the tobacco RIP increased more than 50 per- ly) by 2022, albeit at a very high consumption level cent, and in 26 countries it increased more than 100 (about 120 liters per capita per year). Per capita SSB percent (Paraje and Stoklosa 2017). consumption will also stabilize for UMICs (at about For alcohol, trends are clearer: alcohol has be- 50 liters per capita per year) but increase strongly for come significantly more affordable over the last few LICs and LMICs (World Bank 2020). decades. Between 1990 and 2016, the RIP for beer fell in 30 out of 37 HICs, and in 42 out of 44 LICs Price and affordability and MICs (Blecher et al. 2018). In almost half of the countries, affordability increased both because of Many variables affect consumption, including in- an increase in income and a decrease in beer pric- come and income distribution; advertising and prod- es. The same trend and pattern have been found for uct promotion; changes in social activities, religious a sample of nine Latin American countries (Paraje practices, and cultural norms; along with prices and and Pincheira 2018). affordability. Governments have had success in dis- Affordability has also increased for SSBs in the couraging consumption of health-damaging prod- 1990-2016 period. In 79 of 82 countries (40 of them ucts through a range of policies. These include adver- HICs), the RIP significantly decreased (i.e., less in- tisement and marketing regulations; sales restrictions come was needed to buy 100 liters of Coca-Cola, (e.g., prohibiting sales to minors or in certain places/ for example), mostly due to a higher rate of income times); regulation of places for consumption (e.g., growth, rather than a fall in the real prices of SSBs smoking bans in public spaces); and packaging re- (Blecher et al. 2017). This evidence is consistent quirements (e.g., graphic warnings, front-of-pack with findings in Latin America, where 7 of 12 coun- labeling). However, evidence shows that, for tobac- tries studied showed an increase in the affordability co, alcohol, and SSBs, own-prices and consumers’ of SSBs – in some countries by more than 6 percent incomes are among the main determinants of con- per year – and a decline in affordability in only one sumption (WHO 2017a). The relationship between country: Mexico (Paraje and Pincheira 2018). 190 CHAPTER 7 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E Health impact of the consumption of tobacco, alcohol, and SSBs as other conditions with potentially serious lifelong health consequences, such as low birthweight. A Evidence on the health burden linked to the con- recent meta-analysis finds that smoking is a signifi- sumption of tobacco and alcohol is overwhelming, cant contributor to hospitalization and death from and it is solid and growing for SSBs. Direct epide- COVID-19 infection (Reddy 2021). Though tobacco miological studies in several countries, paired with is one of the biggest killers among the elderly, at least model-based estimates, have led the Global Burden half of tobacco deaths occur in middle age (30-69 of Disease (GBD) Project to estimate that, in 2020, years), leading to a loss of two or more decades of tobacco smoking accounted for 7.7 million deaths life. On average, smoking kills about half of regular globally, of which 87 percent were among current smokers, with an average loss of life for these persons smokers and 80 percent among men (IHME 2020). of about 20 years, or about 10 years for regular smok- In terms of disability-adjusted life years (DALYs), ers overall (Jha and Peto 2014; Jha et al. 2013). tobacco was responsible for 9 percent of DALYs at On the other hand, early cessation can reverse the global level in 2019 – more than 229 million most of the harms of early tobacco use. It has been DALYs (IHME 2020). found that adults who quit smoking at 25 to 34, 35 Alcohol caused an estimated 2.4 million deaths to 44, or 45 to 54 years of age gained about 10, 9, and worldwide in 2019 or 4.3 percent of total risk-re- 6 years of life, respectively, as compared with those lated deaths (IHME 2020). Almost 85 percent of who continued to smoke until their (premature) those deaths were among men. DALYs due to al- death (Jha et al. 2013). cohol represented 3.7 percent of total risk related Alcohol consumption is associated with neu- DALYs. Alcohol was the main cause of death in the ropsychiatric conditions; unintentional and inten- 10-24 year age group (22 percent of all risk-relat- tional injuries; gastrointestinal diseases; several ed deaths), and in some countries (e.g., the United types of cancer; vascular diseases; maternal and States and Scotland) alcohol is identified as one of perinatal conditions; and other negative health im- the leading causes of “deaths of despair”(Substance pacts (Rehm and Monteiro 2005; Rehm et al. 2009; Abuse and Mental Health Services Administration Rehm et al. 2017; Griswald et al. 2018).  Recent (US) and Office of the Surgeon General (US) 2016; studies have shown that, contrary to what some ear- Case and Deaton 2021). Alcohol consumption is lier studies reported, there is no safe, “protective,” directly related to crime and domestic violence, or healthy consumption of alcohol (Griswald et al. which dramatically affect community life and well- 2018; Topiwala et al. 2021; Millwood et al. 2019). being among minors and women. Consumption of SSBs is also associated with Finally, diets high in SSBs were responsible for negative health effects, especially because of the 242,000 deaths at the global level in 2019 (0.4 per- strong link between consumption of SSBs and over- cent of total risk-related deaths) and for more than weight/obesity (Vartanian et al. 2007). Prominent 6 million DALYs (0.25 percent of total risk-related experts in public health have signaled the consump- DALYs) – (IHME 2020). tion of SSBs in particular populations, most notably Tobacco consumption has been causally linked in the Americas, as “the single largest driver of the to “diseases of nearly all organs of the body, to di- obesity epidemic”(Brownell and Frieden 2009). Evi- minished health status, and to harm to the fetus.”(- dence also shows a positive association between SSB Jha 2020; U.S. Department of Health and Human consumption and cardiovascular diseases, type 2 di- Services 2014). Main diseases caused by tobacco abetes mellitus, obesity-related cancer, dental caries, (especially manufactured cigarettes and bidis) are metabolic syndrome, and osteoporosis, among other circulatory diseases (such as ischemic heart diseas- diseases (Imamura et al. 2015; Wang et al. 2015; Yin es and stroke); respiratory diseases (such as COPD et al. 2021; Narain et al. 2016; Hodge et al. 2018). and asthma); a wide array of cancers (lung, trachea, Overall, there is ample evidence that the con- esophagus, bladder, colon, stomach, liver, kidney, and sumption of tobacco, alcohol, and SSBs can severe- others); metabolic conditions (diabetes mellitus, im- ly affect health and wellbeing of people of all ages. paired immune system, and others); infertility; and Children and young adults exposed to these prod- tuberculosis (Jha 2020; U.S. Department of Health ucts can develop chronic illnesses and conditions and Human Services 2014). Tobacco is also the cause that may permanently affect their future productive of permanent adverse consequences for brain devel- capabilities and increase the individual and social opment, if fetuses are exposed to nicotine, as well financial burden of treating those illnesses. Negative COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 7 191 health consequences from consuming these prod- ing-attributable diseases represent about 6.9 percent ucts are present at all ages, even if not immediately of government health budgets in these countries. observed. When manifested at early stages of life, Uruguay and Bolivia were among the most affected, they have long-lasting consequences, while, emerg- spending 16.7 percent and 11.8 percent of their gov- ing in middle or old age, they harm people’s pros- ernment health budgets, respectively, on tobacco-re- pects for healthy longevity. lated costs (Pichon-Riviere 2020). In the case of alcohol, fewer studies are available, Economic costs but those that have been conducted for several coun- tries all show high economic costs. Pooled results for The consumption of products with significant neg- 29 locations (countries, states/regions, cities) found ative health impacts has economic implications. that the total cost of alcohol consumption amounted These consequences can be broadly separated into to US$ PPP 817 per adult, equivalent to 1.5 percent two groups of costs: (i) direct costs attributable to of GDP (of which 68 percent were indirect costs) - the treatment of diseases; and (ii) indirect costs (Manthey et al. 2021). Comparable studies for HICs that can be attributed to premature mortality, loss show that total per capita economic costs related to of productivity due to absenteeism and presentee- alcohol range from US$ PPP 384 in France to US$ ism (i.e., people going to their jobs but being less PPP 837 in the United States, with a weighted av- productive), and opportunity costs for caregivers, erage for four countries (France, the United States, as well as pain and suffering. In the case of alcohol, Canada, and Scotland) of US$ PPP 725 (Rehm et al. for instance, indirect costs also include violence, 2009). On average, 72 percent of those costs are due crime, destruction of property, costs of policing, to lost productivity, while 13 percent are health care prosecution, incarceration, domestic violence, and costs. In two MICs (South Korea and Thailand), the various forms of suffering experienced by drinkers weighted average of total per capita costs was US$ and others. Present economic costs and health costs PPP 293, with 79 percent of that amount attribut- may also have an impact on future economic costs able to lost productivity and 5.6 percent to health and wellbeing, as present illnesses may condition care costs (Rehm et al. 2009). In another study of 14 future incomes. In the case of tobacco, smokers’ HICs, the weighted average of alcohol-related social households also face economic costs reflecting fac- costs was equivalent to 1.6 percent of GDP, with in- tors that include but extend beyond the many annu- dividual country values ranging from 0.6 percent in al deaths from second-hand smoke. These indirect Spain to 2 percent in New Zealand (Mohapatra et al. costs are considered below. 2010). In these countries, indirect costs represented It has been estimated that the economic cost of 72 percent of total social costs. The few studies avail- tobacco consumption at the global level was equiv- able for other low- and middle-income countries are alent to 1.8 percent of world GDP in 2012 - about consistent with these findings. In Chile, the social US$ PPP 1.8 trillion (Goodchild, Nargis, and Tursan costs of alcohol consumption were equivalent to 0.9 d’Espaignet 2017). Of that amount, direct costs (re- percent of GDP in 2014, of which 30 percent were lated to health care expenditures) totaled USD PPP health care-related costs (Margozzini et al. 2018). 467 billion (equivalent to 5.7 percent of global health In South Africa, a study estimated alcohol-related expenditures). In LICs and MICs, direct costs can be costs to be equivalent to between 10 and 12 percent large when compared to total health care costs. For of GDP (Matzopoulos et al. 2013). instance, in Bangladesh and Malaysia, direct care We are unaware of studies estimating economic costs linked to tobacco-attributable diseases have costs specifically for the consumption of SSBs. How- been estimated at 19.8 percent and 16.8 percent of ever, several studies have estimated the economic all health spending, respectively (U.S. National Can- costs of conditions and illnesses strongly associated cer Institute and World Health Organization 2016). with SSB consumption, such as obesity. At the global Total - direct and indirect - costs can also be high, level, it has been estimated that obesity costs US$2 reaching as much as 7.2 percent of GDP in the Phil- trillion per year, or the equivalent of 2.8 percent of ippines, though in most cases the costs range from global GDP (Dobbs et al. 2014). Estimates for 42 0.1 percent to 2 percent of GDP (U.S. National Can- countries in Asia and the Pacific put the economic cer Institute and World Health Organization 2016). costs of overweight and obesity at 0.8 percent of the A recent study conducted in 12 Latin American group’s GDP (Helble and Francisco 2017). Direct countries found that health care costs for smok- costs can reach up to 37 percent of total health costs 192 CHAPTER 7 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E (in Singapore), and indirect costs can reach up to 67 crowding-in of harmful behaviors (Wu et al. 2021). percent of total health costs (in Palau). It has been argued that these effects on household Apart from economic costs related to illnesses spending patterns can harm the accumulation of hu- associated with consuming these products, other man capital (especially through foregone education long-lasting economic effects have been demon- and lower health care expenditures) and, ultimately, strated. In the case of tobacco, evidence from Chi- households’ future well-being and society’s econom- na (Wang et al. 2006), India (John 2008), Cambo- ic growth. Similar effects have been found in Jamaica dia (John 2012), Zambia (Chelwa and van Walbeek in the case of SSBs (Paraje and Gomes 2021). 2014), Chile (Paraje and Araya 2017), Ghana (Ma- Of course, the costs incurred by society due to sa-ud et al. 2020), and a study of 40 LICs and MICs the consumption of tobacco, alcohol, and SSBs are (Do and Bautista 2015) shows that tobacco expendi- not completely recovered if consumption of these tures crowd out other items in household budgets, products goes to zero. Reducing premature deaths such as food, education, and medical care. In certain and chronic illnesses related to consuming these cases, a higher budget share devoted to tobacco is products generates substantial savings. Some of also associated with a higher budget share allocated these savings will be offset by an increase in spend- to alcohol. Across six diverse LMICs, tobacco and al- ing on less preventable diseases that people will ulti- cohol use were associated with crowding-out spend- mately experience. However, helping people reduce ing on acute care and preventative health-related harmful consumption postpones such spending behaviors for children, including vaccination, and and, fundamentally, ensures healthier longevity. TAXATION: ECONOMIC RATIONALES AND EFFECTIVENESS For persons concerned with public health, the social harms of tobacco, alcohol, and SSBs are an adequate justifi- cation for discouraging consumption of these products. Although diverse perspectives exist regarding the appropri- ateness of government intervention to influence individual consumption choices, there are important additional rationales that justify discouraging consumption of these products – especially by raising their prices. These ratio- nales include economic concepts of externalities, internalities, optimal taxation, and cost-effectiveness. Economic rationales for taxing tobacco, alcohol, and SSBs The consumption of tobacco, alcohol, and SSBs is ing diseases with potentially fatal outcomes, though associated with negative externalities since such by no means, the only ones (Jha et al. 2015). Alco- consumption negatively affects third parties’ well- hol consumption causes liver diseases, cancers, car- being. Externalities occur when one individual’s diovascular diseases, traumatic injuries, and others action - consumption, production - affects the well- (Gawryszewski and Monteiro 2014). In the case of being of another individual. In the case of tobacco, SSBs, diabetes, cancers, and cardiovascular diseases for instance, second-hand smoke is related to several are among the main diseases that may lead to prema- diseases (Dunbar et al. 2013; Banderali et al. 2015). ture mortality (Singh 2015). Premature deaths of pro- In the case of alcohol, road accidents, domestic and ductive adults, in whom society has previously invest- street violence, arson, and other negative externalities ed through education, health care, welfare programs, are directly linked to alcohol consumption (Chaloup- and other channels, constitute a loss of valuable col- ka et al. 2019). For all three types of harmful products lective resources with a likely effect on economic de- considered here, the diseases associated with their velopment. At the household level, illnesses caused by consumption are treated in health systems that enjoy the consumption of health-damaging products and partial or total public funding, and these interven- the out-of-pocket expenditures associated with these tions crowd out the treatment of other diseases. ailments may lead to household impoverishment or Diseases caused by the consumption of tobacco, financial hardship that can have long-lasting conse- alcohol, and SSBs often lead to premature mortality. quences and spur a vicious cycle of intergenerational In the case of tobacco, cardiovascular diseases and impoverishment (Jha 1999; U.S. National Cancer In- cancers of several organs are among the main result- stitute and World Health Organization 2016). COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 7 193 Not only are negative externalities present in Organization 2016). Taxes have been found to be the consumption of these products, but so are nega- the most cost-effective tool to curb consumption of tive “internalities,” which arise from individuals ig- tobacco and hence its externalities and internalities noring or not correctly considering harmful health (Jha 1999; International Agency for Research on effects to themselves (Allcott, Lockwood, and Taub- Cancer 2011); among the most cost-effective tools insky 2019). Internalities may be present when con- to decrease alcohol consumption; and cost-effective sumers have imperfect information on the negative to reduce consumption of SSBs (WHO 2017a). Us- effects on their own health of consuming these ing taxes does not preclude or impede using other products and/or when consumers face problems of tools to decrease consumption of these products, self-control and time-inconsistency i.e., underesti- such as marketing restrictions, prohibitions on sell- mating their own future health care costs, relative to ing to certain groups (minors, for instance), label- their future preferences regarding the significance ing of products with health warnings, and others. of such costs (Allcott, Lockwood, and Taubinsky A further economic rationale for taxing tobac- 2019). Educational campaigns, warning labels, re- co, alcohol, and SSBs is unrelated to their health ef- strictions on advertising and sponsorship, and sim- fects. Rather, it focuses on identifying the best way ilar measures correct for internalities, but they are to raise public revenues. Adam Smith discussed this not 100 percent effective. Thus, taxes still have a role point explicitly in The Wealth of Nations, when he to play (Allcott, Lockwood, and Taubinsky 2019). identified tobacco, alcohol, and sugar as products The usual tool used to correct negative external- which were better sources of government revenues ities are the so-called Pigouvian taxes that increase than other goods or services. While this approach marginal (private) costs to the consumer by enough has received relatively little attention, one study ex- to offset the cost of present and future externalities. plicitly explored the idea in relation to alcohol taxes These externalities can be very large and often take (Parry et al. 2009). It found that a revenue-neutral a heavy toll on households where the consumption policy of raising alcohol taxes to replace labor tax- of tobacco, alcohol, and SSBs occur. They include, es would lead to substantial welfare gains. Adopt- for example, lost household income and catastroph- ing this revenue strategy would imply much higher ic health expenditure for the household (Jha 1999; rates of taxes on alcohol than those based purely on U.S. National Cancer Institute and World Health addressing externalities. EFFECTIVENESS OF PRICE POLICIES As noted earlier, affordability – the relationship between consumer income and product price – is the key vari- able affecting consumption of health-damaging products. Per capita incomes and product prices determine af- fordability – the first positively and the second negatively. Per capita incomes are obviously not a relevant policy instrument, but altering product prices can be. Taxation and/or minimum-price policies typically increase prices and, by doing so, decrease affordability and reduce consumption. Studies have consistently shown that taxes on to- the price of an alternative product increases by 1 per- bacco, alcohol, and SSBs do indeed raise prices and cent, thereby indicating whether these are substitute reduce consumption. This effectiveness is a conse- or complementary products; and income elasticities quence of how demand responds to increases in a measure the percentage by which demand increases product’s price (own-price elasticity); to the avail- or decreases when income rises by 1 percent. ability of substitutes (cross-price elasticities); and to changes in income (income elasticity). Own-price Own-price elasticities elasticity measures the percentage by which con- sumption decreases when the product’s price increas- There is a vast literature estimating the effect of to- es by 1 percent. High own-price elasticities imply that bacco prices on demand for entire populations and a small increase in price (e.g., due to a tax increase) subgroups (e.g., youth, individuals with different produces a relatively large reduction in consump- income levels, genders). Less extensive, but still tion. Cross-price elasticities measure the percentage large, is a literature relating tobacco prices to smok- change in consumption of a particular product when ing onset and cessation. 194 CHAPTER 7 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E Studies have measured the own-price elastici- decrease consumption) but lower (in absolute val- ties of tobacco and have centered around values of ue) than for the entire population (-0.3 vs -0.5, re- -0.4 for HICs and around -0.5 for LICs and MICs, spectively (Wagenaar, Salois and Komro 2009). Even though the difference may not be statistically sig- binge drinkers are price-responsive (Burton et al. nificant (US National Cancer Institute and World 2017) but tend to choose cheaper drinks in an effort Health Organization 2016; WHO 2011; Chaloupka to keep up their alcohol consumption. This implies and Powell 2019; Guidon et al. 2015). Overall, short- that policies aimed at increasing the price of cheap- run elasticities are about half the absolute value of er drinks (e.g., minimum unit price policies) can be long-run elasticities (US National Cancer Institute effective at reducing binge and/or heavy drinking and World Health Organization 2016; Becker et al (Jackson et al. 2010; Booth et al. 2008; van Walbeek 1994). About half of the own-price elasticity is due and Chelwa 2021). Finally, evidence on how prices to a decrease in prevalence (i.e., due to people quit- affect alcohol initiation is scarce but suggests that ting), while the other half is due to a reduction in the higher prices delay and, to some extent, prevent ini- number of cigarettes smoked by those who continue tiation. This can have long-lasting effects on future smoking (U.S. National Cancer Institute and World drinking patterns; for example, individuals who ini- Health Organization 2016; Goodchild et al. 2016). tiate at older ages have a lower probability of having Youth are substantially more price-responsive than frequent heavy-drinking episodes (Sornpaisarn et adults in countries at all income levels (U.S. Nation- al. 2015; Paraje, Guindon and Chaloupka 2020). al Cancer Institute and World Health Organization For SSBs, there is mounting evidence on rela- 2016; Nikaj and Chaloupka 2013; Kostova et al. tively high (in absolute terms) and significantly neg- 2011), and young men are more price-responsive ative own-price elasticities. The own-price elasticity than young women (Chaloupka and Pacula 1999). for SSBs is around -1, i.e., a 1 percent increase in SSB In addition, higher prices delayed or even prevent- prices decreases consumption by 1 percent (Teng et ed the initiation of smoking in a variety of coun- al. 2019). In some countries, higher price-respon- tries i.e., the own-price elasticity of smoking onset siveness has been measured, for example in the Unit- was negative and statistically significant (Guindon, ed States (with an own-price elasticity of -1.2) (Pow- Paraje and Chávez 2017; Guindon, Paraje and Cha- ell et al. 2013); Mexico (-1.1 to -1.2) – (Powell et al. loupka 2019; Kostova, Husain and Chaloupka 2017; 2013); Chile (-1.3 to -1.4) – (Caro et al. 2018; Guerre- Vellios and van Walbeek 2016; Shang et al. 2019). ro-López, Unar-Munguía and Colchero 2017); Ecua- This is in large part because young people general- dor (-1.2) – (Paraje 2016); Guatemala (-1.4) – (Cha- ly have lower discretionary incomes than they will on et al. 2018); and South Africa (-1.2) – (Stacey et al. when they are older. Finally, higher tobacco prices 2017). A systematic review of studies in 164 countries increase the likelihood of cessation, as demonstrat- found that price-responsiveness was higher in the ed in both HICs and MICs (Gonzalez-Rozada and lowest-income countries and for the youngest and Montamat 2019; Forster and Jones 2001; DeCicca, oldest adults vis-à-vis middle-aged persons, proba- Kenkel and Mathios 2008; Hyland et al. 2006; Shang, bly reflecting life-cycle changes in incomes. The re- Chaloupka and Kostova 2013; Ross et al. 2013). view found no differences in price-responsiveness The evidence on own-price elasticity for alco- between men and women (Muhammad et al. 2019). hol is also compelling, with values of -0.5 for short- Overall, there is vast and consistent evidence run elasticity and around -0.8 for long-run elasticity that people consume less tobacco, alcohol, and SSBs (Gallet 2007). However, not all types of alcoholic when faced with increases in prices. For tobacco and beverages show the same own-price elasticity. In alcohol, raising prices decreases consumption less general, beer demand is less price-responsive than than proportionally, while in the case of SSBs the wine and spirits (-0.3 vs. -0.6, respectively) – (Powell reductions in consumption tend to be proportional and Chaloupka 2019; Guindon et al. 2022). There is or more than proportional to the price change. little evidence on gender differences in price-respon- siveness (Gallet 2007). Studies show different results Cross-price and income elasticities on own-price elasticities by age, finding no conclu- sive differences between young people and adults A rise in other products’ prices can lead to increas- (Jackson et al. 2010; Stockings et al. 2016). In the es in consumption if products are substitutes, or to United States, alcohol price elasticity among heavy consumption decreases if they are complements. drinkers is significantly negative (i.e., price increases The sign on the cross-price elasticity parameter in- COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 7 195 dicates whether they are substitutes (positive sign) alcoholic beverages, and SSBs are normal goods: or complements (negative sign). that is, their income elasticity is positive, meaning In the case of tobacco, for instance, there is evi- that an increase in income increases demand for the dence of substitution among tobacco products, such product. For tobacco, income elasticity is estimat- as manufactured cigarettes, roll-your-own tobacco, ed at between 0.3 and 0.4, such that a 10 percent bidis, and small cigars, as well as between brands increase in income would lead to a 3 to 4 percent within the same product e.g., cigarettes (Nargis et increase in tobacco demand (Gallet and List 2003; al. 2019; Chaloupka and Powell 2018; Jawad et al. Nargis et al. 2021). In the case of alcohol, a large sys- 2018; Guindon et al. 2011). In addition, there is evi- tematic review found estimated income elasticities dence of electronic cigarettes being substitutes with to be in the vicinity of 0.7 (Gallet 2007). Finally, for manufactured cigarettes (Stoklosa, Drope and Cha- SSBs there is evidence that income elasticities are loupka 2013). Less clear is the relationship between between 0.8 and 1, though the number of studies tobacco and alcohol (Pacula 1998), as some stud- estimating this parameter is still low (Paraje et al. ies have found that they are substitutes (Dee 1999; 2016; Chacon et al. 2018; Stacey et al. 2017). Results Chen et al. 2016), others have found that they are to date on income elasticities for these products— complements, and still other studies have reported suggesting that they are normal goods—indicate mixed results (Markowtiz and Tauras 2009). that increases in affordability lead to higher demand In the case of alcohol, there is no consistent ev- for these products irrespective of what occurs with idence on substitution between different alcoholic own or other products’ prices. This corroborates beverages (possibly due to methodological prob- our earlier discussion of affordability. lems, sample sizes, and other factors) (Ornstein 1980; Edwards 1994), while there is strong evidence Pass-through to prices on substitution within beverage groups e.g., beers (Rojas and Peterson 2008; Gruenewald et al. 2006; Given the strong evidence that demand for tobac- Meng et al. 2014; Srivastava et al. 2015; Ruhm et al. co, alcohol, and SSBs is price-responsive, taxes can 2012). Finally, there is only suggestive evidence on decrease consumption only if they can increase cross-price elasticities between alcohol and illicit prices. Though it is common to assume that a tax drugs (Burton et al. 2017). increase automatically raises prices, in fact, the de- SSBs clearly have substitutes among other bever- gree and speed with which prices rise will depend, ages, such as milk, coffee, tea, bottled water, and diet among other things, on the market structure, the carbonated beverages, with higher degrees of substi- structure of the excise tax system, and the availabil- tution for milk and bottled water (Guerrero-López, ity of non-taxed or differentially taxed substitutes Unar-Munguía and Colchero 2017; Paraje 2016; (which affects the own-price demand elasticity). In Smith, Lin and Lee 2010; Finkelstein et al. 2013). particular, prices will typically rise more in response Some studies have estimated SSB price-responsive- to a tax increase when the market is concentrated, ness in food demand systems, including not only the taxes are charged at the retail level, and when a beverages but solid sugar-rich foods, such as cook- product has few or no close substitutes (and, hence, ies, candies, deserts, and snacks. Findings from these a low own-price demand elasticity). Taxes can be investigations have been mixed. Some products are fully passed to prices (pass-through equal to 1), can substitutes, others are complements, and others are be under-shifted (pass-through less than 1), or can not related (Colchero et al. 2015; Caro et al. 2018). be over-shifted (pass-through greater than 1). Such a pattern would imply that higher SSB prices There is evidence, for instance, that in less com- could lead to an increase in consumption of other un- petitive markets, firms tend to fully pass taxes to pric- healthy food products (salty snacks, candies, cookies, es or even over-shift them. For example, in the Unit- and others), partially offsetting the expected reduc- ed States, the 1983 increase in the Federal Cigarette tion in sugar intake. While this may happen, it has Excise Tax was over-shifted by oligopolistic firms been found that the net effect of SSB price increases exploiting their market power (Harris 1987). Sim- is still an overall decrease in sugar intake (Finkelstein ilar behavior was found in Europe (Delipalla  and et al. 2013; Pell et al. 2021; Leider et al. 2021). O’Donnell. 2001). In the European case, ad-valorem There are considerably fewer studies on income taxes (i.e., a percentage of the price) tend to produce elasticities. Still, the evidence is consistent in show- lower pass-through rates for tobacco when com- ing that the main tobacco products (e.g., cigarettes), pared to specific taxes (i.e., a tax per unit sold). Tax 196 CHAPTER 7 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E structure can also affect pass-through rates. In India and for some categories of SSBs in the UK e.g., those (Jha et al. 2011) and Bangladesh (Nargis et al. 2020), in the top tier of the tax (Scaborough et al. 2020). In a complex tiered tax system has given the tobacco in- the United States, studies of local taxes have found dustry room for passing taxes differentially accord- mixed results, depending on the location of the store ing to price segments (i.e., taxes are under-shifted – e,g., proximity to state borders led to under-shifting for cheaper cigarettes and over-shifted for premium (Cawley et al. 2017), store types, and beverage types cigarettes, thereby leading to downward substitution (Silver et al. 2017; Powell et al. 2020). A recent me- rather than reducing consumption). The same effect ta-analysis found that, on average, the pass-through was also found in Mauritius (Berthet Valdois et al. was 70 percent of the tax, though with considerable 2019). The perspective of further regulations may variation across studies (Powell et al. 2021). also affect the pass-through rate, as producers facing Evidence on price pass-throughs implies that a scenario of tightening restrictions in consumption, large increases in specific taxes are a more effective higher taxation, and other constraints may choose strategy for reducing consumption than a series of to maximize short-run profits by over-shifting taxes smaller tax increases, particularly if they are ad-va- (Becker et al. 1994). There is evidence that the tobac- lorem taxes. A large specific excise tax is less easily co industry uses a number of strategies (e.g., stock- manipulated by industry and, if applied uniformly to piling, that is, oversupplying the domestic market all products, it can narrow the differential between before a tax increase) to delay the change in prices cheaper and more expensive brands. Incremental tax after a tax increase, although prices eventually rise as increases give the industry political room to blame a result of such an increase (Ross, Tesche and Vellios the government for price increases they impose in 2017; Marquez and Moreno-Dobson 2017). order to maximize profits. A larger excise tax increase For alcohol, taxes also seem to be fully passed or makes this “blame and pocket” strategy less feasible, even over-shifted. A review found that taxes are gen- as the public is more likely to be alerted to the larger erally over-shifted in the case of beer and fully shifted expected tax increase (Marquez and Moreno-Dob- for wine and spirits (Nelson and Moran 2020). In the son 2017). More importantly, behavioral research United States, state and federal alcohol taxes appear also find that people are more responsive to large to be over-shifted, especially for beer and spirits. The salient changes in prices than to incremental ones price adjustment also seems to be quite rapid, within because it changes the framing of their consumption three months of the tax change (Young and Bielińs- decision and triggers a greater loss aversion response ka-Kwapisz 2002) Similar results were found in the (Tversky and Kahneman 1981; Babor et al. 1978; Ep- case of beer in Alaska (Kenkel 2005). For the UK, the stein et al. 2012; Kalyanaram and Winer 2022). pass-through rate varies by price level of products. Producers of relatively cheaper alcohol beverages Minimum unit prices: tend to under-shift taxes, while those of relative- cases, evidence, and implications ly more expensive beverages over-shift them (Ally et al. 2014). A similar finding was reported for the Minimum unit prices (MUP) are policies that raise pass-through of alcohol taxes for 27 OECD coun- the “floor” of prices without using taxes. MUPs tries (Shang, Ngo, and Chaloupka 2020). This study raise average prices while reducing the possibility also reported under-shifting or fully shifting taxes in of down-trading (i.e., switching to cheaper brands). the case of wine and certain spirits (e.g., cognac) and This measure has been used in the context of tobac- over-shifting in the case of beer and Scotch whisky. co and alcohol control. Although it is effective when In the case of SSBs, the evidence is mixed. In properly used, it has the disadvantage of not gener- Mexico, the SSB tax was mostly over-shifted for so- ating direct revenues (only indirect ones, via VAT das and fully or mildly under-shifted for other SSBs for instance) and being more costly to enforce than, (Grogger 2016; Colchero et al. 2015). In France, taxes for instance, taxes at the producer level. In the case of were over-shifted for sodas and under-shifted for fruit tobacco, MUP policies might set a minimum price juices and flavored waters (Berardi et al. 2016). In for 20-cigarette packages, while MUPs for alcohol Denmark, three tax changes between 1998 and 2003 have been based on alcohol content (e.g., Scotland’s were either fully shifted or over-shifted (Bergman MUP is £0.50 per 8 grams of pure alcohol) or based and Hansen 2019). However, in other cases, taxes on volume (e.g., £0.50 per liter of beer). have been under-shifted. This is the case of Barbados Some research suggests that, at least in coun- (Alvarado et al. 2017), Chile (Nakamur et al. 2018), tries with the institutional capacity to ensure en- COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 7 197 forcement, MUPs for alcohol may be even more ef- from 2007 to 2014 found that cigarette prices were fective than taxes for reducing consumption among significantly higher in states with MUP policies certain groups. A study in Australia found that an than those without, especially for cheaper products MUP policy was more effective at reducing con- (Huang et al. 2016). Models predict that establish- sumption among heavy drinkers of beer and wine ing a national MUP (at US$ 10) for cigarettes could than a volumetric excise tax i.e., a tax based on alco- significantly reduce sales of cigarettes and increase hol content (Shara et al. 2014). A systematic review cessation (Doogan et al. 2018). found that modelling studies for England and Aus- As MUP do not raise revenues (apart from tralia confirm that an MUP policy is highly effective those that can be generated from VAT) the political at reducing consumption among heavy drinkers, economy to implement them can be more complex. though it would minimally affect moderate drinkers Ministries of Finance, for instance, may not have (Burton et al. 2017). The same effect was found in the incentives to push for MUP as they will not get South Africa (van Walbeek and Chelwa 2021). any substantial extra revenues. High enforcement MUP policies are widely used for tobacco in costs, especially important in poorer countries, can the United States; at least 25 states have implement- also be a deterrent to implement MUPs, as such ed them (Feighery et al. 2005). A study using data costs may not be recovered with extra revenues. RELATED TOPICS Distribution of financial burden and averted health costs: extended cost-benefit analyses The extended cost-benefit analysis (ECBA) methodology incorporates any short-term welfare losses from excise taxes into a framework that includes the medium- and long-term health benefits for those who quit consuming or who consume fewer harmful products. Among other aspects, ECBA accounts for differential behavioral re- sponses across population groups—including different income groups—through the estimation of group-specific price elasticities. This kind of analysis tends to show that, because lower-income individuals tend to be more sensitive to cigarette, alcohol, and SSB pricing, they will benefit disproportionately from the resulting health im- provements, health care expenditure reductions, and higher disposable income for purchasing non-toxic goods and services. Fiscal incidence analyses that do not consider behavioral responses and account only for the direct costs of taxing these products are partial or incomplete because they ignore these other effects. By accounting for both direct and indirect effects in the short and medium terms, ECBA provides a more complete picture of the costs and benefits that taxing tobacco, alcohol, and SSBs entails for household welfare. This contradicts industry claims that such taxes are hardest on the poor. Tobacco, alcohol, and SSBs are unlike other taxable reached similar conclusions (Cobacho et al. 2010). goods. Preventing health and economic harms asso- The ECBA methodology has been applied ciated with consumption of these products generates widely, including in a 13-country study (Global large benefits to current and potential consumers, Tobacco Economics Consortium 2018). At least 13 governments, and societies. In general, most smokers country studies using ECBA highlight that the me- are willing to quit but unable to do so. In the United dium- and long-term benefits of reducing smoking States, for instance, about 68 percent of smokers were can outweigh the short-term spending on taxes, willing to quit, and 55 percent attempted to quit, but resulting in net gains, particularly among poorer only about 7.5 percent of smokers successfully quit households (Del Carmen, Fuchs and Genoni 2018; smoking (U.S. Department of Health and Human Fuchs, Matytsin, and Obukhova 2018; Fuchs and Services 2020). A US study that compares the value Meneses 2018, 2017a, 2017b; Fuchs and Del Car- of statistical life for smokers with the costs of smok- men 2018; Fuchs and Del Carmen and Mukong ing found that “the economic value of the premature 2018; Fuchs and Gonzalez 2019; Fuchs, Orlic, and mortality due to smoking dwarfs the purchase price Cancho 2019; Fuchs, Mandeville, and Alonso-Soria of cigarettes,” reaching a private cost of US$ 222 per 2020). Similar findings are reached by Fuchs, Paz, pack for men and US$ 94 per pack for women in 2006 and Gonzalez Icaza who simulate tax policy chang- dollars (Viscusi and Hersch 2008). A study in Spain es in eight low-, middle- and high-income countries 198 CHAPTER 7 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E (Fuchs, Paz, and Gonzalez 2019). The distribution avoided health care costs, and productivity gains, of elasticities and resultant health and economic these partial studies are biased regarding the distri- benefits from reduced medical expenses and lower butional consequences of taxing harmful products. ‘years of working-life lost’ (YWLL) generally more Complete studies on the distributional conse- than offset the short-term negative effects of tobac- quences of taxing alcohol and SSBs are less com- co taxes on household budgets. For sufficiently high mon but growing in number. For example, a study price shocks and elasticity scenarios, the tax inci- of excise taxes on SSBs in Ukraine found that the dence is progressive. Half the population in these net effect is progressive though small in magnitude: countries could benefit from net positive income raising SSB prices by 20 percent increases disposable gains in the medium to long term if cigarette prices income of the poorest quintile of the population by rose by 50 percent. In a similar study, Postolovska 0.03 percent (Fuchs and Gonzalez 2021). In Ka- et al. find that increasing Armenia’s cigarette excise zakhstan, where the average price elasticity of SSBs tax rate to 75 percent of the retail price could deliv- is estimated to be -0.70, lower income deciles bene- er large health and financial benefits to households, fit more than higher income deciles from the simu- with pro-poor impacts (Postolovska et al. 2017). lated introduction of a 20 percent price increase on The reduction in medical bills tends to be the SSBs (Fuchs, Mandeville, and Alonso-Soria 2020). most significant component driving the net ben- Finally, a recent ECBA in Brazil shows that price in- efits under the ECBA. All income groups benefit creases on alcohol, tobacco, and SSBs have positive from the reduction in medical expenses when taxes and progressive effects, when incorporating the im- discourage smoking, but these benefits are dispro- pact of health taxes on prices, medical expenditures, portionately larger for poor households. In Moldo- and productive lives (Coelho 2020). va, where tobacco-related diseases are the leading No ECBAs are currently available specifically cause of premature adult deaths, just accounting for for alcohol taxation. However, a recent report ana- reductions in medical expenses is enough to offset lyzed the distributional effect of alcohol taxes in the the initial negative price effect on household ex- UK (Bhattacharya 2020). It found that there is no penditures, with a clear progressive pattern (Fuchs evidence to support the idea that alcohol taxes are and Meneses 2018). In Chile, Ukraine, and Russia, regressive. Furthermore, once the use of alcohol tax reducing medical expenses constitutes the largest revenues by the Government is considered, results long-term benefit of the tobacco price increase un- can be strongly progressive, if revenues are used to der the ECBA (Fuchs and Matytsin 2018; Fuchs and finance increased health spending or other pro-poor Meneses 2017a, 2017b). Similarly, tobacco price in- programs. Note that “soft” earmarking (i.e., in which creases yield positive welfare gains from being able the recommended use of the new revenues is not le- to work longer. In Bangladesh, for instance, the gally required) has been found to be helpful for over- main gains of taxing tobacco under the ECBA mod- coming political opposition to higher tobacco excise el arise from extending people’s working lives i.e., taxes (Marquez and Moreno-Dobson 2017). lowering YWLLs (Del Carmen and Fuchs 2018). In the case of alcohol, the relative financial bur- Despite these ECBA results, those lobbying den may be affected not only by income but also, against tax increases continue to cite studies which crucially, by the intensity of drinking. In Australia, look exclusively at the direct short-term distri- alcohol taxes represent a higher burden to heavy butional costs. One such study in Chile showed drinkers irrespective of their incomes and, because that, if cigarette prices increased by 25 percent, the of that, one study suggests that MUP policies or tax- poorest 20 percent of the population would reduce es that increase the cost of the cheapest alcohol can consumption of other goods by 0.35 percent in the be more effective in reducing alcohol consumption, absence of any adjustments in cigarette consump- without having highly regressive effects (Vanden- tion, while the wealthiest 10 percent would face a berg and Sharma 2016). However, this is a partial smaller reduction of 0.09 percent (Fuchs and Mene- study which does not consider the distributional ef- ses 2017). Without considering the health benefits, fects on revenues and health care cost savings. COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 7 199 COMBINATION OF TAXES AND NON-PRICE INTERVENTIONS Price interventions through taxation (and/or MUPs) are not the only policies for reducing consumption of to- bacco, alcohol, and SSBs, although they are generally among the most cost-effective. Other policies that are effec- tive at reducing consumption include: (1) marketing controls (e.g., on advertising, package labelling, promotion, or sponsorship); (2) sales regulations (e.g., licensing, store opening times, age restrictions); (3) consumption regulations (e.g., restrictions on smoking or drinking in public places, work spaces, or other specified locations); (4) consumer education (e.g. mass media campaign, package labelling); and (5) cessation programs and inter- ventions (especially for tobacco, but also for alcohol). WHO has reviewed the evidence on the cost-effec- advertising restrict social cues that encourage smok- tiveness of some of these policies with regard to ing and reduce the tobacco industry’s social and po- reducing tobacco and alcohol consumption (WHO litical influence, further enabling tax increases. Bans 2017a). WHO’s study defines “best-buys” as mea- on smoking in public places, particularly workplaces, sures that cost at most US$ PPP 100 per DALY require smokers to take time from work and to go averted in low- and middle-income countries. outside even in unpleasant weather, in effect raising For tobacco, five “best-buys” were identified. the “cost” of smoking. It has been widely shown that Tobacco taxation leading to price increases was the bans on smoking in public places almost immedi- most cost-effective. Two marketing control measures ately affect health, reducing the incidence of strokes, were jointly ranked second: (a) the implementation myocardial infarctions, and certain respiratory dis- of plain/standardized packaging and/or large graphic eases (Bartecchi et al. 2006; Juster et al. 2007; Naiman health warnings on all tobacco packages; and (b) en- et al. 2010; Sargent et al. 2004; Ferrante et al. 2012). acting and enforcing comprehensive bans on tobac- For alcohol, the WHO “best buys” are: (1) in- co advertising, promotion, and sponsorship. Finally, crease excise taxes on alcoholic beverages; (2) en- cost-effectiveness was demonstrated for two addi- act and enforce bans on exposure to alcohol ad- tional measures: effective mass media campaigns that vertising; and (3) enact and enforce restrictions on educate the public about the harms of smoking/to- the physical availability of retailed alcohol such as bacco use; and eliminating exposure to second-hand through reduced hours of sales (WHO 2017a). smoke in indoor spaces and public transportation. Finally, WHO assessed the cost-effectiveness of Although tobacco taxation was the most cost-ef- measures to reduce unhealthy diets in general, not fective intervention, the measures noted above are interventions focused specifically on SSBs. How- complementary. Public awareness of the hazards ever, WHO’s analysis did cite evidence that taxes of smoking and the benefits of quitting, achieved on SSBs are cost-effective, even though they did through mass media campaigns, graphic warnings not meet the threshold of US$ PPP 100 per DALY on packs, and the other measures cited all encourage averted. For this reason, WHO has actively promot- wider public support for tax hikes, in addition to dis- ed and encouraged the use of SSB taxes as part of couraging consumption (Jha and Peto 2014). Simi- a policy package to tackle the overweight/obesity larly, bans on smoking in public places and bans on pandemic (WHO 2017b; PAHO 2021; WHO 2022). CHALLENGES FROM THE INDUSTRY The tobacco industry has a long and well-researched history of concealing evidence on the toxicity of its products, deceiving the public on the harmful effects of tobacco consumption, and interfering in public policy. As early as the 1950s, tobacco companies concealed evidence on the harmfulness of their products and the addictive properties of nicotine. In the 1970s and 1980s, they denied links between smoking and cancers and the harmfulness of sec- ond-hand smoking (WHO 2019a; Brownell and Warner 2009). More recently, despite robust scientific evidence, they misled the public regarding the relative safety of “light” or “low-tar” cigarettes (WHO 2019a). The tactics of the tobacco industry have been ex- crepancies between what the industry knew and the posed through successful efforts to make the indus- ideas it promoted. For example, the industry has con- try’s own documents public (Tobacco Control 2002). sistently portrayed smokers as well-informed, ratio- These documents have demonstrated the large dis- nal individuals who choose to smoke (Brownell and 200 CHAPTER 7 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E Warner 2009; Kyriakoudes 2016; Risi and Proctor encourage public opposition to proposed taxes; 2020), even when tobacco firms know that their own (d) commissioning research that is biased toward marketing is effective precisely because individuals supporting industry campaigns; (e) deliberately are not well-informed and unfailingly rational. confusing debates about tobacco tax increases with Thus, to market cigarettes, firms use strategies broader tax debates; (f) undermining the credibility that appeal to individuals’ non-rational behaviors. of tobacco control experts; and (g) stimulating and These approaches include branding, associating cig- even participating in smuggling (see below) in the arettes with desirable images, and increasing nico- event of tax increases (Smith et al. 2013). tine content to make smoking even more addictive. The tactics of the alcohol and SSB (or, more ac- Such marketing strategies are used while intentional- curately, food) industries to oppose taxes and regu- ly repressing extensive evidence regarding smokers’ lations are remarkably similar to those used by the ignorance and reluctance to internalize the health tobacco industry. Alcohol, SSB, and food compa- risks from smoking (due to time inconsistency) and nies also shift blame for unhealthy eating or drink- the addictive nature of nicotine (which undermines ing away from themselves as purveyors of products rational choice). Also notable are the industry’s ef- and onto the individuals who are intentionally in- forts to market to youths, whose brain development fluenced by their marketing campaigns (Brownell is known to be incomplete, especially with regard to and Warner 2009). The alcohol and processed-food evaluating long-term consequences and risk (UN industries have also linked regulations and taxes to National Cancer Institute and World Health Orga- the loss of personal freedoms and the notion that nization 2016; Villanti et al. 2019). The one interven- public health policies create “food police” or “food tion major actors in the tobacco industry are will- fascism.” (Brownell and Warner 2009; Larrain 2016). ing to promote is public education campaigns, the Self-regulation is often proposed as an effective and least cost-effective intervention in the WHO study more efficient alternative to state regulation, even reported above (WHO 2017a). Public education is when this has been shown to be ineffective. Like the indeed important, but only as a complement to tax- tobacco industry, the producers of alcohol and SSBs ation and other policies, not as a substitute for them. lobby governments and the public by arguing that Apart from concealing and distorting evidence, taxes do not reduce consumption; that they are re- the tobacco industry has used several arguments to gressive; and that they are “discriminatory” as they deter, impede, or delay increases in tobacco taxes. are levied on a specific group of products, or even A systematic review of the tobacco industry’s to- unconstitutional (Hattersley et al. 2020). bacco tax opposition found that the most common Alcohol and SSB manufacturers promote two and consistent arguments were that tobacco taxes additional key ideas which are also common in to- are: (a) regressive; (b) do not affect consumption; bacco industry strategies. They stress that: (a) the (c) lead to more illicit trade and foster organized etiology of the problem is complex, and therefore crime; and (d) reduce employment and harm busi- no individual product can be blamed; and (b) pop- nesses (Smith et al. 2013). The empirical evidence ulation health measures (such as taxes) are too sim- on regressivity and impact on consumption have ple to address complex public health problems (Pet- been reviewed in preceding sections. Results show ticrew et al. 2017). Often, the solutions proposed by that the tobacco industry’s arguments are unfound- these industries are simplistic, disregarding their ed with regard to these two issues. Evidence also own diagnosis of complexity. Their recommended refutes industry arguments about illicit trade and approaches largely focus on providing more infor- employment, as the next subsections will show. mation, consumer education, and relying on indi- In addition to disseminating these false argu- vidual responsibility (Petticrew et al. 2017). The dis- ments, the tobacco industry has developed a range missal of independent scientific evidence contrary of tactics to oppose, delay, or water down public to their interests as “junk science” is also common. health policies aimed at reducing smoking. The In the case of alcohol, there is less evidence on abovementioned review identifies 21 tactics, in- the strategies used to influence alcohol taxes. Never- cluding: (a) using front groups and friendly “inde- theless, some authors have documented how the in- pendent” experts to give their arguments credibil- dustry uses national and international front groups ity; (b) directly lobbying policy makers; (c) using (for example, the International Centre for Alcohol media resources (including op-ed pieces and letters Policy, ICAP, or the International Alliance for Re- to newspapers) and other publicity campaigns to sponsible Drinking, IARD) to influence public poli- COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 7 201 cy (Collin and Hill 2019). These front groups engage est 40 countries, though the former had an initially in activities to distort evidence on taxation, claiming higher total tax burden (66.5 percent versus 37.6 taxes are a “blunt tool” that do not reduce consump- percent in 2008) (WHO 2019c). Out of 179 coun- tion, are regressive, increase illicit trade, and reduce tries, 91 increased the tax burden of specific taxes, government revenues. They also use strategies to 58 raising this figure by more than 10 percentage meddle in political processes, by presenting alterna- points and 33 by more than 20 percentage points. A tive policies (not related to taxes) and lobbying poli- recent exercise in scoring changes in tobacco taxes cy makers and officials (Collin and Hill 2019). In the found that, out of 170 countries, 51 increased to- case of the SSB industry, the strategies and actions bacco taxes between 2014 and 2018 (Chaloupka et are similar, as the discussion of Mexico’s SSB taxes al. 2021b). The global situation of rising taxes and has demonstrated (Pedroza-Tobias 2021). decreasing consumption does not fit the tobacco in- dustry’s narrative nor the evidence on illicit trade. Taxes and illicit trade Second, the evidence cited by tobacco compa- nies to link tobacco tax increases with illicit trade is Once the tobacco industry lost its ability to deny the weak and generally based on studies financed by the addictiveness and harmfulness of smoking, it sought tobacco industry itself. These studies rarely make other ways to confuse public policy debates, delay their methods and data publicly available for peer action, and stop or limit tax increases. One of the review and scrutiny. By contrast, independent stud- most common arguments the tobacco industry uses ies have used cross-country evidence to show that to challenge tobacco taxes is to claim that these pol- countries with higher taxes have a lower penetration icies will be ineffective and counterproductive be- of illicit trade than those with lower taxes (US Na- cause they will encourage illicit trade, an argument tional Cancer Institute and World Health Organiza- also used regularly by the alcohol industry. While tion 2016; Joossens et al. 2010). The recent cases of the argument is plausible, it is empirically false. the UK, the Philippines, and Botswana provide an Indeed, investigations have found that the transna- example (Blecher 2020; Diosana and Sta. Ana 2020; tional tobacco industry is itself responsible for the World Bank 2019). It is likely that illicit trade is production of about two-thirds of illicitly traded driven less by price increases on tobacco than by the cigarettes (Gilmore et al. 2017). Quite often, tobac- general capacity of tax administration authorities to co companies are directly involved in smuggling, as enforce taxation – and this is true not only for tobac- shown by formal investigations and settlements, first co (US National Cancer Institute and World Health with Canada (CBC News 2008) and later with the Organization 2016). Lack of controls, corruption, European Union, among others (NBC News 2004). and weak administrative capacities may foster illicit Evidence against the tobacco industry’s claims trade, though countries with middling administra- on illicit trade is strong. First, illicit trade has not in- tive capacities, such as the Philippines and Botswana creased in the last decade, despite increasing tobacco (among others), have succeeded in restraining illicit taxes (Paraje and Stoklosa 2021; Joossens and Raw trade. Strengthening excise tax laws (e.g., through 2012; Goodchild 2020). In fact, illicit tobacco prod- harsher penalties, strong governance and control ucts, mostly cigarettes, constitute a stable share in a processes, and other measures) helps to curb illicit shrinking market. As mentioned above, affordability tobacco trade (as illicit trade in any other product). of cigarettes has decreased (as measured by the RIP) Furthermore, small-scale smuggling cannot in 117 out of 168 countries between 2008 and 2018 have a significant effect on sales or smoking. Wide- (Paraje and Stoklosa 2021), and this decrease came spread smuggling requires active or tacit support mostly from higher tobacco taxes. Based on WHO from the tobacco industry. A comparison of France data, the unweighted average of the total tax share and Canada is informative. Both countries mark- (as proportion of final retail price) for the most-sold edly raised taxes, tripling them from the 1990s to brand of cigarettes went from 46.6 percent in 2008 2004. During these years, however, the tobacco in- to 52.4 percent in 2018 (WHO 2019c). Out of 174 dustry in Canada led a well-orchestrated scheme countries with complete data for this period, 118 in- to export tax-free Canadian brand cigarettes into creased the total tax share, with 48 countries raising the United States and then smuggle them back into it by more than 10 percentage points and 22 by more Canada. The industry then used the evidence of than 20 percentage points. There is no significant dif- smuggling (without admitting its own role) to lob- ference on this between the richest 40 and the poor- by the Canadian government to lower taxes and to 202 CHAPTER 7 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E argue against higher taxes in other countries. The er spending on other products will increase and excess consumption that resulted from the reduced raise sales and employment in those other sectors excises spurred by this scheme, as well as directly (Chaloupka and Powell 2018). Furthermore, when from the smuggled (and cheaper) cigarettes, was ap- governments spend the excise tax revenues, they proximately 30-40 billion sticks in the early 1990s. also generate employment. Studies have found that Hence, the failed Canadian response to smuggling shifting demand from the tobacco industry, which would eventually cause some 30,000 to 40,000 to- is relatively capital-intensive, to industries which bacco-attributed deaths due to the tobacco indus- are more labor-intensive can actually increase em- try’s orchestration of smuggling during the 1990s ployment (Marquez and Moreno-Dobson 2017). (CBC News 2008; Jha et al. 2020). A relatively large body of evidence exists on this Third, profit-maximizing smugglers increase for SSB taxes (Hattersley et al. 2020). Though the SSB illicit cigarette prices along with the increases in industry fought against a tax increase in Mexico in licit cigarette prices (Joossens and Raw 2012). This 2014 by claiming that it would reduce employment, implies that, even when increased taxes divert some subsequent evidence showed that the policy did demand to the illicit market, they also push prices not have any impact in terms of employment in the up in the illicit market, discouraging consumption manufacturing sectors affected by the tax (Guerre- (Goodchild et al. 2020). Recent evidence shows that ro-López et al. 2017). Employment in the retail sec- illicit cigarette prices generally follow the prices of tor was also unaffected (it even showed a moderate legal cigarettes (correlation coefficient: 0.87) (Paraje increase). The same is shown for several cities in the and Stoklosa 2021; Drope et al. 2021). United States that have implemented such a tax. In In the case of alcohol, there is less evidence on the case of San Francisco, a recent study shows that, the relationship between taxation and illicit, unre- two years after the implementation of the SSB tax, corded alcohol consumption. Alcohol is less suscep- there was no discernible effect on employment for tible to smuggling than tobacco, because it is heavier the overall economy, private sector, supermarkets and more difficult to transport relative to its value. and other grocery stores, convenience stores, lim- On the other hand, opportunities for artisanal, in- ited-service restaurants, and beverage manufactur- formal production are more widespread (Reuter et ing, when compared to a suitable synthetic group al. 2015). The same can be said for SSB illicit trade, (Marinello et al. 2021). Similar results are obtained where there is no evidence relating higher SSB tax- in studies simulating the impact of a statewide SSB es to increased illicit trade in SSBs (Hattersley et al. tax in California and Illinois (Powell et al. 2014) 2020). SSBs’ low price relative to volume is, proba- In the case of alcohol, a study simulated the ef- bly, a “natural” hurdle for illicit trading. fect on employment of an alcohol tax increase in WHO estimates show that the global share of six US states and found that such a tax would have unrecorded alcohol consumption fell from 28.6 per- a positive impact on employment, mostly because cent for 2005 to 25.5 percent for 2016 (WHO 2011; the resulting fiscal expenditures would spur greater WHO 2018). Estimates of such a share for LICs economic activity (Wada et al. 2017). and LMICs are around 43 percent for 2016, while Finally, in the case of tobacco, there is ample for UMICs and HICs, they are around 17.5 percent evidence that tax and non-tax policies (such as (WHO 2018). Large variations also exist by region. smoke-free policies) do not have a discernible ef- fect on aggregate employment. Studies on this Taxes, economic activity, topic have been conducted for Scotland, the UK, and employment the United States and some of its individual states (Michigan, Indiana), Canada, South Africa, Zim- The tobacco, alcohol, and SSB industries routinely babwe, Bangladesh, Bulgaria, Egypt, and Indonesia assert that excise taxes on these products reduce (US National Cancer Institute and WHO 2016). In economic activity and employment as people pur- most cases, studies found a net gain of jobs under chase less of the taxed goods. This argument, which normal circumstances after a tobacco tax increase. contradicts their claims that taxes do not have an Though aggregate employment is not affected impact on consumption (Petticrew et al. 2017) is by tobacco taxes, members of some specific groups also simplistic and untrue. When taxes reduce con- (e.g., small-scale tobacco farmers) may lose income sumption of these products, it can affect sales and or even go out of business. In most cases, farmers employment in those sectors. However, consum- have long-established practices of shifting among COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 7 203 crops in response to demand and market prices deaths associated with tobacco consumption, and (Marquez and Moreno-Dobson 2017). In other cas- sparing the environmental costs associated with es, crop substitution programs can be implement- tobacco growing, are substantially higher than the ed. Alternatives to tobacco farming are present in costs of implementing these crop substitution pro- countries as diverse as Indonesia, Malawi, Kenya, grams (Lencucha et al. 2022; Lecours et al. 2012). Brazil, Canada (US National Cancer Institute and Non-tax measures, such as smoke-free poli- World Health Organization 2016) and China (Li cies, have also been cited as hurting businesses (e.g., and Tang 2018). Compensatory programs can also bars, restaurants, pubs). However, studies conduct- be financed with the tax revenues. For example, the ed in several countries have shown little effect on Philippines earmarked a significant proportion (15 sales, employment, number of establishments, busi- percent) of incremental revenues from tobacco taxes ness value, or gaming revenue (US National Cancer to help tobacco farmers shift to other crops (Chavez Institute and WHO 2016; Guerrero López 2011). et al. 2016). The benefits of reducing illnesses and CONCLUSIONS In 1776, Adam Smith stated that “sugar, rum, and tobacco are commodities which are nowhere necessaries of life, which have become objects of almost universal consumption, and which are therefore extremely proper subjects of taxation” (Smith 1776). Much time has passed since then, and evidence on the negative effects of the consumption of tobacco, alcohol, and SSBs is overwhelming (something unknown when Smith wrote). In addi- tion, economic theory has demonstrated that taxing products that generate negative externalities not only raises revenues but increases economic efficiency. Evidence collected over the past decades shows that ers lose on average a decade of life when compared tobacco taxation is the single most effective inter- to non-smokers (Jha 2020). Loss of income due to vention to curb tobacco use (Jha 1999; Internation- illness and mortality may affect households’ present al Agency for Research on Cancer 2011). Recently, and future wellbeing, creating a vicious cycle of low- WHO has included tobacco taxes as a “best buy” er human capital investment and poverty. (interventions with the highest cost-effectiveness) to Contrary to the simplistic arguments advanced reduce consumption and the burden of diseases asso- by affected industries, higher taxes on these harmful ciated with tobacco use (WHO 2017a). Similar argu- products are highly progressive (i.e., pro-poor). Be- ments can be made for alcohol taxation. In the case of cause poorer individuals are more price-responsive, SSB taxes, though they have not been included in the they have a higher propensity to reduce consumption “best buy” list, they have been singled out as a cost-ef- or quit altogether when taxes increase. Consequent- fective intervention for reducing SSB consumption. ly, they benefit disproportionately from longer and Reducing consumption of alcohol, tobacco, healthier lives, reduced spending on health care, few- and SSBs and the burden of disease associated with er lost days of work, and longer working lives (Ver- these products is not only about reducing health care guet et al. 2015). Evidence on this for countries at costs, which can be significant and put substantial different income levels is significant and consistent. pressure on health systems, it is also about increas- There is also evidence, especially for tobacco ing the social return on human capital. Chronic ill- and alcohol, that the best tax structure to decrease nesses and conditions associated with the consump- consumption relies on enacting high specific excise tion of these products hinder individuals’ productive taxes and raising them to keep pace with inflation. performance due to absenteeism and presenteeism In the case of tobacco, specific excises on cigarette (the reduction of productivity at work due to ill- “sticks” or grams of other tobacco are preferrable, ness). Meanwhile, in addition to the personal trag- as they discourage substitution to cheaper products, edy of lives lost, premature mortality linked to these which are just as harmful as more expensive ones. products implies that the social resources devoted to In addition, specific taxes limit the industry’s capac- education and health care for the deceased individu- ity to manipulate the pass-through to prices in ways als are also prematurely lost and do not deliver their that undermine the tax’s effect on consumption and full yields. Recall that, in the case of tobacco, smok- profits (International Agency for Research on Can- 204 CHAPTER 7 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E cer 2011). In the case of alcohol, specific excises on these taxes is still far from optimal. Most governments alcohol content are also recommended as the best have only enacted modest tax increases instead of the tool to decrease consumption and associated costs large excise hikes paired with non-price strategies (WHO 2020). In the case of SSBs, specific taxes on which will be most effective in reducing consump- sugar content have also been recommended as the tion. A global assessment of tobacco taxes document- best way to achieve the public health goal of reduc- ed significant progress but noted that it has been slow ing consumption of sugars (Popkin and Ng 2021). and remains inadequate (Chaloupka et al. 2021a). Taxes are an important element of broader ef- A key factor for successfully designing and forts to reduce consumption of tobacco, alcohol, adopting measures to reduce consumption of tobac- and SSBs. They are not a cure-all, however. They co, alcohol, and SSBs is to exclude the targeted in- should be used along with other cost-effective mea- dustries from all stages of decision making (techni- sures. These other measures include mass media cal and political). The industries involved, especially education campaigns; prominent product labelling the tobacco industry, have a well-documented re- showing adverse health effects (especially for tobac- cord of concealing evidence, misleading policymak- co and alcohol); restrictions on opening times (par- ers and the public, and interfering in public policy to ticularly for establishments selling alcohol); restric- oppose, delay, or undermine public health measures. tions on smoking and drinking alcoholic beverages Without reducing the consumption of these in public spaces; labelling of products with health harmful products, society will not be able to promote warnings (for tobacco, alcohol, and SSBs); and oth- healthy longevity. Chronic conditions such as hyper- ers. There is enough evidence on the effectiveness of tension and diabetes are already major sources of ill these measures to amply justify their implementa- health among today’s older adults. These conditions tion, with adaptation and prioritization depending are often correlated with functional dependence and on individual country circumstances. Combina- with other illnesses in such a way that people cannot tions of price and non-price measures have been enjoy a healthy old age. Reducing consumption of highly effective at reducing consumption, as recent these products at any age improves the chances of assessments of the WHO Framework Convention healthier living in the future. For this reason, taxing on Tobacco Control have shown (Muñoz 2021). harmful products is one of the most cost-effective Unfortunately, in too many countries the use of ways to reach a future with healthy longevity. REFERENCES 1. Allcott, Hunt, Benjamin B. Lockwood, and Dmitry Taubinsky. cial Protection and Quality of Life of Older Persons.” Inter-Amer- 2019. “Should We Tax Sugar-Sweetened Beverages? An Over- ican Development Bank. https://doi.org/10.18235/0004287. view of Theory and Evidence.” Journal of Economic Perspectives 5. Babor, Thomas F., Jack H. Mendelson, Isaac Greenberg, and 33 (3): 202–27. https://doi.org/10.1257/jep.33.3.202. John Kuehnle. 1978. “Experimental Analysis of the ‘Happy 2. Ally, Abdallah K., Yang Meng, Ratula Chakraborty, Paul W. Dob- Hour’: Effects of Purchase Price on Alcohol Consumption.” son, Jonathan S. Seaton, John Holmes, Colin Angus, et al. 2014. Psychopharmacology 58 (1): 35–41. https://doi.org/10.1007/ “Alcohol Tax Pass-through across the Product and Price Range: bf00426787. Do Retailers Treat Cheap Alcohol Differently?” Addiction 109 6. Banderali, G., A. Martelli, M. Landi, F. Moretti, F. Betti, G. Radaelli, (12): 1994–2002. https://doi.org/10.1111/add.12590. C. Lassandro, and E. Verduci. 2015. “Short and Long Term Health 3. Alvarado, Miriam, Deliana Kostova, Marc Suhrcke, Ian Ham- Effects of Parental Tobacco Smoking during Pregnancy and bleton, Trevor Hassell, T. Alafia Samuels, Jean Adams, and Lactation: A Descriptive Review.” Journal of Translational Med- Nigel Unwin. 2017. “Trends in Beverage Prices Following the icine 13 (1). https://doi.org/10.1186/s12967-015-0690-y. Introduction of a Tax on Sugar-Sweetened Beverages in Barba- 7. Bartecchi, Carl, Robert N. Alsever, Christine Nevin-Woods, dos.” Preventive Medicine 105 (December): S23–25. https://doi. William M. Thomas, Raymond O. Estacio, Becki Bucher Bar- org/10.1016/j.ypmed.2017.07.013. telson, and Mori J. Krantz. 2006. “Reduction in the Incidence 4. Aranco, Natalia, Mariano Bosch, Marco Stampini, Oliver Azuara of Acute Myocardial Infarction Associated with a Citywide Herrera, Laura Goyeneche, Pablo Ibarrarán, Déborah Oliveira, Smoking Ordinance.” Circulation 114 (14): 1490–96. https://doi. Maria Reyes Retana Torre, William D. Savedoff, and Eric Torres org/10.1161/circulationaha.106.615245. Ramirez. 2022. “Aging in Latin America and the Caribbean: So- 8. Basu, Sanjay, Martin McKee, Gauden Galea, and David Stuckler. COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 7 205 2013. “Relationship of Soft Drink Consumption to Global Over- Prevention — the Public Policy Case for Taxes on Sugared weight, Obesity, and Diabetes: A Cross-National Analysis of 75 Beverages.” New England Journal of Medicine 360 (18): 1805–8. Countries.” American Journal of Public Health 103 (11): 2071–77. https://doi.org/10.1056/nejmp0902392. https://doi.org/10.2105/ajph.2012.300974. 21. Brownell, Kelly D., and Kenneth E. Warner. 2009. “The Perils of 9. Becker, Gary S., Michael Grossman, and Kevin M. Murphy. 1994. Ignoring History: Big Tobacco Played Dirty and Millions Died. “An Empirical Analysis of Cigarette Addiction.” The American How Similar Is Big Food?” Milbank Quarterly 87 (1): 259–94. Economic Review 84 (3): 396–418. https://www.jstor.org/sta- https://doi.org/10.1111/j.1468-0009.2009.00555.x. ble/2118059. 22. Burton, Robyn, Clive Henn, Don Lavoie, Rosanna O’Connor, 10. Berardi, Nicoletta, Patrick Sevestre, Marine Tepaut, and Alexandre Clare Perkins, Kate Sweeney, Felix Greaves, et al. 2017. “A Rapid Vigneron. 2012. “The Impact of a `Soda Tax’ on Prices. Evidence Evidence Review of the Effectiveness and Cost-Effectiveness from French Micro Data.” Applied Economics 48 (41): 3976–94. of Alcohol Control Policies: An English Perspective.” The Lan- https://econpapers.repec.org/paper/bfrbanfra/415.htm. cet 389 (10078): 1558–80. https://doi.org/10.1016/s0140- 11. Bergman, U. Michael, and Niels Lynggård Hansen. 2019. “Are 6736(16)32420-5. Excise Taxes on Beverages Fully Passed through to Prices? 23. Caro, Juan Carlos, Camila Corvalán, Marcela Reyes, Andres Sil- The Danish Evidence.” FinanzArchiv 75 (4): 323. https://doi. va, Barry Popkin, and Lindsey Smith Taillie. 2018. “Chile’s 2014 org/10.1628/fa-2019-0010. Sugar-Sweetened Beverage Tax and Changes in Prices and 12. Berthet Valdois, Julie, Cornelis Van Walbeek, Hana Ross, Hema Purchases of Sugar-Sweetened Beverages: An Observational Soondram, Bhavish Jugurnath, Marie Chan Sun, and Deowan Study in an Urban Environment.” Edited by Claudia Langen- Mohee. 2019. “Tobacco Industry Tactics in Response to Ciga- berg. PLOS Medicine 15 (7): e1002597. https://doi.org/10.1371/ rette Excise Tax Increases in Mauritius.” Tobacco Control 29 (e1): journal.pmed.1002597. tobaccocontrol-2019-055196. https://doi.org/10.1136/tobac- 24. Case, Anne, and Angus Deaton. 2021. Deaths of Despair and the cocontrol-2019-055196. Future of Capitalism. NED - New edition. Princeton University 13. Bhattacharya, Aveek. 2020. “Who Pays the Tab? The Distribu- Press. https://www.jstor.org/stable/j.ctv161f3f8. tional Effects of UK Alcohol Taxes.” London, UK: Institute of 25. Cawley, John, and David E. Frisvold. 2016. “The Pass-through Alcohol Studies. https://www.ias.org.uk/uploads/pdf/IAS%20 of Taxes on Sugar-Sweetened Beverages to Retail Prices: The reports/rp40022020.pdf. Case of Berkeley, California.” Journal of Policy Analysis and Man- 14. Blecher, Evan, Alex Liber, Jeffrey Drope, Binh Nguyen, and agement 36 (2): 303–26. https://doi.org/10.1002/pam.21960. Michal Stoklosa. 2017. “Global Trends in the Affordability of 26. CBC News. 2008. “Tobacco Giants to Pay up to $1.15B over Con- Sugar-Sweetened Beverages, 1990–2016.” Preventing Chronic traband Sales.” CBC News, July 31, 2008. https://www.cbc.ca/ Disease 14 (May). https://doi.org/10.5888/pcd14.160406. news/canada/tobacco-giants-to-pay-up-to-1-15b-over-con- 15. Blecher, Evan, Alex Liber, Corné Van Walbeek, and Laura Ros- traband-sales-1.701089. souw. 2018. “An International Analysis of the Price and Afford- 27. Chacon, Violeta, Guillermo Paraje, Joaquin Barnoya, and Frank ability of Beer.” Edited by Isabel Novo-Cortí. PLOS ONE 13 (12): J. Chaloupka. 2018. “Own-Price, Cross-Price, and Expenditure e0208831. https://doi.org/10.1371/journal.pone.0208831. Elasticities on Sugar-Sweetened Beverages in Guatemala.” Ed- 16. Blecher, Evan, and Corné van Walbeek. 2009. “Cigarette Af- ited by Rachel A. Nugent. PLOS ONE 13 (10): e0205931. https:// fordability Trends: An Update and Some Methodological doi.org/10.1371/journal.pone.0205931. Comments.” Tobacco Control 18 (3): 167–75. https://doi. 28. Chaloupka, Frank, Jeff Drope, Erika Siu, Violeta Vulovic, Maryam org/10.1136/tc.2008.026682. Mirza, Germán Rodriguez-Iglesias, Anh Ngo, et al. 2021a. To- 17. Bloom, David E, Somnath Chatterji, Paul Kowal, Peter bacconomics: Cigarette Tax Scorecard . 2nd ed. Chicago, IL: Lloyd-Sherlock, Martin McKee, Bernd Rechel, Larry Rosen- Health Policy Center, Institute for Health Research and Policy, berg, and James P Smith. 2015. “Macroeconomic Implications University of Illinois Chicago. www.tobacconomics.org. of Population Ageing and Selected Policy Responses.” The 29. Chaloupka, Frank, and Rosalie L. Pacula. 1999. “Sex and Race Lancet 385 (9968): 649–57. https://doi.org/10.1016/s0140- Differences in Young People’s Responsiveness to Price and To- 6736(14)61464-1. bacco Control Policies.” Tobacco Control 8 (4): 373–77. https:// 18. BMJ. 2002. “How to Access Tobacco Industry Documents.” doi.org/10.1136/tc.8.4.373. Tobacco Control 11 (Supplement 1): i39–39. https://doi. 30. Chaloupka, Frank, and Lisa Powell. 2018. “Using Fiscal Policy to org/10.1136/tc.11.suppl_1.i39. Promote Health: Taxing Tobacco, Alcohol, and Sugary Bever- 19. Booth, Andrew, Petra Meier,Tim Stockwell, Anthea Sutton, ages.” Background Paper Prepared for the Task Force on Fiscal Anna Wilkinson, Ruth Wong et al. 2008. “IndependentReview Policy for Health. New York: Bloomberg Philanthropies. of the Effects of Alcohol Pricing and Promotion. Part A: System- 31. Chaloupka, Frank J., Lisa M. Powell, and Kenneth E. Warner. aticReviews.” Sheffield, UK: University of Sheffield. 2019. “The Use of Excise Taxes to Reduce Tobacco, Alco- 20. Brownell, Kelly D., and Thomas R. Frieden. 2009. “Ounces of hol, and Sugary Beverage Consumption.” Annual Review of 206 CHAPTER 7 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E Public Health 40 (1): 187–201. https://doi.org/10.1146/an- Tobacco, Alcohol, and Sugary Beverage Industries.” Back- nurev-publhealth-040218-043816. ground Paper for the Task Force on Fiscal Policy for Health. Uni- 32. Chaloupka, Frank, Jeff Drope, Erika Siu, Violeta Vulovic, Michal versity of Edinburgh. https://data.bloomberglp.com/dotorg/ Stoklosa, Maryam Mirza, Germán Rodriguez-Iglesias, and Hye sites/2/2019/04/Structure-and-Tactics-of-the-Tobacco-Alco- Myung Lee. 2021b. “Tobacconomics Cigarette Tax Scorecard: hol-and-Sugary-Beverage-Industries.pdf. Scoring Component Policy Note – Cigarette Price.” Chicago, 42. DeCicca, Philip, Don Kenkel, and Alan Mathios. 2008. “Cigarette IL: Health Policy Center, Institute for Health Research and Pol- Taxes and the Transition from Youth to Adult Smoking: Smok- icy, University of Illinois Chicago. https://tobacconomics.org/ ing Initiation, Cessation, and Participation.” Journal of Health files/research/695/uic-tobacco-scorecard-component-poli- Economics 27 (4): 904–17. https://doi.org/10.1016/j.jheale- cy-note-cigarette-price-v4.0.pdf. co.2008.02.008. 33. Chaloupke, Frank, and Lisa Powell. 2019. “Health Taxes to Save 43. Dee, Thomas S. 1999. “The Complementarity of Teen Smok- Lives - Background Materials: Case Studies.” Bloomberg Philan- ing and Drinking.” Journal of Health Economics 18 (6): 769–93. thropies Task Force on Fiscal Policy for Health. Chicago, IL: Toba- https://doi.org/10.1016/s0167-6296(99)00018-1. conomics. https://tobacconomics.org/research/health-taxes- 44. Del Carmen, Giselle, Alan Fuchs, and Maria Eugenia Genoni. to-save-lives-background-materials-case-studies/. 2018. “The Distributional Impacts of Cigarette Taxation in Ban- 34. Chavez, Jenina J, Jeff Drope, Qing Li, and Madeiline J Aloria. gladesh.” Policy Research Working Paper No. 8580. Washington, 2016. “The Economics of Tobacco Farming in the Philippines.” D.C.: World Bank. https://doi.org/10.1596/1813-9450-8580. Quezon City: Action for Economic Reforms and Atlanta: Amer- 45. Delipalla, Sophia, and Owen O’Donnell. 2001. “Estimating Tax ican Cancer Society. https://www.cancer.org/content/dam/ Incidence, Market Power and Market Conduct: The European cancer-org/research/economic-and-healthy-policy/the-eco- Cigarette Industry.” International Journal of Industrial Organiza- nomics-of-tobacco-farming-in-the-philippines.pdf. tion 19 (6): 885–908. https://econpapers.repec.org/article/eee- 35. Chelwa, Grieve, and Cornelis van Walbeek. 2014. “Assessing the indorg/v_3a19_3ay_3a2001_3ai_3a6_3ap_3a885-908.htm. Causal Impact of Tobacco Expenditure on Household Spending 46. Diosana, Jo-Ann L., and Filomeno S. Sta. Ana III. 2020. “Case Patterns in Zambia.” ERSA Working Paper 453. Cape Town, South Studies in Illicit Tobacco Trade: The Philippines.” Tobacconom- Africa: Economic Research Southern Africa. https://econrsa. ics, Health Policy Center, Institute for Health Research and Pol- org/wp-content/uploads/2022/06/working_paper_453.pdf. icy, University of Illinois, Chicago. https://portal-uat.who.int/ 36. Chen, Danhong, David Abler, De Zhou, Xiaohua Yu, and Wyatt fctcapps/sites/default/files/kh-media/e-library-doc/2020/07/ Thompson. 2015. “A Meta‐Analysis of Food Demand Elastici- case-studies-in-illicit-tobacco-trade-the-Philippines.pdf. ties for China.” Applied Economic Perspectives and Policy 38 (1): 47. Do, Young Kyung, and Mary Ann Bautista. 2015. “Tobacco Use 50–72. https://doi.org/10.1093/aepp/ppv006. and Household Expenditures on Food, Education, and Health- 37. Cobacho Tornel, Ma Belén, Ángel López Nicolás, and José María care in Low- and Middle-Income Countries: A Multilevel Anal- Ramos Parreño. 2010. “El Coste de Mortalidad Asociado al Con- ysis.” BMC Public Health 15 (1). https://doi.org/10.1186/s12889- sumo de Tabaco En España” [The cost of mortality associated 015-2423-9. with tobacco consumption in Spain]. Revista Española de Salud 48. Dobbs, Richard, Corrine Sawers, Fraser Thompson, James Pública 84 (3): 271–80. https://scielo.isciii.es/scielo.php?pid=S Manyika, Jonathan Woetzel, Peter Child, Sorcha McKenna, and 1135-57272010000300005&script=sci_abstract. Angela Spatharou. 2014. “Overcoming Obesity: An Initial Eco- 38. Coelho, Daniel. 2020. “Distributive Impactof Health Taxes in nomic Analysis.” Discussion Paper. McKinsey Global Institute. Brazil.” Unpublishedreport, World Bank, Washington, DC. 49. Doogan, Nathan J, Mary Ellen Wewers, and Micah Berman. 39. Colchero, M. Arantxa, Juan Carlos Salgado, Mishel Unar-Mun- 2017. “The Impact of a Federal Cigarette Minimum Pack Price guía, Mauricio Hernández-Ávila, and Juan Angel Rivera-Dom- Policy on Cigarette Use in the USA.” Tobacco Control 27 (2): marco. 2015. “Price Elasticity of the Demand for Sugar Sweet- 203–8. https://doi.org/10.1136/tobaccocontrol-2016-053457. ened Beverages and Soft Drinks in Mexico.” Economics & Human 50. Drope, Jeffrey, Roberto Iglesias, Valeska Carvalho Figueiredo, Biology 19 (December): 129–37. https://doi.org/10.1016/j. SzkloAndré Salem, Paulo Borges, Michal Stoklosa, Kevin Weld- ehb.2015.08.007. ing, Hannah Nascimento, and Caroline Cortes. 2022. “Consumo 40. Colchero, M. Arantxa, Juan Carlos Salgado, Mishel Unar-Mun- de Cigarros Ilegais Em Cinco Cidades Brasileiras.” Centro de Es- guía, Mariana Molina, Shuwen Ng, and Juan Angel Rive- tudos sobre Tabaco e Saúde (Cetab). Escola Nacional de Saúde ra-Dommarco. 2015. “Changes in Prices after an Excise Tax to Pública Sérgio Arouca, Fiocruz. Rio de Janeiro: Outras Letras. Sweetened Sugar Beverages Was Implemented in Mexico: 51. Dunbar, Andrew, William Gotsis, and William Frishman. 2013. Evidence from Urban Areas.” Edited by Rachel A. Nugent. “Second-Hand Tobacco Smoke and Cardiovascular Dis- PLOS ONE 10 (12): e0144408. https://doi.org/10.1371/journal. ease Risk.” Cardiology in Review 21 (2): 94–100. https://doi. pone.0144408. org/10.1097/crd.0b013e31827362e4. 41. Collin, Jeff, and Sarah Hill. 2019. “Structure and Tactics of the 52. Edwards, Griffith. 1997. “Alcohol Policy and the Public Good.” COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 7 207 Addiction (Abingdon, England) 92 Suppl 1 (March): S73-79. ened Beverages: The Case of Kazakhstan.” Poverty and Equi- https://pubmed.ncbi.nlm.nih.gov/9167291/. ty Notes No. 24. Washington, D.C.: World Bank. https://doi. 53. Epstein, Leonard H, Noelle Jankowiak, Chantal Nederkoorn, org/10.1596/33859. Hollie A Raynor, Simone A French, and Eric Finkelstein. 2012. 64. Fuchs, Alan, Mikhail Matytsin, and Olga Obukhova. 2018. “To- “Experimental Research on the Relation between Food Price bacco Taxation Incidence: Evidence from the Russian Federa- Changes and Food-Purchasing Patterns: A Targeted tion.” Policy Research Working Paper No. 8626. Washington, D.C.: Review.” The American Journal of Clinical Nutrition 95 (4): World Bank. https://doi.org/10.1596/1813-9450-8626. 789–809. https://doi.org/10.3945/ajcn.111.024380. 65. Fuchs, Alan, and Francisco Meneses. 2017a. “Regressive or Pro- 54. Feighery, Ellen, Kurt Ribisl, NC Schleicher, Leslie Zellers, and N gressive ? The Effect of Tobacco Taxes in Ukraine.” Policy Research Wellington. 2005. “How Do Minimum Cigarette Price Laws Af- Working Paper No. 8227. Washington, D.C.: World Bank. https:// fect Cigarette Prices at the Retail Level?” Tobacco Control 14 (2): documents.worldbank.org/en/publication/documents-re- 80–85. https://www.jstor.org/stable/20747779. ports/documentdetail/934621509107306525/regressive-or- 55. Ferrante, Daniel, Bruno Linetzky, Mario Virgolini, Veronica Schoj, progressive-the-effect-of-tobacco-taxes-in-ukraine. and B Apelberg. 2011. “Reduction in Hospital Admissions for 66. Jha P, Guindon E, Joseph RA,Nandi A, John R, Rao K, et al. A ra- Acute Coronary Syndrome after the Successful Implementa- tional taxation system of bidis andcigarettes to reduce smoking tion of 100% Smoke-Free Legislation in Argentina: A Compar- deaths in India. Economic and Political Weekly.2011;46:44-51. ison with Partial Smoking Restrictions.” Tobacco Control 21 (4): 67. Muñoz M. 2021. “Relación entre Consumo de Tabaco y el 402–6. https://doi.org/10.1136/tc.2010.042325. FrameworkConvention on Tobacco Control: Un Análisis Be- 56. Finkelstein, Eric A., Chen Zhen, Marcel Bilger, James Nonne- fore-After” [Relationship between Tobacco Consumption and maker, Assad M. Farooqui, and Jessica E. Todd. 2013. “Impli- the Framework Convention on Tobacco Control: A Before-After cations of a Sugar-Sweetened Beverage (SSB) Tax When Sub- Analysis]. Santiago, Chile: Universidad Adolfo Ibáñez. stitutions to Non-Beverage Items Are Considered.” Journal of 68. ———. 2017b. “Are Tobacco Taxes Really Regressive?: Evi- Health Economics 32 (1): 219–39. https://doi.org/10.1016/j. dence from Chile.” Washington, D.C.: World Bank. https://doi. jhealeco.2012.10.005. org/10.1596/25969. 57. Forster, Martin, and Andrew M. Jones. 2001. “The Role of 69. ———. 2018. “Tobacco Price Elasticity and Tax Progressivity in Tobacco Taxes in Starting and Quitting Smoking: Duration Moldova.” Policy Research Working Paper No. 8327. Washington, Analysis of British Data.” Journal of the Royal Statistical Soci- D.C.: World Bank. https://doi.org/10.1596/1813-9450-8327. ety: Series a (Statistics in Society) 164 (3): 517–47. https://doi. 70. Fuchs, Alan, Edvard Orlic, and Cesar Cancho. 2019. “Time to org/10.1111/1467-985x.00217. Quit: The Tobacco Tax Increase and Household Welfare in Bos- 58. Fuchs, Alan, and María FernandaGonzález Icaza. 2021. “The nia and Herzegovina.” Washington, D.C.: World Bank. https:// Welfareand Distributional Effects of TaxingSSB to Reduce doi.org/10.1596/31249. the Risks of Obesity inUkraine.” Unpublished report, World 71. Gallet, Craig A. 2007. “The Demand for Alcohol: A Meta-Anal- Bank,Washington, DC. ysis of Elasticities.” The Australian Journal of Agricultural and 59. Fuchs Tarlovsky, Alan, and Fernanda Gonzalez Icaza. 2019. “The Resource Economics 51 (2): 121–35. https://doi.org/10.1111/ Welfare and Distributional Effects of Increasing Taxes on To- j.1467-8489.2007.00365.x. bacco in Viet Nam.” Washington, D.C.: World Bank. https:// 72. Gallet, Craig A., and John A. List. 2003. “Cigarette Demand: A doi. org/10.1596/32062. Meta-Analysis of Elasticities.” Health Economics 12 (10): 821–35. 60. Fuchs, Alan, and Giselle Del Carmen. 2018. “The Distributional https://doi.org/10.1002/hec.765. Effects of Tobacco Taxation: The Evidence of White and Clove 73. Gawryszewski, Vilma Pinheiro, and Maristela G. Monteiro. 2014. Cigarettes in Indonesia.” Policy Research Working Paper No. 8558. “Mortality from Diseases, Conditions and Injuries Where Alco- Washington, D.C.: World Bank. https:// hol Is a Necessary Cause in the Americas, 2007-09.” Addiction doi.org/10.1596/1813-9450-8558. 109 (4): 570–77. https://doi.org/10.1111/add.12418. 61. Fuchs, Alan, Giselle Del Carmen, and Alfred Kechia Mukong. 74. Gilmore, Anna B, Allen W A Gallagher, and Andy Rowell. 2018. 2018. “Long-Run Impacts of Increasing Tobacco Taxes.” Policy “Tobacco Industry’s Elaborate Attempts to Control a Global Research Working Paper No. 8369. Washington, D.C.: World Bank. Track and Trace System and Fundamentally Undermine the Il- https://doi.org/10.1596/1813-9450-8369. licit Trade Protocol.” Tobacco Control 28 (2): 127–40. https://doi. 62. Fuchs, Alan, Fernanda Gonzalez Icaza, and Daniela Paula Paz. org/10.1136/tobaccocontrol-2017-054191. 2019. “Distributional Effects of Tobacco Taxation: A Comparative 75. Global Tobacco Economic Consortium. 2018. “The Health, Pov- Analysis.” Policy Research Working Paper No. 8805. Washington, erty, and Financial Consequences of a Cigarette Price Increase D.C.: World Bank. https://doi.org/10.1596/1813-9450-8805. among 500 Million Male Smokers in 13 Middle Income Coun- 63. Fuchs, Alan, Kate Mandeville, and Ana Cristina Alonso-Soria. tries: Compartmental Model Study.” BMJ 361 (April): k1162. 2020. “Health and Distributional Effects Taxing Sugar-Sweet- https://doi.org/10.1136/bmj.k1162. 208 CHAPTER 7 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E 76. Gonzalez-Rozada, Martin, and Giselle Montamat. 2019. “How org/10.1186/s12889-017-4098-x. Raising Tobacco Prices Affects the Decision to Start and Quit 87. Guindon, G. Emmanuel, Arindam Nandi, Frank J. Chaloupka, Smoking: Evidence from Argentina.” International Journal of En- and Prabhat Jha. 2011. “Socioeconomic Differences in the vironmental Research and Public Health 16 (19): 3622. https:// Impact of Smoking Tobacco and Alcohol Prices on Smoking doi.org/10.3390/ijerph16193622. in India.” Social Science Research Network. Rochester, NY. No- 77. Goodchild, Mark, Nigar Nargis, and Edouard Tursan d’Espaig- vember 1, 2011. https://papers.ssrn.com/sol3/papers.cfm?ab- net. 2017. “Global Economic Cost of Smoking-Attributable Dis- stract_id=1954506. eases.” Tobacco Control 27 (1): 58–64. https://doi.org/10.1136/ 88. Guindon, G. Emmanuel, Guillermo R. Paraje, and Frank J. Cha- tobaccocontrol-2016-053305. loupka. 2018. “The Impact of Prices and Taxes on the Use of 78. Goodchild, Mark, Jeremias Paul, Roberto Iglesias, Annerie Tobacco Products in Latin America and the Caribbean.” Amer- Bouw, and Anne-Marie Perucic. 2020. “Potential Impact of ican Journal of Public Health 108 (S6): S492–502. https://doi. Eliminating Illicit Trade in Cigarettes: A Demand-Side Perspec- org/10.2105/ajph.2014.302396r. tive.” Tobacco Control Published Online First: (November 3, 89. ———. 2019. “Association of Tobacco Control Policies with 2020). https://doi.org/10.1136/tobaccocontrol-2020-055980. Youth Smoking Onset in Chile.” JAMA Pediatrics 173 (8): 754. 79. Goodchild, Mark, Anne-Marie Perucic, and Nigar Nargis. 2016. https://doi.org/10.1001/jamapediatrics.2019.1500. “Modelling the Impact of Raising Tobacco Taxes on Public 90. Guindon, G. Emmanuel, Guillermo R. Paraje, and Ricardo Health and Finance.” Bulletin of the World Health Organization Chávez. 2017. “Prices, Inflation, and Smoking Onset: The Case 94 (4): 250–57. https://doi.org/10.2471/blt.15.164707. of Argentina.” Economic Inquiry 56 (1): 424–45. https://doi. 80. Griswold, Max G, Nancy Fullman, Caitlin Hawley, Nicholas org/10.1111/ecin.12490. Arian, Stephanie R M Zimsen, Hayley D Tymeson, Vidhya Ven- 91. Guindon, G. Emmanuel, Kevin Zhao, Tooba Fatima, Sophiya kateswaran, et al. 2018. “Alcohol Use and Burden for 195 Coun- Garasia, Nicholas Quinn, N. Bruce Baskerville, and Guillermo tries and Territories, 1990–2016: A Systematic Analysis for the Paraje. 2022. “Prices, Taxes and Alcohol Use: A Systematic Global Burden of Disease Study 2016.” The Lancet 392 (10152): Umbrella Review.” Addiction 117 (12): 3004–23. https://doi. 1015–35. https://doi.org/10.1016/s0140-6736(18)31310-2. org/10.1111/add.15966. 81. Grogger, Jeffrey T. 2016. “Soda Taxes and the Prices of Sodas 92. Harris, Jeffrey E. 1987. “The 1983 Increase in the Federal Cig- and Other Drinks: Evidence from Mexico.” SSRN Electronic Jour- arette Excise Tax.” Tax Policy and the Economy 1 (January): 87– nal. https://doi.org/10.2139/ssrn.2725043. 111. https://doi.org/10.1086/tpe.1.20061764. 82. Gruenewald, Paul J., William R. Ponicki, Harold D. Holder, and 93. Hattersley , Libby, Alan Fuchs, Alberto Gonima , Lynn Silver, and Anders Romelsjo. 2006. “Alcohol Prices, Beverage Quality, and Kate Mandeville . 2020. “Evidence Summary: Business, Employ- the Demand for Alcohol: Quality Substitutions and Price Elas- ment, and Productivity Impacts of Sugar-Sweetened Beverag- ticities.” Alcoholism: Clinical and Experimental Research 30 (1): es Taxes 1.” Washington, D.C.: World Bank. https://documents1. 96–105. https://doi.org/10.1111/j.1530-0277.2006.00011.x. worldbank.org/curated/zh/972421592559284445/pdf/Busi- 83. Guerchet, Maëlenn M., Mariella Guerra, Yueqin Huang, Peter ness-Employment-and-Productivity-Impacts-of-Sugar-Sweet- Lloyd-Sherlock, Ana Luisa Sosa, Richard Uwakwe, Isaac Acosta, ened-Beverages-Taxes-Evidence-Summary.pdf. et al. 2018. “A Cohort Study of the Effects of Older Adult Care 94. Hattersley, Libby, Alessia Thiebaud, Lynn Silver, and Kate Man- Dependence upon Household Economic Functioning, in Peru, deville. 2020. “Countering Common Arguments against Taxes Mexico and China.” Edited by Stefano Federici. PLOS ONE 13 (4): on Sugary Drinks.” Health, Nutrition and Population Knowledge e0195567. https://doi.org/10.1371/journal.pone.0195567. Brief. Washington, D.C.: World Bank. https://openknowledge. 84. Guerrero Lopez, C. M., J. A. Jimenez Ruiz, L. M. Reynales Shige- worldbank.org/handle/10986/34361. matsu, and H. R. Waters. 2011. “The Economic Impact of Mexico 95. Heijdra, Ben J., and Laurie S. M. Reijnders. 2016. “Human Cap- City’s Smoke-Free Law.” Tobacco Control 20 (4): 273–78. https:// ital Accumulation and the Macroeconomy in an Ageing Soci- doi.org/10.1136/tc.2010.036467. ety.” De Economist 164 (3): 297–334. https://doi.org/10.1007/ 85. Guerrero-López, Carlos M., Mariana Molina, and M. Arantxa s10645-016-9283-2. Colchero. 2017. “Employment Changes Associated with the 96. Helble, Matthias, and Kris Francisco. 2017. “The Imminent Obe- Introduction of Taxes on Sugar-Sweetened Beverages and sity Crisis in Asia and the Pacific: First Cost Estimates.” Working Nonessential Energy-Dense Food in Mexico.” Preventive Med- Paper No. 743. Manilla, The Philippines: Asian Development icine 105 (December): S43–49. https://doi.org/10.1016/j. Bank. https://www.adb.org/publications/imminent-obesity-cri- ypmed.2017.09.001. sis-asia-and-pacific-first-cost-estimates. 86. Guerrero-López, Carlos M., Mishel Unar-Munguía, and M. 97. Hodge, Allison M, Julie K Bassett, Roger L Milne, Dallas R English, Arantxa Colchero. 2017. “Price Elasticity of the Demand for and Graham G Giles. 2018. “Consumption of Sugar-Sweetened Soft Drinks, Other Sugar-Sweetened Beverages and Energy and Artificially Sweetened Soft Drinks and Risk of Obesity-Re- Dense Food in Chile.” BMC Public Health 17 (1). https://doi. lated Cancers.” Public Health Nutrition 21 (9): 1618–26. https:// COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 7 209 doi.org/10.1017/s1368980017002555. in India.” Economic and Political Weekly XLVI (42). https://www. 98. Huang, Jidong, Jamie F Chriqui, Hillary DeLong, Maryam Mirza, tobacconomics.org/files/research/73/A-rational-taxation-sys- Megan C Diaz, and Frank J Chaloupka. 2016. “Do State Mini- tem-of-bidis-and-cigarettes-to-reduce...India-2011.pdf. mum Markup/Price Laws Work? Evidence from Retail Scan- 107. Jha, Prabhat, Catherine Hill, Daphne C N Wu, and Richard ner Data and TUS-CPS.” Tobacco Control 25 (Suppl 1): i52–59. Peto. 2020. “Cigarette Prices, Smuggling, and Deaths in France https://doi.org/10.1136/tobaccocontrol-2016-053093. and Canada.” The Lancet 395 (10217): 27–28. https://doi. 99. Hyland, A, FL Laux, C Higbee, G Hastings, H Ross, FJ Chaloupka, org/10.1016/s0140-6736(19)31291-7. GT Fong, and KM Cummings. 2006. “Cigarette Purchase Pat- 108. Jha, Prabhat, Mary MacLennan, Frank. J. Chaloupka, Ayda terns in Four Countries and the Relationship with Cessation: Yurekli, Chintanie Ramasundarahettige, Krishna Palipudi, Wi- Findings from the International Tobacco Control (ITC) Four told Zatonksi, Samira Asma, and Prakash C. Gupta. 2015. “Global Country Survey.” Tobacco Control 15 (suppl_3): iii59–64. https:// Hazards of Tobacco and the Benefits of Smoking Cessation and doi.org/10.1136/tc.2005.012203. Tobacco Taxes.” Disease Control Priorities, Third Edition (Volume 100. Imamura, Fumiaki, Laura O’Connor, Zheng Ye, Jaakko Mursu, 3): Cancer. https://doi.org/10.1596/978-1-4648-0349-9_ch10. Yasuaki Hayashino, Shilpa N Bhupathiraju, and Nita G Forouhi. 109. Jha, Prabhat, and Richard Peto. 2014. “Global Effects of Smok- 2015. “Consumption of Sugar Sweetened Beverages, Artificially ing, of Quitting, and of Taxing Tobacco.” New England Journal Sweetened Beverages, and Fruit Juice and Incidence of Type of Medicine 370 (1): 60–68. https://doi.org/10.1056/nejm- 2 Diabetes: Systematic Review, Meta-Analysis, and Estimation ra1308383. of Population Attributable Fraction.” BMJ 351 (July): h3576. 110. Jha, Prabhat, Chinthanie Ramasundarahettige, Victoria Lands- https://doi.org/10.1136/bmj.h3576. man, Brian Rostron, Michael Thun, Robert N Anderson, Tim 101. Institute of Health Metrics and Evaluation (IHME). 2020. “Global McAfee, and Richard Peto. 2013. “21st-Century Hazards of Burden of Disease Study 2019 (GBD 2019) Results.” Institute for Smoking and Benefits of Cessation in the United States.” The Health Metrics and Evaluation (IHME). Seattle, United States. New England Journal of Medicine 368 (4): 341–50. https://doi. 2020. http://ghdx.healthdata.org/gbd-results-tool. org/10.1056/NEJMsa1211128. 102. International Agency for Research on Cancer (IARC). 2011. 111. Jha, Prahbat, and Frank J Chaloupka. 1999. “Curbing the Epi- “Effectiveness of Tax and Price Policies for Tobacco Control.” demic: Governments and the Economics of Tobacco Control.” IARC Handbooks of Cancer Prevention Volume 14. Lyon, France: Washington, D.C.: World Bank. https://documents1.worldbank. International Agency for Research on Cancer. https://publica- org/curated/en/914041468176678949/pdf/multi-page.pdf. tions.iarc.fr/Book-And-Report-Series/Iarc-Handbooks-Of-Can- 112. John, Rijo M, Hana Ross, and Evan Blecher. 2011. “Tobacco Ex- cer-Prevention/Effectiveness-Of-Tax-And-Price -Poli- penditure and Its Implications for Household Resource Alloca- cies-For-Tobacco-Control-2011. tion in Cambodia.” Tobacco Control 21 (3): 341–46. https://doi. 103. Jackson, Rachel, Maxine Jackson, Fiona Campbell, Josie Messi- org/10.1136/tc.2010.042598. na, Louise Guillaume, Petra Meier, Elizabeth Goyder, Jim Chil- 113. John, Rijo M. 2008. “Crowding out Effect of Tobacco Expendi- cott, and Nick Payne. 2010. “Interventions on Control of Alcohol ture and Its Implications on Household Resource Allocation Price, Promotion and Availability for Prevention of Alcohol Use in India.” Social Science & Medicine 66 (6): 1356–67. https://doi. Disorders in Adults and Young People.” ScHARR Public Health org/10.1016/j.socscimed.2007.11.020. Collaborating Centre, University of Sheffield. https://www. 114. Joossens, Luk, David Merriman, Hana Ross, and Martin Raw. nice.org.uk/guidance/ph24/evidence/interventions-on-con- 2010. “The Impact of Eliminating the Global Illicit Cigarette trol-of-alcohol-price-promotion-and-availability-for-preven- Trade on Health and Revenue.” Addiction 105 (9): 1640–49. tion-of-alcohol-use-disorders-371568349. https://doi.org/10.1111/j.1360-0443.2010.03018.x. 104. Jawad, Mohammed, John Tayu Lee, Stanton Glantz, and Chris- 115. Joossens, Luk, and Martin Raw. 2012. “From Cigarette Smug- topher Millett. 2018. “Price Elasticity of Demand of Non-Ciga- gling to Illicit Tobacco Trade.” Tobacco Control 21 (2): 230–34. rette Tobacco Products: A Systematic Review and Meta-Anal- https://doi.org/10.1136/tobaccocontrol-2011-050205. ysis.” Tobacco Control 27 (6): 689–95. https://doi.org/10.1136/ 116. Juster, Harlan R., Brett R. Loomis, Theresa M. Hinman, Mat- tobaccocontrol-2017-054056. thew C. Farrelly, Andrew Hyland, Ursula E. Bauer, and Guthrie 105. Jha, Prabhat. 2020. “The Hazards of Smoking and the Bene- S. Birkhead. 2007. “Declines in Hospital Admissions for Acute fits of Cessation: A Critical Summation of the Epidemiological Myocardial Infarction in New York State after Implementa- Evidence in High-Income Countries.” ELife 9 (1). https://doi. tion of a Comprehensive Smoking Ban.” American Journal org/10.7554/elife.49979. of Public Health 97 (11): 2035–39. https://doi.org/10.2105/ 106. Jha, Prabhat, Emmanuel Guindon, Renu Joseph, Arindam ajph.2006.099994. Nandi, Rijo John, Kavita Rao, Frank Chaloupka, Jagdish Kaur, 117. Kalyanaram, Gurumurthy, and Russell S. Winer. 2022. “Behavior- Prakash C Gupta, and M Govinda Rao. 2011. “A Rational Taxa- al Response to Price: Data-Based Insights and Future Research tion System of Bidis and Cigarettes to Reduce Smoking Deaths for Retailing.” Journal of Retailing 98 (1): 46–70. https://ideas. 210 CHAPTER 7 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E repec.org/a/eee/jouret/v98y2022i1p46-70.html. 2030: A Modelling Study.” The Lancet 393 (10190): 2493–2502. 118. Kenkel, Donald S. 2005. “Are Alcohol Tax Hikes Ful- https://doi.org/10.1016/s0140-6736(18)32744-2. ly Passed through to Prices? Evidence from Alaska.” 129. Margozzini, Paula, Guillermo Paraje, Pedro Zitko, Manuel Espi- American Economic Review 95 (2): 273–77. https://doi. noza, Rubén Rojas, Carlos Balmaceda, Tomas Abbot, Paula Be- org/10.1257/000282805774670284. dregal, and Patricia Berrios. 2017. “Estudio Del Costo Económi- 119. Kostova, Deliana, Muhammad J Husain, and Frank J Cha- co Y Social Del Consumo de Alcohol En Chile [Study on the loupka. 2016. “Effect of Cigarette Prices on Smoking Initia- Economic and Social Cost of Consuming Alcohol in Chile].” tion and Cessation in China: A Duration Analysis.” Tobacco Santiago, Chile: Departamento de Salud Pública, Facultad de Control 26 (5): 569–74. https://doi.org/10.1136/tobaccocon- Medicina, Pontificia Universidad Católica de Chile. https:// trol-2016-053338. medicina.uc.cl/wp-content/uploads/2019/03/COSTO-ALCO- 120. Kostova, Delliana, Hana Ross, Evan Blecher, and Sara Markow- HOL_Actualizacio%CC%81n-2018_Informe.pdf. itz. 2011. “Is Youth Smoking Responsive to Cigarette Prices? Ev- 130. Marinello, Samantha, Julien Leider, and Lisa M. Powell. 2021. idence from Low- and Middle-Income Countries.” Tobacco Con- “Employment Impacts of the San Francisco Sugar-Sweetened trol 20 (6): 419–24. https://doi.org/10.1136/tc.2010.038786. Beverage Tax 2 Years after Implementation.” Edited by Guiller- 121. Kyriakoudes, Louis M. 2016. “The Tobacco Industry’s Deadly mo Paraje. PLOS ONE 16 (6): e0252094. https://doi.org/10.1371/ Distortions of History.” Tobacco Control 25 (5): 491–91. https:// journal.pone.0252094. doi.org/10.1136/tobaccocontrol-2015-052712. 131. Marquez, Patricio V., and Blanca Moreno-Dodson. 2017. “To- 122. Larrain, Luis. 2017. “Fascismo, Alimentos Y Libertad de Ex- bacco Tax Reform at the Crossroads of Health and Develop- presión [Facism, Food, and Freedom of Expression].” Libertad Y ment: A Multisectoral Perspective .” Washington, D.C.: World Desarrollo, January 5, 2017. https://lyd.org/opinion/2017/01/ Bank. https://doi.org/10.1596/28494. columna-luis-larrain-diario-financiero-fascismo-alimentos-lib- 132. Masa-ud, Abdul Gafar, Grieve Chelwa, and Corné van Walbeek. ertad-expresion/. 2020.“DoesTobacco Expenditure Influence Household Spending 123. Lecours, Natacha, Guilherme E G Almeida, Jumanne M Ab- Patterns in Ghana?: Evidence from the Ghana 2012/2013 dallah, and Thomas E Novotny. 2012. “Environmental Health Living Standards Survey.” Tobacco Induced Diseases 18 (June). Impacts of Tobacco Farming: A Review of the Literature: Table https://doi.org/10.18332/tid/120936. 1.” Tobacco Control 21 (2): 191–96. https://doi.org/10.1136/to- 133. Matzopoulos, Richard G, Sarah Truen, Brett Bowman, and Jo- baccocontrol-2011-050318. anne Corrigall. 2014. “The Cost of Harmful Alcohol Use in South 124. Leider, Julien, Vanessa Oddo, and Lisa Powell. 2021. “A Review Africa.” South African Medical Journal 104 (2): 127–32. https:// of the Effects of U.S. Local Sugar- Sweetened Beverage Taxes doi.org/10.7196/samj.7644. on Substitution to Untaxed Beverages and Food Items.” P3RC 134. Matzopoulos, Richard Gregory, Sarah Truen, Brett Bowman, Research Brief No. 123. Chicago, IL: Policy, Practice, and Preven- and Joanne Corrigall. 2013. “The Cost of Harmful Alcohol Use in tion Research Center (P3RC), University of Illinois, Chicago. South Africa.” South African Medical Journal 104 (2): 127. https:// https://p3rc.uic.edu/wp-content/uploads/sites/561/2021/12/ doi.org/10.7196/samj.7644. Rvw-Effcts-US-SSB-Taxes-Sbsttn-to-Untxd-Bev-Foods_Rsrch- 135. Meng, Yang, Alan Brennan, Robin Purshouse, Daniel Hill-Mc- Brf-No.-123_Nov-2021.pdf. Manus, Colin Angus, John Holmes, and Petra Sylvia Meier. 125. Lencucha, Raphael, Jeffrey Drope, Peter Magati, and Gumilang 2014. “Estimation of Own and Cross Price Elasticities of Alco- Aryo Sahadewo. 2022. “Tobacco Farming: Overcoming an Un- hol Demand in the UK—a Pseudo-Panel Approach Using the derstated Impediment to Comprehensive Tobacco Control.” Living Costs and Food Survey 2001–2009.” Journal of Health Tobacco Control 31 (2): 308–12. https://doi.org/10.1136/tobac- Economics 34 (March): 96–103. https://doi.org/10.1016/j.jhe- cocontrol-2021-056564. aleco.2013.12.006. 126. Li, Virginia C., and Songyuan Tang. 2018. “China’s New Road for 136. Millwood, Iona Y., Robin G. Walters, Xue W. Mei, Yu Guo, Ling Tobacco Control: Tobacco Crop Substitution.” American Journal Yang, Zheng Bian, Derrick A. Bennett, et al. 2019. “Conventional of Public Health 108 (10): 1316–17. https://doi.org/10.2105/ and Genetic Evidence on Alcohol and Vascular Disease Aetiol- ajph.2018.304620. ogy: A Prospective Study of 500 000 Men and Women in Chi- 127. Manthey, Jakob, Syed Ahmed Hassan, Sinclair Carr, Carolin Kil- na.” The Lancet 393 (10183): 1831–42. https://doi.org/10.1016/ ian, Sören Kuitunen-Paul, and Jürgen Rehm. 2021. “What Are S0140-6736(18)31772-0. the Economic Costs to Society Attributable to Alcohol Use? A 137. Mohapatra, Satya, Jayadeep Patra, Svetlana Popova, Amy Systematic Review and Modelling Study.” PharmacoEconomics Duhig, and Jürgen Rehm. 2009. “Social Cost of Heavy Drinking 39 (7): 809–22. https://doi.org/10.1007/s40273-021-01031-8. and Alcohol Dependence in High-Income Countries.” Inter- 128. Manthey, Jakob, Kevin D Shield, Margaret Rylett, Omer S M national Journal of Public Health 55 (3): 149–57. https://doi. Hasan, Charlotte Probst, and Jürgen Rehm. 2019. “Global Al- org/10.1007/s00038-009-0108-9. cohol Exposure between 1990 and 2017 and Forecasts until 138. Muhammad, Andrew, Birgit Meade, David R Marquardt, and COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 7 211 Dariush Mozaffarian. 2019. “Global Patterns in Price Elasticities https://doi.org/10.15288/jsa.1980.41.807. of Sugar-Sweetened Beverage Intake and Potential Effective- 150. Pacula, Rosalie Liccardo. 1998. “Adolescent Alcohol and Marijua- ness of Tax Policy: A Cross-Sectional Study of 164 Countries by na Consumption: Is There Really a Gateway Effect?” NBER Work- Sex, Age and Global-Income Decile.” BMJ Open 9 (8): e026390. ing Paper No. 6348. Cambridge, MA: National Bureau of Econom- https://doi.org/10.1136/bmjopen-2018-026390. ic Research (NBER). https://www.nber.org/papers/w6348. 139. Naiman, A., R. H. Glazier, and R. Moineddin. 2010. “Association of 151. Pan American Health Organization (PAHO). 2021. “Sug- Anti-Smoking Legislation with Rates of Hospital Admission for ar-Sweetened Beverage Taxation in the Region of the Amer- Cardiovascular and Respiratory Conditions.” Canadian Medical icas.” Washington, D.C.: Pan American Health Organization. Association Journal 182 (8): 761–67. https://doi.org/10.1503/ https://doi.org/10.37774/9789275122990. cmaj.091130. 152. Paraje, Guillermo. 2016. “The Effect of Price and Socio-Econom- 140. Nakamura, Ryota, Andrew J. Mirelman, Cristóbal Cuadrado, ic Level on the Consumption of Sugar-Sweetened Beverages Nicolas Silva-Illanes, Jocelyn Dunstan, and Marc Suhrcke. (SSB): The Case of Ecuador.” Edited by Aurora García-Gallego. 2018. “Evaluating the 2014 Sugar-Sweetened Beverage Tax in PLOS ONE 11 (3): e0152260. https://doi.org/10.1371/journal. Chile: An Observational Study in Urban Areas.” Edited by Clau- pone.0152260. dia Langenberg. PLOS Medicine 15 (7): e1002596. https://doi. 153. Paraje, Guillermo R., G. Emmanuel Guindon, and Frank J. Cha- org/10.1371/journal.pmed.1002596. loupka. 2020. “Prices, Alcohol Use Initiation and Heavy Episodic 141. Narain, A., C. S. Kwok, and M. A. Mamas. 2016. “Soft Drinks and Drinking among Chilean Youth.” Addiction, July. https://doi. Sweetened Beverages and the Risk of Cardiovascular Disease org/10.1111/add.15167. and Mortality: A Systematic Review and Meta-Analysis.” Inter- 154. Paraje, Guillermo, and Daniel Araya. 2018. “Relationship be- national Journal of Clinical Practice 70 (10): 791–805. https:// tween Smoking and Health and Education Spending in Chile.” doi.org/10.1111/ijcp.12841. Tobacco Control 27 (5): 560–67. https://doi.org/10.1136/tobac- 142. Nargis, Nigar, A.K.M. Ghulam Hussain, Mark Goodchild, Anne cocontrol-2017-053857. C.K. Quah, and Geoffrey T. Fong. 2020. “Tobacco Industry 155. Paraje, Guillermo, and Fabio S. Gomes. 2022a. “Expenditures Pricing Undermines Tobacco Tax Policy: A Tale from Bangla- on Sugar-Sweetened Beverages in Jamaica and Its Association desh.” Preventive Medicine 132 (March): 105991. https://doi. with Household Budget Allocation.” BMC Public Health 22 (1). org/10.1016/j.ypmed.2020.105991. https://doi.org/10.1186/s12889-022-12959-7. 143. Nargis, Nigar, AKM Ghulam Hussain, Mark Goodchild, Anne CK 156. ———. 2022b. “Expenditures on Sugar-Sweetened Beverages Quah, and Geoffrey T Fong. 2019. “A Decade of Cigarette Taxa- in Jamaica and Its Association with Household Budget Alloca- tion in Bangladesh: Lessons Learnt for Tobacco Control.” Bulletin tion.” BMC Public Health 22 (1). https://doi.org/10.1186/s12889- of the World Health Organization 97 (3): 221–29. https://doi. 022-12959-7. org/10.2471/BLT.18.216135. 157. Paraje, Guillermo, and Pablo Pincheira. 2018. “Asequibilidad 144. Nargis, Nigar, Michal Stoklosa, Ce Shang, and Jeffrey Drope. 2020. de Cerveza Y Bebidas Azucaradas Para 15 Países de América “Price, Income, and Affordability as the Determinants of Tobacco Latina.” Revista Panamericana de Salud Pública 42. https://doi. Consumption: A Practitioner’s Guide to Tobacco Taxation.” Nicotine org/10.26633/rpsp.2018.49. & Tobacco Research, July. https://doi.org/10.1093/ntr/ntaa134. 158. Paraje, Guillermo, Michal Stoklosa, and Evan Blecher. 2022. 145. National Research Council. 2015. “Understanding the U.S. Illicit “Illicit Trade in Tobacco Products: Recent Trends and Com- Tobacco Market: Characteristics, Policy Context, and Lessons ing Challenges.” Tobacco Control 31 (2): 257–62. https://doi. from International Experiences.” Washington, D.C.: The National org/10.1136/tobaccocontrol-2021-056557. Academies Press. https://doi.org/10.17226/19016. 159. Parry, Ian, Sarah West, and Ramanan Laxminarayan. 2009. 146. NBC News. 2004. “Philip Morris, EU Resolve Smuggling “Fiscal and Externality Rationales for Alcohol Policies.” The B.E. Charges.” NBC News, July 9, 2004. https://www.nbcnews.com/ Journal of Economic Analysis & Policy 9 (1): 1–48. https://econ- id/wbna5402238. papers.repec.org/article/bpjbejeap/v_3a9_3ay_3a2009_3ai_ 147. Nelson, Jon P., and John R. Moran. 2019. “Effects of Alcohol 3a1_3an_3a29.htm. Taxation on Prices: A Systematic Review and Meta-Analysis of 160. Pedroza-Tobias, Andrea, Eric Crosbie, Melissa Mialon, Ange- Pass-through Rates.” The B.E. Journal of Economic Analysis & Poli- la Carriedo, and Laura A Schmidt. 2021. “Food and Beverage cy 20 (1). https://doi.org/10.1515/bejeap-2019-0134. Industry Interference in Science and Policy: Efforts to Block 148. Nikaj, Silda, and Frank J. Chaloupka. 2013. “The Effect of Prices Soda Tax Implementation in Mexico and Prevent Internation- on Cigarette Use among Youths in the Global Youth Tobac- al Diffusion.” BMJ Global Health 6 (8): e005662. https://doi. co Survey.” Nicotine & Tobacco Research 16 (Suppl 1): S16–23. org/10.1136/bmjgh-2021-005662. https://doi.org/10.1093/ntr/ntt019. 161. Pell, David, Oliver Mytton, Tarra L. Penney, Adam Briggs, Steven 149. Ornstein, SI. 1980. “Control of Alcohol Consumption through Cummins, Catrin Penn-Jones, Mike Rayner, et al. 2021. “Changes Price Increase.” Journal of Studies on Alcohol 41 (9): 807–18. in Soft Drinks Purchased by British Households Associated with 212 CHAPTER 7 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E the UK Soft Drinks Industry Levy: Controlled Interrupted Time Se- of Smoking on COVID‐19 Severity: A Systematic Review and ries Analysis.” BMJ 372 (372). https://doi.org/10.1136/bmj.n254. Meta‐Analysis.” Journal of Medical Virology 93 (2). https://doi. 162. Petticrew, Mark, Srinivasa Vittal Katikireddi, Cécile Knai, Rebec- org/10.1002/jmv.26389. ca Cassidy, Nason Maani Hessari, James Thomas, and Heide 172. Rehm, Jürgen, Gerhard E. Gmel, Gerrit Gmel, Omer S. M. Hasan, Weishaar. 2017. “‘Nothing Can Be Done until Everything Is Sameer Imtiaz, Svetlana Popova, Charlotte Probst, et al. 2017. Done’: The Use of Complexity Arguments by Food, Beverage, “The Relationship between Different Dimensions of Alcohol Alcohol and Gambling Industries.” Journal of Epidemiology and Use and the Burden of Disease-an Update.” Addiction 112 (6): Community Health 71 (11): 1078–83. https://doi.org/10.1136/ 968–1001. https://doi.org/10.1111/add.13757. jech-2017-209710. 173. Rehm, Jürgen, Colin Mathers, Svetlana Popova, Montarat 163. Pichon-Riviere, A, A Alcaraz, A Palacios, B Rodríguez , L M Rey- Thavorncharoensap, Yot Teerawattananon, and Jayadeep Pa- nales-Shigematsu , and M Pinto. 2020. “The Health and Eco- tra. 2009. “Global Burden of Disease and Injury and Economic nomic Burden of Smoking in 12 Latin American Countries and Cost Attributable to Alcohol Use and Alcohol-Use Disorders.” the Potential Effect of Increasing Tobacco Taxes: An Economic Lancet (London, England) 373 (9682): 2223–33. https://doi. Modelling Study.” The Lancet Global Health 8 (10): e1282–94. org/10.1016/S0140-6736(09)60746-7. https://doi.org/10.1016/S2214-109X(20)30311-9. 174. Rehm, Jürgen, and Maristela Monteiro. 2005. “Alcohol Con- 164. Popkin, Barry M., and Shu Wen Ng. 2021. “Sugar-Sweetened sumption and Burden of Disease in the Americas: Impli- Beverage Taxes: Lessons to Date and the Future of Taxation.” cations for Alcohol Policy.” Revista Panamericana de Salud PLOS Medicine 18 (1): e1003412. https://doi.org/10.1371/jour- Pública 18 (4-5): 241–48. https://doi.org/10.1590/s1020- nal.pmed.1003412. 49892005000900003. 165. Postolovska, Iryna, Rouselle F. Lavado, Gillian Tarr, and Stephane 175. Rhum, Christopher J., Alison Snow Jones, Kerry Anne McGeary, Verguet. 2017. “Estimating the Distributional Impact of Increas- William C. Kerr, Joseph V. Terza, Thomas K. Greenfield, and Ravi ing Taxes on Tobacco Products in Armenia: Results from an S. Pandian. 2012. “What U.S. Data Should Be Used to Measure Extended Cost-Effectiveness Analysis.” Washington, D.C.: World the Price Elasticity of Demand for Alcohol?” Journal of Health Bank. https://doi.org/10.1596/26386. Economics 31 (6): 851–62. https://doi.org/10.1016/j.jheale- 166. Powell, Lisa M, JF Chriqui, T Khan, Roy Wada, and Frank J co.2012.08.002. Chaloupka. 2012. “Assessing the Potential Effectiveness of 176. Risi, Stephan, and Robert N Proctor. 2019. “Big Tobacco Focuses Food and Beverage Taxes and Subsidies for Improving Pub- on the Facts to Hide the Truth: An Algorithmic Exploration of lic Health: A Systematic Review of Prices, Demand and Body Courtroom Tropes and Taboos.” Tobacco Control, September, Weight Outcomes.” Obesity Reviews 14 (2): 110–28. https://doi. tobaccocontrol-2019-054953. https://doi.org/10.1136/tobac- org/10.1111/obr.12002. cocontrol-2019-054953. 167. Powell, Lisa M., Julien Leider, and Pierre Thomas Léger. 2020. 177. Rojas, Christian, and Everett B. Peterson. 2008. “Demand for Dif- “The Impact of the Cook County, IL, Sweetened Beverage Tax ferentiated Products: Price and Advertising Evidence from the on Beverage Prices.” Economics & Human Biology 37 (May): U.S. Beer Market.” International Journal of Industrial Organization 100855. https://doi.org/10.1016/j.ehb.2020.100855. 26 (1): 288–307. https://doi.org/10.1016/j.ijindorg.2006.12.003. 168. Powell, Lisa M., Roy Wada, Joseph J. Persky, and Frank J. Cha- 178. Ross, Hana, Deliana Kostova, Michal Stoklosa, and Maria Leon. loupka. 2014. “Employment Impact of Sugar-Sweetened Bev- 2013. “The Impact of Cigarette Excise Taxes on Smoking Cessa- erage Taxes.” American Journal of Public Health 104 (4): 672–77. tion Rates from 1994 to 2010 in Poland, Russia, and Ukraine.” https://doi.org/10.2105/ajph.2013.301630. Nicotine & Tobacco Research 16 (Suppl 1): S37–43. https://doi. 169. Powell, Lisa, Samantha Marinello, and Julien Leider. 2021. “A org/10.1093/ntr/ntt024. Review and Meta-Analysis of Tax Pass-through of Local Sug- 179. Ross, Hana, Jean Tesche, and Nicole Vellios. 2017. “Undermin- ar-Sweetened Beverage Taxes in the United States.” P3RC Research ing Government Tax Policies: Common Legal Strategies Em- Brief No. 120. Chicago, IL: Policy, Practice, and Prevention Research ployed by the Tobacco Industry in Response to Tobacco Tax In- Center (P3RC), University of Illinois, Chicago. https://p3rc.uic. creases.” Preventive Medicine 105 (December): S19–22. https:// edu/wp-content/uploads/sites/561/2021/09/Rvw-Meta-Anal- doi.org/10.1016/j.ypmed.2017.06.012. Tax-PssThrgh-SSB-Taxes_Rsrch-Brf-No.-120_Jul-2021.pdf. 180. Sargent, Richard P, Robert M Shepard, and Stanton A Glantz. 170. Prince, Martin J, Fan Wu, Yanfei Guo, Luis M Gutierrez Robledo, 2004. “Reduced Incidence of Admissions for Myocardial In- Martin O’Donnell, Richard Sullivan, and Salim Yusuf. 2015. “The farction Associated with Public Smoking Ban: Before and after Burden of Disease in Older People and Implications for Health Study.” BMJ : British Medical Journal 328 (7446): 977–80. https:// Policy and Practice.” The Lancet 385 (9967): 549–62. https://doi. doi.org/10.1136/bmj.38055.715683.55. org/10.1016/s0140-6736(14)61347-7. 181. Scarborough, Peter, Vyas Adhikari, Richard A. Harrington, 171. Reddy, Rohin K., Walton N. Charles, Alexandros Sklavounos, Ahmed Elhussein, Adam Briggs, Mike Rayner, Jean Adams, Ste- Atul Dutt, Paul T. Seed, and Ankur Khajuria. 2020. “The Effect ven Cummins, Tarra Penney, and Martin White. 2020. “Impact COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 7 213 of the Announcement and Implementation of the UK Soft 191. Smith, Katherine E., Emily Savell, and Anna B. Gilmore. 2013. Drinks Industry Levy on Sugar Content, Price, Product Size and “What Is Known about Tobacco Industry Efforts to Influence Number of Available Soft Drinks in the UK, 2015-19: A Con- Tobacco Tax? A Systematic Review of Empirical Studies.” Tobac- trolled Interrupted Time Series Analysis.” PLOS Medicine 17 (2): co Control 22 (2): e1–1. https://doi.org/10.1136/tobaccocon- e1003025. https://doi.org/10.1371/journal.pmed.1003025. trol-2011-050098. 182. Shang, Ce, Frank J. Chaloupka, Prakash C. Gupta, Mangesh S. 192. Smith, Travis A., Biing-Hwan Lin, and Jong-Ying Lee. 2010. “Tax- Pednekar, and Geoffrey T. Fong. 2019. “Association between ing Caloric Sweetened Beverages: Potential Effects on Bever- Tobacco Prices and Smoking Onset: Evidence from the TCP age Consumption, Calorie Intake, and Obesity.” USDA-ERS Eco- India Survey.” Tobacco Control 28 (Suppl 1): s3–8. https://doi. nomic Research Report No. 100. Washington, D.C.: United States org/10.1136/tobaccocontrol-2017-054178. Department of Agriculture (USDA). https://doi.org/10.2139/ 183. Shang, Ce, Frank J. Chaloupka, and Deliana Kostova. 2013. ssrn.2118636. “Who Quits? An Overview of Quitters in Low- and Middle-In- 193. Sornpaisarn, Bundit, Kevin D. Shield, Joanna E. Cohen, Robert come Countries.” Nicotine & Tobacco Research 16 (Suppl 1): Schwartz, and Jürgen Rehm. 2015. “Can Pricing Deter Adoles- S44–55. https://doi.org/10.1093/ntr/ntt179. cents and Young Adults from Starting to Drink: An Analysis of 184. Shang, Ce, Anh Ngo, and Frank J. Chaloupka. 2020. “The Pass- the Effect of Alcohol Taxation on Drinking Initiation among through of Alcohol Excise Taxes to Prices in OECD Countries.” Thai Adolescents and Young Adults.” Journal of Epidemiology The European Journal of Health Economics 21 (6): 855–67. and Global Health 5 (4): S45–57. https://doi.org/10.1016/j. https://doi.org/10.1007/s10198-020-01177-w. jegh.2015.05.004. 185. Sharma, Anurag, Brian Vandenberg, and Bruce Hollingsworth. 194. Srivastava, Preety, Keith R. McLaren, Michael Wohlgenant, and 2014. “Minimum Pricing of Alcohol versus Volumetric Taxation: Xueyan Zhao. 2015. “Disaggregated Econometric Estimation of Which Policy Will Reduce Heavy Consumption without Ad- Consumer Demand Response by Alcoholic Beverage Types.” versely Affecting Light and Moderate Consumers?” PLoS ONE Australian Journal of Agricultural and Resource Economics 59 (3). 9 (1): e80936. https://doi.org/10.1371/journal.pone.0080936. https://ideas.repec.org/a/ags/aareaj/283211.html. 186. Silver, Lynn D, Shu Wen Ng, Suzanne Ryan-Ibarra, Lindsey 195. Stacey, Nicholas, Aviva Tugendhaft, and Karen Hofman. 2017. Smith Taillie, Marta Induni, Donna R Miles, Jennifer M Poti, “Sugary Beverage Taxation in South Africa: Household Expen- and Barry M Popkin. 2017. “Changes in Prices, Sales, Consumer diture, Demand System Elasticities, and Policy Implications.” Spending, and Beverage Consumption One Year after a Tax on Preventive Medicine 105 (December): S26–31. https://doi. Sugar-Sweetened Beverages in Berkeley, California, US: A Be- org/10.1016/j.ypmed.2017.05.026. fore-And-after Study.” PLoS Medicine 14 (4): e1002283. https:// 196. Stockings, Emily, Wayne D Hall, Michael Lynskey, Katherine I doi.org/10.1371/journal.pmed.1002283. Morley, Nicola Reavley, John Strang, George Patton, and Louisa 187. Singh, Gitanjali M., Renata Micha, Shahab Khatibzadeh, Stephen Degenhardt. 2016. “Prevention, Early Intervention, Harm Re- Lim, Majid Ezzati, and Dariush Mozaffarian. 2015. “Estimated duction, and Treatment of Substance Use in Young People.” The Global, Regional, and National Disease Burdens Related to Sug- Lancet Psychiatry 3 (3): 280–96. https://doi.org/10.1016/s2215- ar-Sweetened Beverage Consumption in 2010.” Circulation 132 0366(16)00002-x. (8): 639–66. https://doi.org/10.1161/circulationaha.114.010636. 197. Stoklosa, Michal, Jeffrey Drope, and Frank J. Chaloupka. 2016. 188. Singh, Gitanjali M., Renata Micha, Shahab Khatibzadeh, Peilin “Prices and E-Cigarette Demand: Evidence from the European Shi, Stephen Lim, Kathryn G. Andrews, Rebecca E. Engell, Majid Union.” Nicotine & Tobacco Research: Official Journal of the So- Ezzati, and Dariush Mozaffarian. 2015. “Global, Regional, and ciety for Research on Nicotine and Tobacco 18 (10): 1973–80. National Consumption of Sugar-Sweetened Beverages, Fruit https://doi.org/10.1093/ntr/ntw109. Juices, and Milk: A Systematic Assessment of Beverage Intake 198. Substance Abuse and Mental Health Services Administration in 187 Countries.” Edited by Michael Müller. PLOS ONE 10 (8): (US), and Office of the Surgeon General (US). 2016. Facing Ad- e0124845. https://doi.org/10.1371/journal.pone.0124845. diction in America: The Surgeon General’s Report on Alcohol, Drugs, 189. Slade, Tim, Cath Chapman, Wendy Swift, Katherine Keyes, and Health. Washington, D.C.: US Department of Health and Zoe Tonks, and Maree Teesson. 2016. “Birth Cohort Trends in Human Services. https://pubmed.ncbi.nlm.nih.gov/28252892/. the Global Epidemiology of Alcohol Use and Alcohol-Related 199. Teng, Andrea M., Amanda C. Jones, Anja Mizdrak, Louise Harms in Men and Women: Systematic Review and Metare- Signal, Murat Genç, and Nick Wilson. 2019. “Impact of Sugar‐ gression.” BMJ Open 6 (10): e011827. https://doi.org/10.1136/ Sweetened Beverage Taxes on Purchases and Dietary Intake: bmjopen-2016-011827. Systematic Review and Meta‐Analysis.” Obesity Reviews 20 (9): 190. Smith, Adam. 1776. An Inquiry into the Nature and Causes of the 1187–1204. https://doi.org/10.1111/obr.12868. Wealth of Nations. McMaster University Archive for the History 200. Topiwala, Anya, Klaus P. Ebmeier, Thomas Maullin-Sapey, and of Economic Thought, number smith 1776. https://ideas.repec. Thomas E. Nichols. 2021. “No Safe Level of Alcohol Consump- org/b/hay/hetboo/smith1776.html. tion for Brain Health: Observational Cohort Study of 25,378 214 CHAPTER 7 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E UK Biobank Participants.” MedRxiv Pre-print (May 12, 2021). MacLennan, Shane M Murphy, Elizabeth D Brouwer, Rachel https://doi.org/10.1101/2021.05.10.21256931. A Nugent, Kun Zhao, Prabhat Jha, and Dean T Jamison. 2015. 201. ———. 2022. “Alcohol Consumption and MRI Markers of Brain “The Consequences of Tobacco Tax on Household Health and Structure and Function: Cohort Study of 25,378 UK Biobank Finances in Rich and Poor Smokers in China: An Extended Participants.” NeuroImage: Clinical 35 (January): 103066. https:// Cost-Effectiveness Analysis.” The Lancet Global Health 3 (4): doi.org/10.1016/j.nicl.2022.103066. e206–16. https://doi.org/10.1016/s2214-109x(15)70095-1. 202. Tversky, Amos, and Daniel Kahneman. 1981. “The Framing of De- 212. Villanti, Andrea C., Shelly Naud, Julia C. West, Jennifer L. Pear- cisions and the Psychology of Choice.” Science 211 (4481): 453–58. son, Olivia A. Wackowski, Raymond S. Niaura, Elizabeth Hair, 203. U.S. Department of Health and Human Services. 2014. “The and Jessica M. Rath. 2019. “Prevalence and Correlates of Nic- Health Consequences of Smoking: 50 Years of Progress. A Re- otine and Nicotine Product Perceptions in U.S. Young Adults, port of the Surgeon General.” Atlanta, GA: U.S. Department of 2016.” Addictive Behaviors 98 (November): 106020. https://doi. Health and Human Services, Centers for Disease Control and org/10.1016/j.addbeh.2019.06.009. Prevention, National Center for Chronic Disease Prevention 213. Viscusi, W. Kip, and Joni Hersch. 2008. “The Mortality Cost to and Health Promotion, Office on Smoking and Health. https:// Smokers.” Journal of Health Economics 27 (4): 943–58. https:// www.ncbi.nlm.nih.gov/books/NBK179276/. doi.org/10.1016/j.jhealeco.2008.01.001. 204. U.S. National Cancer Institute, and World Health Organiza- 214. Wada, Roy, Frank J. Chaloupka, Lisa M. Powell, and David tion. 2016. “The Economics of Tobacco and Tobacco Control.” H. Jernigan. 2017. “Employment Impacts of Alcohol Taxes.” National Cancer Institute Tobacco Control Monograph 21. NIH Preventive Medicine 105 (December): S50–55. https://doi. Publication No. 16-CA-8029A. Bethesda, MD: U.S. Department org/10.1016/j.ypmed.2017.08.013. of Health and Human Services, National Institutes of Health, 215. Wagenaar, Alexander C, Matthew J Salois, and Kelli A Kom- National Cancer Institute; and Geneva, Switzerland: World ro. 2009. “Effects of Beverage Alcohol Price and Tax Levels on Health Organization. http://cancercontrol.cancer.gov/brp/ Drinking: A Meta-Analysis of 1003 Estimates from 112 Studies.” tcrb/monographs/21/index.html. Addiction (Abingdon, England) 104 (2): 179–90. https://doi. 205. United Nations. 2022. “Population - Standard Projections.” org/10.1111/j.1360-0443.2008.02438.x. World Population Prospects 2022. United Nations Department 216. Wang, Hong, Jody L. Sindelar, and Susan H. Busch. 2006. “The of Economic and Social Affairs (UNDESA). 2022. https://popu- Impact of Tobacco Expenditure on Household Consumption lation.un.org/wpp/Download/Standard/Population/. Patterns in Rural China.” Social Science & Medicine 62 (6): 1414– 206. United States Public Health Service Office of the Surgeon Gen- 26. https://doi.org/10.1016/j.socscimed.2005.07.032. eral, and National Center for Chronic Disease Prevention and 217. Wang, Meng, Min Yu, Le Fang, and Ru-Ying Hu. 2014. “Associ- Health Promotion (US) Office on Smoking and Health. 2020. ation between Sugar-Sweetened Beverages and Type 2 Dia- “Smoking Cessation: A Report of the Surgeon General.” Wash- betes: A Meta-Analysis.” Journal of Diabetes Investigation 6 (3): ington (DC): US Department of Health and Human Services. 360–66. https://doi.org/10.1111/jdi.12309. https://pubmed.ncbi.nlm.nih.gov/32255575/. 218. World Bank. 2019. “Confronting Illicit Tobacco Trade : A Global 207. Van Walbeek, Cornelis, and Grieve Chelwa. 2021. “The Case Review of Country Experiences.” World Bank Group Global Tobac- for Minimum Unit Prices on Alcohol in South Africa.” South co Control Program. Washington, D.C.: World Bank. https://doc- African Medical Journal 111 (7): 680. https://doi.org/10.7196/ uments1.worldbank.org/curated/en/677451548260528135/ samj.2021.v111i7.15430. pdf/133959-REPL-PUBLIC-6-2-2019-19-59-24-WBGTobaccoIl- 208. Vandenberg, Brian, and Anurag Sharma. 2015. “Are Alcohol licitTradeFINALvweb.pdf. Taxation and Pricing Policies Regressive? Product-Level Effects 219. ———. 2020. “Taxes on Sugar-Sweetened Beverages: In- of a Specific Tax and a Minimum Unit Price for Alcohol.” Alco- ternational Evidence and Experiences.” Washington, D.C.: hol and Alcoholism 51 (4): 493–502. https://doi.org/10.1093/ World Bank. https://thedocs.worldbank.org/en/doc/d9612c- alcalc/agv133. 480991c5408edca33d54e2028a-0390062021/original/ 209. Vartanian, Lenny R., Marlene B. Schwartz, and Kelly D. Brownell. World-Bank-2020-SSB-Taxes-Evidence-and-Experiences.pdf. 2007. “Effects of Soft Drink Consumption on Nutrition and 220. ———. 2023. “World Development Indicators.” World Bank Health: A Systematic Review and Meta-Analysis.” American Open Data. World Bank. 2023. https://data.worldbank.org/. Journal of Public Health 97 (4): 667–75. https://doi.org/10.2105/ 221. World Health Organization (WHO). 2014. “Global Status Re- ajph.2005.083782. port on Alcohol and Health 2014.” Geneva: Switzerland: World 210. Vellios, Nicole, and Corné van Walbeek. 2016. “Determinants of Health Organization. https://www.who.int/publications/i/ Regular Smoking Onset in South Africa Using Duration Analy- item/global-status-report-on-alcohol-and-health-2014. sis.” BMJ Open 6 (7): e011076. https://doi.org/10.1136/bmjop- 222. ———. 2017a. “Tackling NCDs: ‘Best Buys’ and Other Recom- en-2016-011076. mended Interventions for the Prevention and Control of Non- 211. Verguet, Stéphane, Cindy L Gauvreau, Sujata Mishra, Mary communicable Diseases.” Geneva: Switzerland: World Health COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 7 215 Organization. https://www.who.int/publications-detail-redi- 2021: Addressing New and Emerging Products.” Geneva: Swit- rect/WHO-NMH-NVI-17.9. zerland: World Health Organization. https://www.who.int/ 223. ———. 2017b. “Taxes on Sugary Drinks: Why Do It?”Geneva: Swit- publications/i/item/9789240032095. zerland: World Health Organization. https://iris.who.int/bitstream/ 230. ———. 2022. “Sugar-Sweetened Beverage Taxes in the WHO handle/10665/260253/WHO-NMH-PND-16.5Rev.1-eng.pdf. European Region: Success through Lessons Learned and Chal- 224. ———. 2018. Global Status Report on Alcohol and Health 2018. lenges Faced.” Geneva: Switzerland: World Health Organization. Geneva: Switzerland: World Health Organization. https://apps. https://www.who.int/europe/publications/i/item/WHO-EU- who.int/iris/bitstream/handle/10665/274603/978924156563 RO-2022-4781-44544-63081. 9-eng.pdf. 231. Wu, Daphne C., Geordan Shannon, Luz Myriam Reynales-Shi- 225. ———. 2019a. “Tobacco Industry: Decades of Deception and gematsu, Belen Saenz de Miera, Blanca Llorente, and Prabhat Duplicity.”Geneva, Switzerland: World Health Organization. https:// Jha. 2021. “Implications of Household Tobacco and Alcohol applications.emro.who.int/docs/FS-TFI-198-2019-EN.pdf?ua=1. Use on Child Health and Women’s Welfare in Six Low and 226. ———. 2019b. WHO Global Report on Trends in Prevalence of Middle-Income Countries: An Analysis from a Gender Perspec- Tobacco Use 2000-2025. Third Edition. Geneva, Switzerland: tive.” Social Science & Medicine 281 (July): 114102. https://doi. World Health Organization. https://www.who.int/publica- org/10.1016/j.socscimed.2021.114102. tions/i/item/who-global-report-on-trends-in-prevalence-of- 232. Yin, Jiawei, Yalun Zhu, Vasanti Malik, Xiaoqin Li, Xiaolin Peng, tobacco-use-2000-2025-third-edition. Fang Fang Zhang, Zhilei Shan, and Liegang Liu. 2021. “Intake of 227. ———. 2019c. “WHO Report on the Global Tobacco Epidem- Sugar-Sweetened and Low-Calorie Sweetened Beverages and ic, 2019 : Offer Help to Quit Tobacco Use.” Geneva, Switzerland: Risk of Cardiovascular Disease: A Meta-Analysis and Systematic World Health Organization. Review.” Advances in Nutrition 12 (1): 89–101. https://pubmed. 228. ———. 2020. “Alcohol Pricing in the WHO European Region: ncbi.nlm.nih.gov/32696948/. Update Report on the Evidence and Recommended Policy Op- 233. Young, Douglas J., and Agnieszka Bielińska-Kwapisz. 2002. “Al- tions.” Geneva, Switzerland: World Health Organization. cohol Taxes and Beverage Prices.” National Tax Journal 55 (1): 229. ———. 2021. “WHO Report on the Global Tobacco Epidemic 57–73. https://doi.org/10.17310/ntj.2002.1.04. 216 CHAPTER 8 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E Gender and Aging Chapter 8 Gender Gaps in Health and Well-Being of Older Adults: A Review of the Burden of Non-communicable Diseases and Barriers to Health Care for Women and Men Chapter 9 Gendered Responsibilities, Elderly Care, and Labor Supply: Evidence from Four Middle-Income Countries 8 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 8 217 Gender Gaps in Health and Well-Being of Older Adults A review of the burden of non-communicable diseases and barriers to healthcare for women and men Seemeen Saadat a, Meriem Boudjadja a, Sameera Altuwaijri a a Health, Nutrition and Population Global Practice, World Bank 218 CHAPTER 8 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E INTRODUCTION Women and men experience aging differently. Evidence from across the globe suggests that aging is not “gen- der-neutral”, but rather “gender-imbalanced”. Along with physical and physiological differences, social and gen- der norms contribute to different experiences for women and men that manifest across the life cycle (Box 8.1). These differences, such as in education, age at marriage, employment opportunities, as well as caregiving and living arrangements contribute to gender-based differences in socio-economic status as well as health outcomes in people as they age (Carmel 2019; Hosseinpoor et al. 2012; Weber et al. 2019). While aging brings on the onset of non-communica- der gaps in availability of equitable, affordable, qual- ble diseases (NCDs) for both women and men, their ity healthcare services. health outcomes vary due in part to gender-based Health systems play an important role in prevent- differences. Several studies highlight higher male ing and addressing age related disability and disease. mortality at earlier ages due to heart disease or dia- However, health interventions for aging populations betes in high- and middle-income countries, for ex- such as prevention and management of NCDs do not ample. On the other hand, women in these countries necessarily adopt a gender- responsive or transforma- also have similar levels of NCDs but tend to live lon- tive approach. There is a neutrality that assumes that ger with a lower quality of life.1 Women’s longer life women and men have similar needs for healthcare expectancy compared to men means a greater con- and have similar access and capacity to afford health centration of women in older age groups and more services and, especially as they age. Yet, evidence on years lived with disability. Women also go through women’s health and aging acknowledges the  spe- menopause as they age, which takes a physical and cific challenges older women such as increased risk mental toll, and is linked to an increased risk of con- of NCDs during menopause transition or social and tracting non-communicable diseases (Hess et al. financial constraints to access of health services (An- 2012). At the same time, older women, especially in agnostis and Stevenson 2023; Carmel 2019; WHO lower income countries, are less likely than men to 2015); and with the growing proportion of aging pop- receive proper healthcare due to their greater finan- ulations, the prevalence and severity of these issues cial vulnerability – a direct consequence of gender is expected to increase (UN Women 2012). Health gaps over the life cycle that affect women’s patterns systems need to be at the very least responsive to the of labor force participation, income generation op- specific issues related to aging and longer life expec- portunities, and decision-making capacities (Quick, tancies of women and men on the one hand, and their Jay, and Langer 2014; Cotlear 2011). social and economic vulnerabilities on the other such A multitude of factors contribute to differenc- as ability to pay and limited mobility. es in health outcomes of women, men, and other With the progression of aging, older adults also gender identities, and therefore healthcare needs. require long-term care (LTC). While healthcare en- As highlighted above, epigenetic, and social deter- compasses organized delivery of medical care, LTC is minants of health interact with gender (also consid- defined as the “day-to-day help with activities such as ered a determinant of health) to influence knowl- washing and dressing or help with household activ- edge, attitudes, and perceptions about health and ities such as cleaning and cooking” that is likely ac- health seeking behaviors, the ability to access and companied with some type of medical care.2 As with use health services, the capacity to pay for health- healthcare, long-term care policies and programs care, and consequently health outcomes. Gender’s require a gender lens. Gender-based barriers to care interaction with race, ethnicity, or sexual orienta- that influence women’s and men’s health are also re- tion can further exacerbate barriers to health (Cené flected in the access and use of LTC arrangements. et al. 2016; Liu et al. 2019; O’Hanlan et al. 2004). For example, as women tend to live longer than men, Understanding and responding to these issues in they may need LTC services for a longer period of policy and programs is essential for reducing gen- time. However, their capacity to pay for services may 1   On average life expectancy at birth for women is about 5 years higher than that for men, and in MICS the difference is generally about 8 to 10 years. 2   https://www.oecd.org/els/health-systems/long-term-care.htm COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 8 219 BOX 8.1 Gender gaps and the life course As a social determinant of health, gender influences an individual’s way of living, well-being and health, and patterns of seek- ing and receiving healthcare (e.g., Heise et al. 2019; Weber et al. 2019). Both the biological sex* and social constructs of gender influence health outcomes across the life cycle. This is seen in different demographic and health indicators observed in many countries, such as the distorted sex ratios at birth; in gender-based nutrition differentials; in the difference in health seeking behaviors for children based on their gender; in the quality of care; in the high proportion of maternal deaths despite being preventable and the excess high adult mortality among both women and men across different countries. Gender roles are established from an early age and perpetuated across the life course. From the day they are born, children re- ceive signals on what is expected of them and how they should behave based on their gender. These gender norms are passed on by parents and family, teachers and peers, as well as signaling from others in the community (Munoz-Boudet et al. 2013). Gender norms translate into power hierarchies and vulnerabilities, acting as powerful motivators to encourage or prohibit specific individual attitudes and behaviors by sanctioning deviation from these shared expectations. Gender and health are fundamentally connected. Pre-determined gender roles and expectations about behavior create inequities in health and access to health services before birth and continue through the entire life course of a person. Women’s limited em- powerment, for example, in some countries makes it harder for them to access reproductive, maternal and child health services for themselves and their children freely. Similarly, men may deliberately ignore illness or avoid seeking medical help, to avoid being per- ceived as weak by their peers. These patterns influence the health outcomes of women and men and persist over their life course. A life course approach to healthcare supports healthy longevity and can help reduce gender gaps. Healthy longevity refers to the human life span lived to the best possible quality of health that helps to build and preserve human capital, especially as people age (O’Keefe and Haldane 2024). Healthy longevity happens across the life course. While the seeds for healthy longevity are planted even before a person is born, as people age, healthy longevity means avoiding death and serious disability in middle age, en- abling a high level of mental and social functioning through middle and older ages, and a socially connected and reasonably pain-free, short period of time before death. Understanding and addressing gender gaps across the human life course such as reducing early marriages, facilitating use of reproductive and maternal health services, reducing delays in seeking healthcare for female children when sick, or ensuring access to affordable health services and mental health care, especially for vulnerable populations can help to promote healthy longevity for not only for women, but also their children, as well as men, and other gender identities. * The differences in the biological determinants of health and illness include differential genetic vulnerability to illness, reproductive and hormonal factors, and differences in physiological characteristics that can contribute to gender gaps in health over the life-cycle. be more limited than that of men. Moreover, it is of- outcomes for populations aged 45 years and older; ten women who provide LTC, whether through for- (ii) examine the barriers to access and use of health- mal or informal arrangements. Thus, LTC provision care, especially for women aged 45 years and above; can be a source of female employment (when formal and (iii) review the challenges associated with avail- arrangements are in place) or it can pull women from ability of health and long-term care. the labor markets (when informal arrangements are The chapter discusses challenges related to employed). These dynamics make it all the more im- healthy longevity and healthcare from a gendered portant to approach LTC through a gender lens. perspective, especially for older women, and com- plements the discussion on LTC in other studies in Objectives the series. It draws heavily from the evidence base on high income and aging societies for LTC but it’s This study is part of a series of papers on healthy lon- focus is broader and covers, to the extent possible, gevity. It examines the sex disaggregated prevalence gender gaps in health outcomes and health care for and mortality related to four key NCDs for older lower- and middle-income countries. adults and discusses the key barriers to healthcare A life course approach to policy making is im- for older women. The paper aims to: (i) present a de- portant for healthy longevity. So is consideration of scriptive analysis of gender-based gaps in key NCD gender, especially from the point of view of distribu- 220 CHAPTER 8 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E BOX 8.2 The Healthy Longevity Framework and Gender The Healthy Longevity Initiative (HLI) examines the links between human capital formation ((knowledge, skills, and other in- dividual aspects that contribute to productivity) and its maintenance during adulthood and as people age. Gender is an im- portant aspect of human capital formation, affecting health, education and employment opportunities and outcomes. The influence of gender is clear within the context of the HLI framework (O’Keefe and Haldane 2024): • Decisions or choices made for girls and boys earlier in life, and across the life course influence health outcomes in older ages. Early marriage and childbirth, for example, affects women’s health and life opportunities for the rest of their lives, and child- hood health and nutrition influence adult health outcomes. • As populations age, human capital accumulation increases at a decreasing rate and it starts to deplete among the very el- derly, especially because of a decline in their health. Women and men can experience differences in depreciation of human capital due to a combination of health and social factors. tional impacts. Cross-sector interventions, including er documents dating back to 1990 when these were those that target child and adolescent health, improve relevant to the topic at hand. The data analysis uses access to life-saving medicines and technologies, and the Global Burden of Disease (IHME) database and promote healthy behaviors contribute to better health World Development Indicators. across the life cycle, including older age cohorts. Inter- Literature was identified using string combina- ventions that reduce gender gaps in healthy longevi- tions of key words (Table 8.1). Databases searched ty benefits both men and women. Affording women included pub med, google scholar, google, and the and girls access to reproductive and maternal health, World Bank’s “imagebank” and “open repository” for example, supports better reproductive health (e.g., databases. In addition, a search request was also sub- in managing menstruation, prevention of STIs) and mitted to the World Bank Library search services on pregnancy outcomes for mother and child, and reduc- “menopause”. Menopause (and reproductive health), es the burden of reproductive disability (e.g., uterine while not covered in this study, were used as identifier prolapse or fistula, or endometriosis) with long term terms for papers on health issues among middle-aged benefits, especially as women age. Similarly, such an women who are not of reproductive age. The string approach can contribute to reducing the burden of “women, aging, and health” provided over 10,000 re- early mortality due to NCDs, especially among men. sults because of the width of topics covered under it (epidemiology, supply, and demand). The search was Methodology refined with additional terms (as listed above. On the demand side, this still yielded over 500 articles. Article The study is based on a literature review of demand titles and abstracts were then reviewed, and 91 were and supply issues combined with secondary data selected as relevant to the topic at hand with about analysis to build a profile of health outcomes and half used in the final write up. On the supply side, an barriers in access to care. The literature review is initial 245 papers were identified and reviewed, with based on peer reviewed journal articles, white pa- roughly 65 included in the final analysis. Seminal re- pers, and published reports on the topic for the time ports and other supporting evidence, when identified period 2006-2023, with the inclusion of some earli- were also included in the review. TABLE 8.1  Key search teams for the review (not an exhaustive list) Aging Health/healthcare Access Gender Elderly Reproductive health Affordability Caregivers Women Demand for healthcare Quality Gender differences Men Utilization of healthcare Mobility Female Supply side care Menopause Long-term care/LTC Informal care giving COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 8 221 For the quantitative review, the study focuses on Structure of the review women and men aged 45 years and above. We use the Global Burden of Disease for health data and the The structure for this review follows the interactions World Development Indicators database for data on individuals, and in particular, women may have income for the year 2019. We examine the propor- with the healthcare system (Figure 8.1). The review tion of deaths attributable to a disease, disease prev- focuses on four main areas that overlap health sec- alence, and years of life lived with disability (YLD) tor demand and supply issues: due to disease in three age groups: 45-59 years, 60-79 years, and 80 years and above.34 As most women ex- a. Women’s health needs as they age focusing on perience menopause during age 45-59 years, we also gender differences in mortality and prevalence of include this age group. The data analysis examines non-communicable diseases (NCDs) in women gender gaps in four major NCDs affecting older age and men aged 45 years and above. groups: cardio-vascular disease, diabetes, cancers, and mental health. For mental health outcomes, b. Women’s access to and utilization of healthcare, we consider prevalence and YLD due to depressive including informal and long-term care. This also disorders as well as self-harm since they are both covers issues related to affordability such as the ef- connected to mental health and have considerably fects of out-of-pocket expenditures or insurance different outcomes for women and men. on demand for services. FIGURE 8.1 Structure and Linkages for Women and Aging in Health Review 3   The authors acknowledge the limitations of the GBD dataset. However, for the purpose of this study, the protocols employed for data collec- tion and analysis are sufficient and the dataset offers more complete information than other sources alone. 4   YLD is a widely used measure to convey the impact of disease on the quality of life prior to being resolved or death from the disease. While the methods used to calculate YLD have some limitations, it is a useful way of conveying the negative impact or the non-monetary cost of living with a disease. In most cases in this study, the YLD follows a similar pattern to the prevalence data and are shown only to strengthen the message that higher prevalence is linked to greater burden of disease. 222 CHAPTER 8 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E c. Provision of health and long-term care services cio-economic status (which in turn also influences for aging women including availability and qual- accessibility and affordability). Access to services de- ity of services covering both formal and infor- pends on factors such as availability of services (sup- mal care; and ply), ease of physical access/transport, individuals’ mobility, and their awareness of said services. Wheth- d. Women as providers of care. This focuses espe- er available services are used depends on socio-eco- cially on the role women play as informal care nomic factors such as wealth and education, cultural givers in the context of long-term care. norms, as well as quality of services. At the same time, availability of services and quality of services are inter- Demand for health, and long-term care services, is linked and can affect demand for these services. The associated with individuals’ health status, their need framework looks these issues from a gendered per- for services, their ability to afford it, and the ease with spective, focusing on aging women and men. It helps which they can access and use these. An individual’s to identify where the gaps are in terms of demand and health status is linked to many factors including their supply of services, why these exist, and consequently health and nutrition across the lifecycle, and their so- how to address them (London 2008; WHO 2009). GENDER DIFFERENTIATED HEALTH OUTCOMES OF AGING POPULATIONS Men and women experience aging differently due to biological and socio-economic reasons. Women for example, have a longer life expectancy at birth compared to men which can increase their chances of living with illness or dis- ability over longer periods. For instance, there are a greater number of blind women than men, and women bear a higher burden of Alzheimer’s disease compared to men because of their longer life spans (Dubal 2020; WHO 2007). Moreover, women’s reproductive cycle has a significant impact on their health and well-being (e.g., physiological changes due to childbirth or menopause) compared to men’s. Socio-economic inequalities and gender norms that prevent women from generating income or accumulating savings, especially in lower income countries can adverse- ly influence women’s health further through restricting their access and use of health services, especially in LMICs (e.g., Dhar 2001; Prus and Gee 2003 ; Szanton et al. 2008; Zhan 2005). In addition, women’s limited mobility and dependence on other family members, particularly males, restricts their access to timely care. On the other hand, gender norms can also create barriers for men to access timely healthcare, which can contribute to worsening health or early and avoidable mortality (Etienne 2018; McGraw, White, and Russell-Bennett 2021; Novak et al. 2019). Examining data on major NCDS for women and deaths among women) than men, with the exception men aged 45 and above shows across countries of South Asia, which is likely driven by the larger shows some common threads for women and men. proportion of CVD mortality among males in India, These are discussed below: especially for those under age 60 years.5 When directly comparing CVD deaths for Cardiovascular diseases women and men, the picture is more mixed. Figure 3a and 3b show the ratio of female to male deaths due Cardiovascular diseases (CVD) are the leading cause to CVD by income groups and at the regional level. of mortality, responsible for roughly 18 million As Figure 3a shows, in high- and low- income groups deaths each year (WHO 2021). This is especially the the ratio of female to male burden of mortality due case for older adults, both women and men: in 2019, to CVD is higher, but for middle income groups, a CVD accounted for 41.6 percent and 38.2 percent greater proportion of men than women are reported of all deaths globally among women and men aged to have died due to CVD. When disaggregated by re- 45 years and above respectively (IHME 2020). The gions (Figure 3b), the ratio of female to male deaths largest burden of CVD deaths is in upper-middle in- due to CVD is especially high for Eastern and Cen- come countries, particularly in the Middle East fol- tral Europe and sub-Saharan Africa for those aged 45 lowed by Europe in Central Asia (Figure 8.2: 2a and years and older, while it is at parity in North Amer- 2b), with proportionally larger number of deaths ica. In other regions, at the aggregate level, the ratio due to CVD among women (when compared to all indicates a higher burden on men (IHME 2020). 5   IHME 2020. Global Burden of Disease database (accessed 11-16-23) COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 8 223 FIGURE 8.2 Cardiovascular Mortality among older adults by regional and income aggregates Source: IHME Global Burden of Disease 2020 Globally, over 500 million people are estimated North Africa regions. In the Middle East and North to live with cardio-vascular diseases. This represents Africa and South Asia regions, prevalence is high- a nearly a fifth of the total burden of all disease prev- er for men than women, with roughly 8 women for alence among older adults (ages 45 years and above), every 10 men living with CVD (ages 45 years and with about 11 women for every 10 men living with above). On the other hand, in the East Asia and Pacif- CVD. The largest burden is in higher income coun- ic, there are about 13 women for every 10 men, and tries and the North America region followed by East- in Eastern Europe and Central Asia about 12 women ern Europe and Central Asia and the Middle East and for every 10 men live with CVD (IHME 2020). FIGURE 8.3 Ratio of female to male CVD mortality by regional and income aggregates Source: IHME Global Burden of Disease 2020 224 CHAPTER 8 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E When data is disaggregated further by age 45-59 years, gender gaps are minimal, but there are groups, the story gets more interesting. Figure a handful of countries, particularly in the MENA re- 8.4 presents the breakdown by three indicators on gion where the differences between women and men mortality, prevalence, and years lived with disabili- in terms of CVD mortality are stark, and significant- ty (YLD) for CVD by sex (with blue for female and ly impact male mortality – the outliers in figure 8.4; orange for male) and age groups at the country lev- and (3) as people age, the burden of CVD prevalence el6. The data indicate three things: (1) patterns of and deaths is greater among women, particularly for CVD prevalence, mortality, and YLD for women the oldest age group. and men are fairly similar; (2) for the younger ages FIGURE 8.4 Cardio-Vascular Disease by Sex for Older Populations (ages 45 years and above) Source: IHME Global Burden of Disease 2020; World Development Indicators (WDI) Database 6   Disaggregation by country allows to see the variations within regions which in some cases are driven by a handful of countries. COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 8 225 The figures on YLD indicate a similar pattern as elderly Canadians in Alberta. They found that these prevalence. However, the gender gap in YLD factors significantly affected women’s access to timely begins to enlarge during ages 60 - 79 years, with care (measured as receiving cardiac catheterization) women, especially in lower- and middle- income but not so for men. They conclude that special atten- countries exhibiting higher YLD. tion needs to be paid towards improving the quality Studies on CVD and gender align with this evi- of care for women, especially those in lower income dence and indicate the increasing burden of disease groups. The Lancet Women and Cardiovascular Dis- on women. For example, Thielke and Diehr (2012) ease Commission also underscored these issues and used data on adults aged 65 years and above from recommended a 10-point agenda that promotes bet- the US Cardiovascular Study (n=5888) to under- ter data collection and research, inclusion of women stand the probability of transitioning from one state in clinical trials, training of healthcare personnel, and of health to the next or dying. They found that when heart health services and programs aimed at women men were sick, they died more often than women in in different settings, in poor socio-economic settings a similar state, but that they also remained healthi- (Vogel et al. 2021). er than women while alive and were more likely to recover from being sick. While earlier studies found Diabetes mellitus that heart disease was more prevalent among women at older ages than men (Maas and Appelman 2010), Gender based gaps are also clearly visible in the recent Lancet Women and Cardiovascular Dis- diabetes outcomes. Diabetes is among the top ten ease Commission indicates that it is also increasing NCDs globally, accounting for roughly 1.55 million among younger women (Vogel et al. 2021). deaths of which 95 percent were among persons aged Moreover, women are less likely to receive ap- 45 years and above in 2019 (IHME 2020). Aggre- propriate diagnosis or care because of misconcep- gate level data show that slightly more women than tions and poorer understanding of the incidence of men die due to diabetes among the 45 and older age disease among women, women’s symptoms of heart groups (3.6 percent vs. 2.9 percent). When com- disease, and other socio-economic and gender-based paring the burden of diabetes proportional to other barriers to care (Papadopoulou and Kaski 2013; Pe- causes of deaths by sex, the burden is higher among ters and Woodward 2022; Vogel et al. 2021). Fabreau women than men. This is evident among different et al (2015) for example, examined the relationship country income groups and across all regions except between sex, neighborhood, and other socio-eco- for North America (Figure 8.5 – 5a and 5b). The larg- nomic factors on acute coronary syndromes among est burden of deaths from diabetes is in lower mid- FIGURE 8.5 Deaths due to Diabetes, Ages 45+ (by Income Groups and Regional Aggregates) Source: IHME Global Burden of Disease 2020 226 CHAPTER 8 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E FIGURE 8.6 Ratio of female to male mortality due to diabetes by regional and income aggregates Source: IHME Global Burden of Disease 2020 dle-income countries (4.7 percent for women and ery 10 men living with diabetes). When aggregated by 3.9 percent for men), while the largest gender gap is income groups, a similar gap exists for high income in upper-middle income countries (3.4 percent for and lower-middle income countries. women and 2.4 percent for men). At the regional lev- However, when further disaggregated by age, el, the largest gender gap is in the Latin America and the data show growing gender gaps across countries Caribbean region (7.1 percent for women; 5.7 per- – a patten similar to that for CVD outcomes. Figure cent for men) followed by the Middle East and North 8.7 presents the data on deaths, prevalence, and years Africa (4.8 percent for women; 3.6 percent for men). lived with disability due to diabetes at the country A similar pattern emerges when directly com- level for three age groups: 45-59, 60-79, and 80+ paring male and female deaths due to diabetes at years. For each outcome, countries follow a similar the aggregate level. Figure 8.6 presents this com- trend. For ages 45- 59 years, the burden of disease is parison. Proportionally, nearly 12 women for every mostly similar for women and men, with a few outli- 10 men above the age of 45 die due to diabetes in ers. However, for those between ages 60 and 79 years, upper-middle income countries (Figure 6a). In oth- a shift begins to happen with gender gaps growing er income groups the gaps are not as large and in between women and men. While in the oldest age low-income countries are reversed. At the regional group of 80 years and above there is a clear shift to- level (Figure 6b), the largest gaps are in the East- wards a greater burden of disease for women. ern Europe and Central Asia region with about 13 These gaps represent the growing vulnerability women for every 10 men over the age of 45 dying of women as they age and compound their poorer due to diabetes. In Sub-Saharan Africa and North health outcomes. For instance, older women with di- America, the gap is reversed with about 11 men for abetes have been found to have higher risk of CVD every 10 women dying from diabetes. mortality compared to men (Clemens et al., 2020; Nearly 460 million people live with diabetes glob- Juutilainen et al., 2004; Pan et al., 2011) or lose any ally of which 80 percent are aged 45 years or above. Di- biological advantage they may have compared to abetes accounts for about 16 percent of all the disease men in developing a risk of CVD (Wannamethee et prevalence among older adults, with about 97 women al. 2012). Recent evidence on COVID-19 outcomes for every 100 men living with diabetes. This is mainly also suggests that older women’s biological advantage driven by a higher prevalence of diabetes among men for survival compared to men disappeared if they in North America, South Asia and Middle East and had multimorbidity such as diabetes or heart disease North Africa regions (roughly about 9 women for ev- (Kautzky-Willer, 2021). Moreover, in addition to risk COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 8 227 FIGURE 8.7 Gender Gaps in Diabetes for Older Populations (ages 45 years and above) Source: IHME Global Burden of Disease 2020; World Development Indicators (WDI) Database factors such as obesity and genetics that can lead to tion to men, but when this is no longer the case, they diabetes in both women and men, women face an may feel lonely, isolated, or a burden. This - coupled additional risk from gestational diabetes which can with loss of a partner and poorer social networks predispose them to developing diabetes after birth or (compared to women) this can contribute to poor later in life. Women with diabetes are more likely to mental and physical health. In their study of diabe- experience poor physical function due to lower bone tes patients in Germany, Kramer et al (2012), found strength (Nillson et al 2016). One recent qualitative significant differences in the utilization of healthcare examination of health-related quality of life for per- between elderly women and men, with men having sons with diabetes in Spain, for example, also found fewer visits and prescriptions which the authors as- that women had a lower quality of life than men, and cribe to socio-economic characteristics, and perhaps older women (over 75 years of age) had the worst an over-utilization of health services by women. It quality of life (Gálvez Galán et al., 2021). may also be an under-utilization of services by men Socio-economic conditions and gender norms as highlighted in other literature on male behavior, at- also contribute to these gender gaps that also affect titudes about masculinity, and use of health services. men. For example, in their study in India, Perkins et In addition, challenges such as financial or mo- al (2016) found that while women’s health was poor- bility barriers can prevent access to timely care. An ex- er than men’s in general, widowed men had a higher amination of expenditure data between 2002 and 2011 likelihood of developing diabetes within the first 4 for diabetes treatment in the United States showed that years of their wife’s death. Social and gender norms women paid significantly and consistently more in are likely to influence behaviors. For instance, older out-of-pocket payments for prescriptions and office as men are more likely to have poor mental health es- well as home-based care when compared to men (Wil- pecially after retirement. Gender norms prevalent in liams et al. 2017). While majority of the population is most of the world place the task of income genera- covered by health insurance in the United States, in 228 CHAPTER 8 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E a lower-middle-income country such as Bangladesh, gender gaps are also narrower, with slightly greater this can be cost prohibitive where payments are out of mortality among women as a proportion of all causes pocket and the annual cost of diabetes care is roughly of mortality for women compared to men (also Fig- around US$ 865/- (Afroz et al. 2019). ures 8a and 8b). While research on the determinants of gender differences in cancer related outcomes in Cancers low and lower-income countries is lacking, a larger share of other diseases, including maternal mortality, Globally, an estimated 10 million people die of can- as well as known barriers to access to healthcare for cers each year (WHO 2022). The myriad of cancers women such as availability of cancer screenings and affect men and women differently. Some are unique treatment services, lack of information, limited mo- to women (e.g., breast cancer) or men (e.g., prostate bility and financial constraints that create barriers and cancer) while others affect both (e.g., lung, skin, thy- may account for reversed gender gaps in these regions roid) but to different extents based on genetic, envi- (Black, Hyslop, and Richmond, 2019; Getachew et al. ronmental, and behavioral factors (Dorak & Karpu- 2019; Mascara & Constantinou 2021; Okoronkwo et zoglu, 2012; Hatano et al. 2022; Lipsky et al. 2021). al. 2015; Rivera-Franco & Leon-Rodriguez, 2018). Unlike CVD and diabetes, evidence suggests When comparing the number of deaths for that overall, mortality from cancers is higher for men women to that of men directly at the global level, than women among older age groups. Figure 8.8 there are 134 men for every 100 women above the age presents the data on cancer mortality as a percent- of 45 years that die from cancers (IHME 2020). Fig- age of all-cause mortality among women and men ure 8.9 shows the breakdown of gender gaps by re- ages 45 years and above. Across regions, the largest gion and income group when comparing the deaths gender gaps are in Eastern Europe and Central Asia due to cancers among women and men in absolute (ECA) and the East Asia, and the Pacific (EAP) re- numbers, which shows near parity in high income gions followed by North America (Figure 8a). This is countries where in other income groups, gender gaps also reflected in Figure 8b which shows that across increase with a greater burden of deaths for men. all cancers for ages 45 and above the largest gender When comparing geographical regions, gender gaps gaps are in high- and upper-income countries. are most significant in East Asia and the Pacific and However, in low- and lower-middle income the Middle East and North Africa regions, where for countries (mainly in Sub-Saharan Africa and South every 100 women there are roughly 162 and 142 men Asia), where the burden of cancer mortality is lower, respectively that also die from cancers. FIGURE 8.8 Deaths due to Cancers, Ages 45+ (by Income Groups and Regional Aggregates) Source: IHME Global Burden of Disease 2020 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 8 229 FIGURE 8.9 Ratio of female to male mortality due to cancers by regional and income aggregates Source: IHME Global Burden of Disease 2020 Interestingly, when the data is further disag- rates among women reach similar levels as that of gregated to the country level by sex and age groups, men, particularly in upper-middle income countries. a consistent cross-country pattern emerges. The Female cancers make up nearly 60 percent of the burden of cancer related mortality is relatively cancers prevalent among women aged 45 and above. low, and similar for women and men aged 45 to 59 This includes breast, uterine, ovarian, and cervical can- years, especially for LMICs. Gender gaps appear in cers, with breast cancer being the most prevalent (44.7 mortality from cancers in the 60 to 79 age group, percent) and leading cause of mortality (27 percent) as cancer mortality increases (Figure 8.10), with a from cancer among women (IHME 2020). By far the larger proportion of death among men compared to largest burden of these cancers is in the 45-59 years age women across all countries. However, for the old- group (68 percent) i.e., towards the end of woman’s re- est age group (80 + years) the gaps close again, but productive cycle. The prevalence drops to 45 percent mortality (as expected) is especially high for all but for those ages 80 years and above. On the other hand, low-income countries, likely because of the lower cancers such as prostate and testicular cancers make (albeit growing) burden of NCDs and a lower life up around 30 percent of the cancers prevalent among expectancy in low-income countries. men over 45 years, with the largest prevalence among On the other hand, global prevalence of can- men 60-79 years (IHME 2020). This includes breast cers and years lived with disability (YLD) presents a cancer, which although rare among men, can occur. slightly different pattern, but again with some consis- Annex 8.1 presents the breakdown of mortality and tency in gender outcomes across three age groups of prevalence of cancers by type among women and men. older adults. Globally, 103 women for every 100 men Gender gaps in prevalence and mortality for live with cancer. Gender gaps exist among the 45 to cancers that are common to both sexes are complex. 59 years age group with clearly higher prevalence and Bronchial, Tracheal, and Lung cancers together are YLD among women compared to men (Figure 8.10, the leading cause of cancer-related mortality for men last column). These gaps tend to close for older pop- (25.5 percent of total cancers among men aged 45 ulations, particularly among LMICs. In upper-mid- years and above) but responsible for only 9.6 percent dle income countries men begin to exhibit higher of cancers among women in the same cohort (IHME prevalence and years lived with disability compared 2020). Similarly, 2 men for every 1 woman have oral to women among the 60 to 79 years age group. How- cancer (Lipsky et al. 2021). Men’s more frequent use of ever, for those aged 80 years and above, gender gaps tobacco and alcohol, poorer oral hygiene, and delayed not only minimize but also reverse as the prevalence care-seeking compared to women as well as higher 230 CHAPTER 8 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E FIGURE 8.10 Gender Gaps in Cancers among Older Populations (ages 45 and above) Source: IHME Global Burden of Disease 2020; World Development Indicators (WDI) Database likelihood of multimorbidity are well-documented were less likely to have a partner (70% versus 83%, risk factors behind these outcomes (Dorak & Kar- p < 0.01) (Oertelt-Prigione et al. 2021). Hatano et al. puzoglu, 2012; Jha 2020). On the other hand, wom- 2022 highlight the differences in end-of-life experi- en who smoke are 3 times more likely to develop it ences for women and men in a sample of patients in than men who smoke (Barta, Powell, and Wisnivesky palliative care in East Asian countries (n=1632). Their 2019). Thyroid cancers are also more prevalent among findings show that men were more likely to have mul- women than men (2.2 vs. 1.1 percent of all cancers for tiple diseases and experience illness such as bronchitis those ages 45 years and above), which may be linked and fever compared to women (43% vs 36%, p=0.05; to post-menopausal changes in women’s metabolism 42% vs 30%, p<0.01), while women stayed longer in and weight gain (Cui et al. 2021; Dorak & Karpuzog- palliative care units than men (28 days vs 21 days, lu, 2012; Rahbari et al. 2010; Suteau et al. 2021). p<0.01). Understanding these types of differences can Men’s and women’s lived experience with cancer help to design more responsive/effective policies and can also be different. A study that looked at sex dif- interventions to provide care for women and men. ferences in the prevalence of physical and emotional symptoms as well as loss of functioning in 5339 can- Mental health and self harm cer survivors in the Dutch PROFILES registry, iden- tified gender gaps with female survivors experiencing One in five persons globally suffers from a mental greater net loss in physical and cognitive functioning health issue. Among these, depressive disorders are and male survivors with greater net loss in role and the most common and merit special attention among social functioning when compared to an age and sex aging populations. As figure 8.11 shows, prevalence matched reference population. Female patients in the levels of depression and the years lived with disabil- cohort were also less likely to be employed (20% vs ity (YLD) due to depressive disorders are consis- 24%, p < 0.01), received less formal education; and tently higher for women compared to men. This gap COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 8 231 persists across all countries, for all ages 45 years and women in these countries, as well as the prevalence above and income groups. While there are variations in rates globally. However, these decline for older age these gaps, in some countries such as Viet Nam, groups and are among the lowest for ages 80 years Ukraine, and Pakistan the female burden of major and above. On the other hand, among those aged depressive disorders is up to double that of men, and 80 years and above, Japanese women have the these gaps persist in older age groups. Two outliers to highest rates of depressive disorders compared to this overall trend are United Arab Emirates and Qatar, other countries, and this is roughly three times that where for ages 45 to 59 years, the prevalence of of depression among men in this age group in depressive disorders is considerably higher than that of Japan (13.17% vs. 4.96%)- (IHME 2020). FIGURE 8.11 Gender Gaps in Depressive Disorders among Older Adults Source: IHME 2020; WDI While women tend to have higher rate of depres- women, although the gaps begin to close those over sive disorders than men, the rate of deaths due to 80 years of age in lower income countries (Figure self-harm among older men is higher. As a proxy for 8.12, first column, bottom graph). In some cases, par- deaths related to depressive disorders, male deaths ticularly in Middle Eastern countries, the mortality due to self-harm significantly outpace those among from self-harm among men is extremely high, even 232 CHAPTER 8 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E higher than that of males in other countries (Figure harm (e.g., cutting or burning of one’s own skin) 8.12). This is likely because of more violent methods is equally prevalent among older women and men, of self-harm that men employ which are more likely with some exceptions (Figure 8.12, middle column), to result in death than women. As epidemiological and in fact increases for women ages 80 years and studies on middle eastern (and other) countries point above. The years lived with disability (Figure 8.12, to higher numbers of suicides attempts by women (El last column) follow a similar pattern as prevalence, Halabi 2020; Mergl et al. 2015). except for those aged 60-79 years where for middle Self-harm takes various forms, and not all end income countries, the years lived with disability for in death. Data also show that prevalence of self- women are slightly higher than those of men. FIGURE 8.12 Gender gaps in self-harm among older adults (ages 45 and above) Source: IHME 2020; WDI COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 8 233 Gender norms likely contribute to how wom- ute to unequal power dynamics and inequalities en and men approach mental health and seeking such as limited control over finances or limited de- support to address their depression, or stress. The cision-making power, often beginning during ado- roles men and women engender during their life- lescence, can lead to depression (Bone, Lewis, and cycles and the social support systems they can ac- Lewis 2020; Rungreangkulkij et al. 2019). cess, and use contribute to their mental health and In a similar vein, the negative effect of male well-being (Adams et al. 2015; Cheung and Mui perceptions on their value may be linked to men’s 2021; Mudege and Ezeh 2009; Obuobi-Donkor, greater propensity for self-harm leading to suicide Nkire, and Agyapong 2021; Treviño-Siller et al. (Rivera and Scholar 2020). For example, results from 2006). A recent study in India, for example, found the Australian Longitudinal Study on Male Health that 56 percent of older adults who were widowed (Ten to Men, n = 13,884) found that “self-reliance” and living alone were significantly more likely to suf- was the dominant dimension of masculinity linked fer from depression [AOR: 1.56; CI 1.28,1.91] com- to suicidal ideation (AOR 1.34; 95% CI 1.26–1.43) pared to those who were married and living with among men aged 18 to 55 years (Pirkis et al. 2017). others (Srivastava et al. 2021). In often cases, this For older adults, after retirement or loss of job men is women (Perkins et al. 2016; Pilania et al. 2019) are more likely to feel isolated and less valuable as and is a global experience (Beutel et al. 2018; Keily, members of society. Earlier work by Dhar (2001) Brady, and Byles 2019; Lin et al. 2021). in India found that social and familial support Although genetics can predispose people to de- was linked to better health, especially among men. pressive disorders, women’s lived experiences also Kendig et al. (2014) find that Australian men per- contribute to their higher rates of the disease. One ceived low social activity as linked to aging poorly. in three women for example have experienced inti- Some studies suggest that as women’s roles tend to mate partner violence in their lives (WHO 2021b). be more complex than men’s, and go beyond the Exposure to rape has been found to contribute to labor markets, their ties within their communities the development of psychological problems, such as can be beneficial in overcoming social isolation. depression, dissociation, substance abuse, feelings While women face persistently greater barriers in of guilt and shame, and suicidal ideation (Covers access to healthcare and their quality of life is often et al. 2021). Childbirth can also contribute to de- lower than that of men (Carmel 2019; Griffin, Loh, pression in women and girls. One study found that and Hesketh 2012), their ability to cope with social new mothers of low socioeconomic status were 11 and individual changes that come with aging may times more likely to develop postpartum depression an important factor in their lower rates of suicides symptoms than women of higher socioeconomic compared to men (Carlo et al. 2019). status (Carberg 2019). Gender norms that contrib- AGING, GENDER, AND DEMAND FOR SERVICES Older women’s access to healthcare services Access to health services refers to the ease with which an individual can get the healthcare they need. On the demand side, a key question is how the distributional effects of factors such as income, mobility, distance, and availability of services affect access to health services for aging populations, and especially women. As people age, their reliance on others to help care for them is expected to increase, especially if they have a disability or their health deteriorates. In addition, they may depend on other household members for financial support or transport, especially in the case of women. One of the ways to estimate access to healthcare (Bertakis et al. 2000), others find just the opposite is through delays in health seeking behavior. Evi- results (Cameron et al. 2010; Song and Bian 2014; dence on gender differences in seeking healthcare Azad et al. 2020). All evidence however points to among aging populations is mixed. While some women’s lower economic status and higher costs for studies find evidence of higher utilization of health healthcare (Azad et al. 2020; Bertakis et al. 2000; services, most often in higher income countries Cameron et al. 2010). Several studies focusing on 234 CHAPTER 8 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E higher income countries highlight wealth/income At the same time, women’s burden of care influ- as having a greater role in determining delays to ac- ences their own access to, and utilization of health cessing and using healthcare than gender. Relatively services. For instance, as Albertini (2016) points lower gender inequality, health insurance coverage, out, adult women are the risk of becoming “sand- and longer employment histories may contribute wiched” between caring for the young and the el- to this. However, there are other disparities that derly. Using data from three waves of the SHARES can be overlooked (Cameron et al. 2010). For ex- survey in Europe (n=79359) the study also finds, ample, even though in their study on older adults rather counter-intuitively, that being more educated in Japan, Murata et al (2010), find that income and and having higher market incomes increases the risk time are more relevant to health seeking behavior of women to be “caught in between the care needs than gender, they also report that significantly more of the younger and the older family members” and men had a regular source of health care compared leaving the labor market. Certainly, evidence from to women (~ 23% compared to 17%). This is likely Eastern Europe highlights the social pressure on because of more regular employment (e.g., women women to fulfill this role (Bussulo, Koettl, and Sin- may step out of the labor market for childbirth of not 2015). This is also highlighted in evidence from care provision). It may then be argued that gender East Asia, where the “provision of family-based care differences in income or wealth then do tend to play [by women] is associated with fewer working hours a role in how women and men seek and use health and a higher probability of exit from the workforce”. services differently. Income inequalities across the Similarly, in China, 28 percent of people aged 45 + life course can compound over time, and in the ab- years (mainly women) provide care for grandchil- sence of equitable social protection, making it more dren, and 14% provide care for the elderly (World costly to access healthcare for older women. Bank 2016). Literature on women’s access to healthcare fo- Often as people age, their need for reliance on cuses mainly on younger populations or minority family and other sources of support increases. In groups. Few studies examine the pathways and bar- low- and lower-middle income countries, it is not riers to access for older women beyond financial is- uncommon for elderly to live with their adult chil- sues. Transportation, or the lack of it, is one of the dren and their families or to rely on them financial- key challenges for women in general, and especially ly. This is especially the case for older and widowed older women. Women’s experiences point to multi- women, who may rely on their families for residen- ple transportation challenges such as availability of tial, financial, or other support. Due to women’s suitable transport, especially in rural areas, cost of lower labor force participation, gender-based pay the transport, safety, or poor infrastructure making gaps, and often limited autonomy, they are relatively travel difficult (Hamiduzzaman et al. 2017; Naud et more likely to be dependent on family for their res- al. 2019). Women’s care giving responsibilities for idence, mobility, and accessing the health sector or example can limit their mobility. While this may es- informal care. While this intergenerational support pecially be a challenge in lower income countries, has benefits, it can also erode the agency of elderly evidence points across the board barrier, with persons, and can lead to foregoing the healthcare older women from diverse countries such as Viet needs of elderly to save resources for other, young- Nam and Switzerland experiencing trans- er members of a household, especially in poorer portation barriers (Hamiduzzaman et al. 2017). households (Hossen and Westhues 2010; Lena et One study of female veterans in the United al. 2009), direct neglect or abuse (Jamuna 2003; States (n=3,608), for example, found that while Mansur et al. 2010), or stress from being in state of there were considerable gender gaps in access to dependence on children or other relatives (Jeon et healthcare, measured as delays in health seeking al. 2013). For instance, in their study of 9615 adults behavior, transportation was the leading cause from seven States in India, Perkins et al. (2016) find among women over age 65 years for such delays that widowed women were more likely to have poor (36 percent), followed by childcare or other health outcomes compared to married women. This responsibilities (18 percent), and affordability was linked to hypertension and stress. Widowhood (11.5 percent) - (Washington et al. 2011). had no similar effect on men. Another study in Kar- Another systematic review in the United States nataka (n=415) found that the elderly in the study attributed the transport gap as a barrier for 3 to 21 suffered from hypertension followed by arthritis, percent of the older population, across different diabetes, asthma, cataract, and anemia. In terms of settings in the country (Syed et al. 2013). COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 8 235 attitudes, about 48 percent felt they were not happy, women compared to men (Guerra, Alvarado and and 68 percent said that elderly were often neglect- Zunzunegui, 2008; Thorpe et al. 2008). ed (Lena et al. 2009). Women, especially in lower income settings, Social and gender norms also play a role in tend to face greater financial constraints to health ser- setting different expectations from and for wom- vices compared to men, making it harder to seek ap- en and men, which contribute to their limitations. propriate and timely healthcare. In India, widowhood While for women, gender-based socio-economic is significantly associated with women’s use of lower constraints such as limited decision-making pow- cost public health services in older ages but not so for er, norms that limit their mobility and/or financial men or married women, who are more likely to use resources may be larger barriers to access, norms higher quality and more expensive private healthcare surrounding masculinity tend to be greater barriers (Hossain et al. 2021). As highlighted earlier, depen- to men in seeking help or treatment. dency on others particularly financial and tradeoffs such as care for other household members often play Affordability of health services into these decisions. Evidence from a small-scale for older women study (n=152) in seven urban settings in Europe highlighted that while poverty is a barrier for access A main barrier to access to healthcare, especially to prescribed medication, it is significantly more so among the aging is the high cost of care. Research, for women than men (Stankuniene et al. 2015). The especially in high income countries, has shown that interaction of race or ethnicity with gender and other poverty is a main barrier to access. Those with lower socio-economic characteristics exacerbate women’s incomes tend to have a higher level of illness, either poor health outcomes. In their study of bladder can- self-assessed or otherwise. This is likely the result of cer survival among Floridians, Brookfield et al. (2009) avoiding health care services because of the costs in- found that while white, non-Hispanic men with high- volved. For example, using data from the Aichi Ge- er incomes had the highest survival rates, being from rontological Evaluation Study (AGES, n=15,302), poor communities, female, and African American Murata et al. (2010) found that low-income older had the lowest survival outcomes (n=31,100). persons (aged 65 and above) experienced more ill- Evidence shows that overall wealth, and not ness that required medical attention (such as hy- just income, is an important indicator of health out- pertension) than their richer counterparts. They comes for older adults. People with more material also had lower levels of self-perceived health than wealth also report a higher level of subjective health those in high income groups. Although the study (Adena and Myck 2014; Hornby-Turner, Peel, and did not find any gender differences, poverty seemed Hubbard 2017; Nummela et al. 2007). Adena and to be the driving factor for access to healthcare, Myck (2014), for example, examine the effect of with low-income group 1.4 times less likely to com- poverty over a 4-to-5-year period among Europe- plete needed treatments compared to those in the ans aged 50+ from 12 countries using the Survey of high-income groups. Health, Ageing and Retirement in Europe (SHARE). Poverty is clearly linked to health outcomes They found that while relative income had no effect for the elderly, evidence on gender differences in on changes in health, broader measures of poor ma- outcomes is often mixed. This is likely because of terial conditions, such as subjective poverty or low the extent of interaction with other socio-economic wealth, significantly increased the probability of factors. Prus and Gee (2003), in their study of el- transition to poor health and reduced the chance of derly in Canada found that socio-economic status recovery from poor health. Using a sample of Medi- significantly affects women’s health but not men’s. care recipients in the US, Fitzpatrick et al. (2004) Szanton et al (2008) find that elderly women in the found that women perceived higher barriers to US, ages 70-79 years who reported financial stress healthcare access compared to men independent of were 60% more likely to die within the next 5 years other factors such as poverty and race. compared to those who did not. More recently, Women’s quality of life as they age can deteriorate Palta et al. (2015) found that among older wom- considerably due to illness and lack of resources. Con- en, financial stress was linked to higher oxidative trol over financial resources and access to care are two stress. Earlier studies in Brazil and US found that important dimensions of healthy aging (McMaughan, socioeconomic inequalities across the life course Oloruntoba, and Smith 2020; MIszkurka et al. 2012; contribute to higher levels of disability among older Read, Grundy and Foverskov 2015). For women, this 236 CHAPTER 8 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E is also rooted in their voice and agency. Traditionalism, Simultaneously, it is important to ensure health negative attitudes towards aging, and stigma can also and other data systems are not gender neutral. Women contribute to self-perceived lower quality of life (Sun especially in lower income and more traditional com- et al. 2022). For example, a recent Finnish study found munities may be left out of formal markets and social that in the aftermath of COVID-19 women reported systems because of the lack of proper documentation more symptoms and a lower quality of life than men and non-reporting. For example, in communities with (Lindahl et al. 2022). With a significant proportion of more stereotypical gender norms such as those that the population over the age of 60 by 2050, including in promote female seclusion or ‘purdah’, not all female sub-Saharan Africa, women will also constitute a larg- births and deaths are recorded in civil registration and er proportion of the elderly. vital statistics (CRVS) systems. This excludes women Interventions that empower women across their from formal spaces and social protection programs life course and support their access to resources such due to lack of appropriate identification papers. Sim- as income, social protection measures, and health ilarly, health information and management systems insurance that enable them to take control over their may underreport on women’s and men’s health when lives are pivotal for strengthening their capacity to data are not sex-disaggregated and when collection is seek, access, and use health service. incomplete (Muñoz et al. 2020). SUPPLY OF HEALTH AND CARE SERVICES FOR AGING POPULATIONS The following section examines issues related to the supply of care services. Although broader health system issues are of relevance for the aging, as they are for all age groups; the availability, cost, and quality of long-term care in particular is important in understanding their health and well-being as populations age. Moreover, long term care is inclusive of the medical and healthcare needs of the elderly. In this vein, this section focuses on the supply of care for aging population from the perspective of long-term care. Availability of services Sharma, and Zhao 2023; Rocks et al. 2020), another The current majority of aging economies are high- or key barrier for elderly, especially women. Models of middle- income countries, with considerable varia- bringing care closer to women, such as communi- tion in the quality and availability of services. In high- ty-based nutrition programs aimed at pregnant and er income countries of Western Europe and North lactating mothers and their children, and provision America, there is generally greater availability of key of door-to-door family planning services have played health services and long-term care arrangements an important role in improving maternal and child compared to middle income countries. Availability health outcomes in low- and middle-income coun- of health services is dependent on a number of fac- tries (Cortez et al. 2014; Kuruvilla et al. 2014). tors such as availability of health insurance, adequate Similarly, at the country level, long-term care supply and distribution of the health workforce, how (LTC) arrangements are varied, depending on gov- essential and specialized healthcare is provided, and ernment policy as well as social norms (Ngai and Pis- availability of essential drugs and medicines. sarides, 2009) and influence how and when women Policy decisions on how health and other ser- and men access and use these services - the way that vices are organized affects women’s access to these LTC is defined, how LTC is arranged and who is re- services. For instance, availability of health services sponsible for its delivery. It should be noted that the may be limited due to sub-optimal distribution of aim of LTC is to prevent deterioration and adjust to physicians across regions or clustering of service de- stages of decline and is broader than medical or acute livery infrastructure (Chan et al. 2005; Mobley et al. care. Whereas the primary goal of medical or acute 2006), which may contribute to access barriers for care is to return an individual to a previous function- older people, and women in particular. Integrated ing level or to at least ease the burden of pain, the goals primary healthcare (PHC) that brings services closer of LTC are much more complex and encompass social to the populations they serve have the potential of ad- and emotional well-being. This necessitates a broader dressing some of these challenges. Several studies also array of services and arrangements for care provision. show that this can also reduce the cost of care (Hou, LTC services can be organized in a broad variety, COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 8 237 ranging from family care, public institutions, private contributions. Under this Scandinavian care regime, residences and assisted-living environments depending characterized by predominance of public care services, on the needed level of care (Frank 2012). “Some coun- families who can provide care would be recompensed tries (Denmark, Finland, Norway, or Sweden) chose for doing so. In contrast, Southern European countries to frame care provision as a societal task rather than (Cyprus, Greece, Italy, Malta, Portugal, or Spain) would a private matter where comprehensive care provision represent the other extreme, with almost exclusively is done through public care providers, both in institu- family care provision, saving on public expenditure tional and in social care. This public responsibility is through the continuation of traditional caregiving roles funded through general taxation and social insurance by the family” (Hoff, Feldman, and Vidovicova 2010). BOX 8.3 Understanding Long-Term Care Arrangements In May 2016, at the World Health Assembly, 194 countries agreed that every country should have a long-term-care system (WHO 2017a). Yet, many countries are far from instituting any formal LTC arrangements. Many countries, especially the lower income ones, are mostly reliant on informal family care; in other countries social norms contribute to the same reliance. Larger cohorts of younger populations and family networks that can provide this care, also facilitate family care. In other countries, particularly higher income ones, with larger cohorts of aging populations, and declining working age populations, there are also publicly and privately funded formal long-term care arrangements that aim to support the needs of elderly populations. However, to date there is no formal systematic method to categorize and compare LTC systems cross nationally. Applebaum (2013) building on previous work by Colombo et al. (2011) and Kraus et al. (2010), offers a five-category typology to classi- fy approaches to LTC services and support arrangements, combining issues of financial and functional disability requirements with the supply, balance, and array of long-term services available (Table 8.2 below). This classification approach draws heavily on the OECD financial categories —universal versus means-tested or mixed—availability of services and the amount of out- of-pocket expenditures by individuals. In addition, they have combined these factors with additional delivery system indicators, such as the availability of residential care and the balance of formal and informal LTC services. TABLE 8.2 Typology to Classify a Country’s Long-Term Care Services and Supports System Group 1 Group 2 Group 3 Group 4 Group 5 Public insurance funding Mixture of public All funding for long-term Funds are means tested but quite No public funds are available for long-term care- insurance and means- services is means-tested limited in availability available for long-term services tested funding available care-services HCBS (home- and HCBS widely available HCBS commonly HCBS are of limited availability HCBS not available community-based services) available widely available Institutional care widely Institutional care widely Institutional care widely Institutional care somewhat available Institutional care rarely available available available available Housing with services Housing with services Housing with services Housing with services is of limited Housing with services not widely available widely available available availability available Cash payments often Cash payments generally Cash payments available Cash payments not available Cash payments not available for long-term available on a limited basis available services Informal care is one Informal care is an important Informal care is a critical Very heavy reliance on informal care Exclusive reliance on component of the system part of the system element of the system informal care Examples: Germany, Japan, Examples: Australia, France, Examples: Estonia, Italy, Examples: Argentina, Brazil, China, Examples: Bangladesh, Korea, the Netherlands Ireland, Spain, Switzerland Poland, Romania, United Egypt, India, Mexico, South Africa, Ghana, Kenya, Nepal States Thailand Source: OECD 2011 238 CHAPTER 8 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E These various approaches reflect the contrast- lution of its health status. Affordability of healthcare, ing solutions to the dilemma of growing demand then, is dependent on these national level decisions, for care with ultimate responsibility for caregiving and household wealth which influence the level of assigned to different societal actors. The differences out-of-pocket payments for different services. Given in interpretation of whether LTC issues should be the high burden of NCDs for older populations, af- relevant to the public sphere vs. the private sphere fordability of services plays an important role in the have implications in terms of development of long- type of services the population 45 and older receives. term care systems (OECD 2011). Regardless of Health and other social sector expenditures these differences, the form this system takes varies, generally form a very small proportion of public depending on each country’s context, available re- expenditures in lower income countries, with high- sources and societal choices about the distribution er out of pocket payments for health care. Figure of the overall costs of care. 8.13 shows the breakdown of health expenditure by source. In high income countries majority of The cost of health and LTC services: healthcare is covered through domestic public ex- supply and affordability penditures, whereas for countries in lower income groups, private expenditures make up the larger Availability of services is tied to the public and proportion of expenditures, and the health sys- private spending on health care. Decisions about tem (particularly in low-income countries) is also healthcare financing and the mix of public and pri- dependent on external resources. This has impli- vate healthcare spending depend on factors such as cations for availability of health services and their GDP growth, social protection, and private insur- affordability for those who are poorer and more ance, health infrastructure and technologies, and vulnerable. As discussed earlier, in majority of the demographics such as life expectancy, infant mortal- cases, this puts women at a disadvantage since they ity, the age structure of the population, and the evo- lack similar resources and opportunities as men. FIGURE 8.13 Breakdown of Health Expenditure per capita in PPP, 2019 Source: World Bank, Health Nutrition and Population Statistics 2019 (Accessed December 15, 2022) COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 8 239 Figure 8.14 shows pre-COVID out of pock- quality of care, as well as the variations in out-of- et healthcare expenditures as a percentage of total pocket payments for those seeking LTC care. 8 health expenditures and per capita. While the per The cost of LTC is generally high relative to the capita expenditure in lower income countries is income of individuals. Formal LTC, especially, is of- low (Figure 14b), the proportion of out of pocket is ten unaffordable in the absence of social protection. considerably high at 42 percent of the total health All OECD countries have some form of social pro- expenditure. On the other hand, in higher income tection for LTC, but even where coverage is compre- countries, which also have a higher prevalence of hensive, people pay some of the cost out of pocket NCDs and larger cohorts of older populations, the (OOP). In Germany and the United Kingdom, for ex- per capita cost of healthcare is relatively very high ample, OOP payments amount to approximately 25 but because of greater public expenditure and avail- and 30 percent of total LTC expenditures, respective- ability of private health insurance, the out-of-pock- ly (Kraus 2016). Moreover, formal coverage for home et expenditures are considerably lower (13 percent). care for moderate or severe needs is often insufficient Affordability of long-term care services also var- and expensive. This contributes to large OOP pay- ies dramatically. Although data are limited to aging ments. Levels of difficulty with daily tasks was found economies, mainly of the OECD, there is consider- to be strongly associated with unmet needs. Com- able variation between them. Public spending on pared with recipients of spousal care, men receiving long-term care ranges from more than 4% of GDP in care from one non-spousal caregiver, and men and the Netherlands to less than half a percent of GDP in women receiving care from any other configuration, countries such as Israel, Latvia, and Poland (OECD had higher odds of unmet needs (Potter 2017). Un- n.d.7) This variability is reflected in the supply and less there are familial or other arrangements for in- FIGURE 8.14 Out of Pocket Expenditures, 2019 Source: World Bank, Health Nutrition and Population Statistics Database (Accessed December 15, 2022) 7   http://www.oecd.org/els/health-systems/long-term-care.htm 8   Comparing LTC expenditures is a difficult since National LTC systems differ widely in terms of eligibility criteria, level of benefits, institution- al variety and regional heterogeneity (Geyer, 2020). 240 CHAPTER 8 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E Improving the quality of services formal care, many people are unable to afford LTC in their own home, leaving them with unmet needs or the quality of health and long-term care services  is at risk of early institutionalization (Muir 2017). crucial to the quality of life of people who rely on these In lower income countries with less developed so- services. Quality of healthcare is defined as, “the extent cial protections systems and limited coverage for LTC, to which health care services provided to individuals and OOP tend to be even higher, with many families de- patient populations improve desired health outcomes. In pending on more traditional, home-based family care. order to achieve this, health care must be safe, effective, A 2016 global study found that almost two out of three timely, efficient, equitable and people-centred” (WHO). older citizens live in a country where families rather The WHO also conceptualizes the quality of LTC as than governments have the legal duty to provide LTC “care that achieves gains in health, and is responsive to services for older family members who cannot cover the legitimate expectations of LTC recipients” (Brodsky their needs on their own. While in several Asian coun- et al. 2003). Quality of care, as perceived or experi- tries the nuclear family is held responsible, in many enced, can affect its use. Quality of care is affected by African countries a more widely defined family is ob- a multitude of factor from provider skill, knowledge, ligated (Scheil-Adlung 2015). However, these require- and behavior (Box 8.4) to supply chain interruptions, ments can overlook the capabilities of family caregiv- old and malfunctioning equipment to poor infra- ers and increase pressure on them in terms of physical structure. Thus, if the quality of public healthcare is and emotional stress, limiting social interactions, and perceived as poor, people will avoid it and prefer to use increasing financial obligations, especially for women other sources such as private sector providers if they (Glinskaya et al. 2023). Chapter 14 on adequacy pen- can. Whether for health services or LTC, it is import- sions and healthcare provides additional discussion on ant that systems are safe, effective, and meet the needs financial coverage for health and long-term care.” of their users so that these expectations are met. BOX 8.4 Provider behaviors and access to care for older adults Healthcare providers are a crucial part of service delivery and play an important role in influencing peoples’ health seeking behaviors. Poor patient-provider communication can for example, affect the quality of care. A handful of studies also point to the lack of communication or limited discussion about subjects that may be considered intimate or taboo such as sexual health or incontinence among older, middle-aged women and their providers. While women’s socio-economic characteristics play a role in women themselves approaching the topic with their healthcare providers, the limited evidence that exists also suggests that providers very seldom initiate the discussion (Bergeron et al. 2017; Chapman et al. 2019; Hughes and Lewinson 2015). Provider behaviors and attitudes towards specific groups of older women may also put them at a disadvantage. Research iden- tifies a consistent pattern of gender bias in diagnosis and clinical treatment for the same illnesses. One population-based study on trauma treatment found that when adults surviving trauma arrived at hospitals, women were more likely to be directed towards non-trauma care compared to men, indicating a potential gender bias, although the authors could not rule out other factors (Gomez et al. 2016). In Sweden, hospital level data on waiting times and triage priority for older adults (mean age 54 years, n = 135,417) indicated presence of gender bias in combination with age and socioeconomic status (Roberston 2014). Us- ing data from primary care practices associated with two academic medical institutions Madonis, Skelding, and Robert (2017) note that women are less likely to receive an early diagnosis of CVD than their male counterparts. In addition, they are less likely to receive appropriate, timely interventions despite having worse outcomes than men after acute coronary syndromes and higher postintervention complications (Cook et al. 2009; Pagidipat and Peterson, 2016; Vaccarino et al. 2009). In a similar vein, Fitzpatrick et al. (2004) reported that physician attitude was a greater barrier to seeking healthcare rather than transport or cost among the elderly in their study of Medicare beneficiaries in the United States. Fabreau et al. (2015) also point towards a need for improving the quality of care for women, especially those with low incomes. Recognizing these challenges and taking steps such as provider education to reduce bias, periodic skills training and refreshers, and ensuring that health information is available and clearly communicated to patients can counter this gap in quality of care. COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 8 241 Ensuring quality of care achieves these ideals ted to hospital (Hamada et al. 2019). necessitates standardized protocols, appropriate However, few countries systematically measure regulation, and accountability. For example, regula- whether their LTC systems are safe, effective, and tory standards guiding accreditation of care homes centered around the needs of service users. Indica- and qualification requirements for care providers tors of LTC quality, including measures of clinical can improve quality of formal LTC services. Cer- quality, user satisfaction or quality of life surveys are tainly, Regulatory oversight and quality assurance useful for assessing service provision. However, be- is more visible and rigorous in institutional care cause of the costs and logistics involved, only a few settings, while home or community-based care set- countries measure quality of care; and where mea- tings are subject to little or no regulation (OECD/ sures are available, they only capture narrow aspects European Union 2013). When available, a 2022 of clinical quality rather than providing a complete study found that the complex process of achieving profile of the quality of life experienced by LTC us- care: the functional fragmentation and geographical ers. This leaves space for quality of care to decline, dispersion of care institutions made the care-seek- especially when issues such as negative provider at- ing process challenging and confusing for older titudes (discussed earlier), or treatment of elderly in adults to access (Gong et al. 2022). Similarly, trained LTC homes or institutions are not captured in data physicians, nurses, and other hospital staff, that can or addressed in practice. This is especially of con- provide comprehensive health services in a timely cern for countries where implementation of quality manner can improve the quality of care. A recent controls is already weak. study on treating elderly in Japan for instance found OECD countries, where aging is a reality to- that elderly patients with pneumonia admitted to day, have taken different pathways to improving the a hospital had better quality of care compared to quality of care, and offer examples of how to bridge those not at the hospital. From a service provision the quality gap. Quality oversight especially related perspective, the care was both cost effective and to institutional care – and incentives to improve timelier (e.g., administration of drugs or change of quality are the most common. To bridge the gap be- IV tube) as compared to that for elderly not admit- tween clinical care quality and quality of life of the BOX 8.5 Improving Quality of Care through Standards and Practice in the OECD Compulsory accreditation of LTC facilities – Although all countries have legislation regarding adequate and safe care, only in two-thirds of OECD countries, accreditation or certification of care facilities is compulsory or common practice (e.g., England, Spain, Ireland, France, Australia, Germany, Portugal, the United States, Switzerland). Some countries (e.g., Poland) issue certifi- cates for care establishments where minimum standards are often used as evaluation criteria for the authorization of provision. However, enforcement of standards has often lagged behind, given the difficulties and costs ·entailed for both the authorities and providers to supervise/comply with them (Chawla, Betcherman, and Banerji 2007). Standards to normalize care practice - Standardized assessment tools can be helpful to develop care plans and interven- tions, promote consistency in care and prevent adverse events. Examples include: the Resident Assessment Instrument (RAI) used in Belgium, Canada, Finland, Iceland, Italy, the United States, and Spain; the Autonomie, Gérontologie, Groupes Iso-Res- sources (AGGIR) scale in France*; and the Katz Index of Independence in Activities of Daily Living in Belgium and elsewhere. Public reporting on LTC quality is also mandatory in some countries, such as the United States, Japan, England, Germany, Portugal, the Netherlands, and Canada. Evidence about the effectiveness of such tool to encourage providers to improve their standards is however mixed (Yeh 2021). Australia set up a new quality system, including a review of accreditation procedures for providers, improved monitoring and dealing with complaints. The Czech Republic and the Slovak Republic implemented oversight policies, authorization, and com- pliance of providers’ quality, while Germany enhanced quality supervision and aims at enhancing quality management of providers, together with improved consumer voice. Ireland published in 2009 the National Quality Standards for Residential Care Settings for older people, while Austria produced a handbook on dementia and Luxembourg has installed a committee on quality of care (OECD 2011). * The AGGIR grid is designed to assess an elderly person’s degree of dependence or autonomy. 242 CHAPTER 8 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E end user, a third of the countries employ market in- further education, training and professional growth centives for providers and users, allow users to make courses are provided within the nursing home care. decisions on service provision, for example through New caregivers go through a training where they cash benefits. The Netherlands implements a Care are assessed on specific caregiver assessment cri- Innovation Platform, aimed at the development, teria including aptitude for teamwork, attitude to structural dissemination, and implementation of elders, empathy, commitment, and independence. innovations in (long-term) care provision. In some Box 8.5 highlights some ways in which the OECD is countries, caregivers are offered the possibility of addressing quality of care for the elderly. WOMEN AS LONG-TERM CARE PROVIDERS With population aging, especially in middle- and high- income countries, the need for long-term care (LTC) is growing. This is a labor-intensive sector comprised of a formal workforce as well as family care givers. Care arrangements across countries vary, with greater reliance on informal, family-based care in lower income coun- tries, and a greater mix of formal and informal arrangements in higher income ones. In all cases, women form a large proportion of the care providers. In 2020, a study that focused on LTC workers found that essential long- term care workers commonly held second jobs and double- or triple-duty caregiving roles (Van Houtven, De- Pasquale, and Coe 2020). Information on female long-term caregivers however ment that the largest amount of intergenerational time is limited in scope. While data exist on women’s bur- transfers are provided by middle-aged (45–65 years den of caregiving for the elderly in several countries, old) women to both the younger and older genera- in most cases these are higher income countries and tions within the family in OECD countries. In China, formal caregiver. Less is available on the impact of an estimated 14 percent of people aged 45 years and caregiving on the lives of these women, or the types over provide care for the elderly, with an additional 28 of support they need. The little evidence that exists percent engaged in providing care for grandchildren. however, points to increase in vulnerability of female While women assume a crucial role in intrafamily caregivers, particularly informal caregivers. This sec- care services, constituting the chief contributors of tion highlights some of the key issues that have been long-term care and childcare, their care contributions identified in the literature review on female caregiv- at home are not recognized as “social achievements” ers in the context of LTC. and are not monetarily compensated by the Chinese Whether through formal or informal arrange- welfare state (Wang and Liu 2020).  Women aged 55- ments, women are often the providers of long-term el- 59 years provide about 16 percent of elderly care, with derly care. According to one estimate about 67 percent another 45 percent providing childcare (World Bank of the healthcare workforce globally is comprised of 2016). In Colombia, Indonesia, and Poland, women women (WHO 2017b). About 60 percent of informal aged 40 to 59, are 2 to 3 times more likely to provide healthcare providers in OECD countries are women care for older parents than men (Gatti et al. 2024). In (OECD 2013), and a much higher proportion (90.7 the United States, the majority of care providers are percent) of LTC providers are women (Figure 8.14). also women, with a mean age of 46 years (Stone and An earlier study found that up to 80 percent of all Dawson 2008; Yamada 2002; Montgomery et al. 2005). health care and 90 percent of HIV/AIDS related illness Many are even older, with the proportion of home care care was provided at home, with little recognition of, workers over 65 years, roughly triple the mean age in or support for the care provider (Östlin 2009). Wom- other industries (Montgomery et al. 2005). en’s concentration in caregiving is related, at least in With population aging, the demand for LTC part, to the underlying gender-based division of labor is expected to increase. According to estimates, by and social norms that assign certain tasks, activities, 2050, there will be roughly 1.5 billion persons aged and occupations as more suitable for one gender or 65 years or above (United Nations 2019). This is go- the other. Occupations such as nursing and caregiving ing to contribute to increase in the demand for for- have thus been viewed as more suitable for women. mal and informal LTC arrangements. For example, In many cases, women provide this care as they in the United States about 71 percent of long-term are aging themselves. Albertini et al. (2016) docu- care providers are direct care workers. It is estimat- COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 8 243 FIGURE 8.15 Share of female LTC workers in select OECD countries (% of total, latest available year) Source: Adapted from OECD 2019. ed that with population aging, the demand for their sociated with nursing and less-skilled health- and services will increase by 48 percent by 2030 if all care-related functions (ILO/WHO 2022). else remains the same (U.S. Department of Health One of the key features of the care economy is and Human Services 2017). In China, the demand its draw on migrant labor to provide care services. for LTC services has already increased dramatical- There are significant flows of care workers from ly due to the aging population with an increase in low- and middle-income countries to high-income population over 65 years from 96 million in 2003 to countries globally. For instance, 20 percent of care about 150 million in 2018. New studies examining providers in the United States were migrants, main- disability patterns in 23 lower- and middle-income ly of Hispanic or Latino origin (Montgomery et al. countries predict that the rate of severe activity lim- 2005). Similarly, in Italy 83 percent of care providers itations among older adults will remain relatively were migrants (Polverini and Lamura 2004). Sever- stable over the next three decades. Nevertheless, the al countries are both, countries of origin and des- total number of individuals facing activity limita- tination for the flow of migrant care workers. For tions is expected to rise due to the significant growth instance, while Polish care workers have migrated in the older population (Glinskaya et al. 2024). for work across Europe (in particular, in Germany, Given the trend for female employment in this Ireland, Italy, Sweden, and the United Kingdom), field, it is likely that increased demand will draw the country - at least the metropolitan area around more women into the care economy. Comprising Warsaw - has also experienced an influx of mi- nursing assistants, home health aides, personal grant care workers from Belarus and the Ukraine care aides, and psychiatric aide, direct care workers (Hoff 2011). These flows have also resulted in subtle play a crucial role in providing the majority of paid changes in the gendered experience of migration: hands-on care, including supervision, medication while in the past migration was typically associat- administration and emotional support to the elder- ed with young male workers, increasing numbers ly and disabled. Notably, they are responsible for of contemporary migrants are female working in 8 out of every 10 hours of care provided (Dawson eldercare (Leeson and Harper 2006). and Surpin 2001). An ILO study of 2020 confirmed While elderly care is a source of employment that although in all countries women continue to for women, it is generally associated with low waged be overrepresented in occupational categories as- work. In Japan, the United Kingdom, and the Unit- 244 CHAPTER 8 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E ed States, caregivers typically earn between 50-70 force is at least double that of the formal care work- percent of the average wage. Similarly, work-related force in Denmark for example, and in some cases, it benefits for LTC workers, such as pension or childcare is estimated to be more than ten times the size of the benefits, tend to be limited (Hickey, Sawo, and Wolfe formal-care workforce such as in Canada, New Zea- 2022). In addition, caregivers are frequently subject to land, United States, and The Netherlands (ENNHRI long and irregular working hours, while their work- 2017). In fact, when formal services are unable to ing situations are often precarious since part-time and meet the demand for care because of high costs, or short-term employment contracts are common (Wit- because of unavailability of services, it often falls on ters 2011). Almost half (45 percent) of LTC workers in the shoulders of women in the household to provide OECD countries work part-time, over twice the share care (e.g., Chawla, Betcherman, and Banerji 2007). in the economy as a whole (OECD 2020). Home-based support relies greatly on informal care In Europe, where more information on LTC is provision (Sunwoo 2017), many older adults receive available, evidence points to a high turnover of care care from one or more family caregiver, and some providers. Difficult working conditions such as long caregivers may help more than one older adult. The shifts; having to manage complex health and phys- circumstances of individual caregivers and the care- ical care needs of elderly and providing end of life giver context are extremely variable with either fam- care to patients, which can take a mental and phys- ily caregivers living with, nearby, or far away from ical toll on the care giver, as well as being exposed the person receiving care. Nevertheless, the family to illness or aggressive behaviors from patients caregiver’s involvement is determined primarily by with mental health conditions such as dementia, a personal relationship rather than by financial re- combined with and low pay often generate high muneration. The care they provide may be episodic, turnover among workers, and create difficulties in daily, occasional, or of short or long duration and attracting highly motivated, highly qualified care can be provided in a variety of settings such as nurs- staff. Overall, labor intensity in the LTC sector var- ing homes, residential care facilities, and individual ies widely throughout Europe. Sweden and Germa- homes, and can include a variety of tasks.9 ny have the highest ratio of formal workers against The balance between informal and formal care service users (1.1 and 1.0 respectively), compared differs substantially from country to country and with 0.19 and 0.15 in Estonia and the Czech Repub- changes over time within countries as it is connect- lic where formal LTC is still relatively under-de- ed to particular political, economic, demographic, veloped. Moreover, qualification levels tend to be and cultural factors. In southern European countries, lower on average in relation to the healthcare sec- informal care remains the dominant source of home tor overall, and the ratio of nurses to care workers care. Without the work of these unpaid care givers, is also low (Witters 2011). Poor job quality limits home care would be totally unsustainable, and many retention in LTC with a large body of evidence acute needs would remain unattended. The situation suggesting that shift work is associated with a wide is different in northern European countries where range of health risks, such as anxiety, burnout, and informal care is less common because municipal- depressive disorders (OECD 2020). ities provide extensive personal care and domestic Stressful work conditions contribute to care- services (such as home help in Denmark, although giver absenteeism. For example, in the US, it is esti- private organizations may provide this with munic- mated that this absenteeism is responsible for US$ ipal funding), and informal care tends to focus on 25.2 billion in lost productivity for the economy providing companionship and social support. based on the average number of workdays missed While the care economy has given a boost to per working caregiver, assuming US$200 million in women workers and female migration, several stud- lost productivity per day (Witters 2011). ies also find that care workers are also vulnerable The gap in care supply is often made up through to specific risks. In Europe for instance, care poli- the provision of informal supply by family and cies have exposed women migrants to three inter- friends. The size of the informal family care work- connected levels of inequalities through: (a) the 9   The scope of services ranges from in kind service provided to family members on a non-paid basis to bathing, eating, dressing, running errands, help with medication management, paying bills, transportation, meal preparation, and health maintenance tasks and managing financial affairs. COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 8 245 gendered and racialized everyday experiences of sions and the size of these pensions when they are migrant women workers; (b) the polices of national eligible to receive them. In turn, this has implication welfare states to find ways of reducing the care costs for their economic wellbeing. Stepping out of the la- associated with increased women’s employment and bor market can reduce their lifetime income signifi- an ageing society; and (c) in a transnational political cantly and the amount of pension they qualify for. economy of care in which big business and states Moreover, those engaged in elderly care lack predict- drain poorer countries of their care resources by ability in timing, duration, and intensity that comes drawing on their labor force (Williams 2011). with care, making it more difficult to negotiate a care Care work is labor intensive and has conse- and paid employment, particularly full-time em- quences for the care provider. One estimate puts the ployment (Keck and Saraceno 2009). mean hours of informal care at 46 hours per week For those women who remain in employ- and the mean length of time spent on care related ment while engaged in informal care, productivity tasks is 60 months. Most often this care is provid- (and thus, earnings) can be substantially reduced ed by adult daughters or daughters in law, who may through different channels, such as change in choice have to forego their own employment to provide the of employment to accommodate care demands, in- care (Albertini 2016; Karpinska et al. 2016; World cluding informal jobs that are less stable and offer Bank 2015). In 2018, the ILO estimated that at this lower wages and impacts on physical and mental pace, it will take 210 years to close the gender gap health. Even timing of retirement is moderated by in unpaid care work (ILO 2018). Women may also gender. For example, wives caring for their hus- make alternate work arrangements such as part- bands have retirement odds 5 times greater than time or less demanding jobs. Evidence from the US women who are not caregivers, whereas husbands suggests that compared to their male counterparts, caring for their wives are substantially slower to women caregivers are more likely to make alternate retire (Dentinger and Clarkberg 2002). Changes work arrangements: taking a less demanding job in the labor force participation of women do not (16% women vs. 6% men), giving up work entirely appear to alter their caregiving responsibilities. In (12% women vs. 3% men), and losing job-related fact, time-demanding support obligations are more benefits (7% women vs. 3% men).10 likely to fall on the shoulders of women, requiring In fact, informal care can be a major constraint them to multitask regardless of whether or not they to female labor force participation. A large body of were employed (Moen, Robison, and Fields 1994). evidence from different countries indicates that in- Female caregivers also suffer health challenges tensive (i.e., 20+ hours per week) informal care work and other hardships due to caregiving. In low- and has a negative impact on the likelihood of staying middle-income countries, despite having received in the workforce (OECD 2011; Schmitz and West- inadequate attention health risks among informal phal 2017). A study conducted in Poland, Colombia, caregivers were found to result in greater health Egypt, and Indonesia reveals that there is a notable burdens, particularly mental health, for women that decrease in the likelihood of employment among were more prominent and manifested in more di- both men and women who provide parental care, verse ways for women relative to men (Bhan, Rao, with the impact being more pronounced for wom- and Raj 2020). For example, the prevalence of men- en, in particular (Gatti et al. 2024). Women may face tal health issues is 20 percent higher in caregivers higher barriers to labor market entry or re-entry be- than in non-caregivers. Caregivers, especially those cause of ageism or may be more likely stay out of caring for men, were often distressed, and remained the labor force to cater to the needs of children (or so over time (Li et al. 2022). This is partly because grandchildren) and elderly relatives - the so-called older men have more health, functional and behav- “sandwich generation”. Indeed, this peak of inelastic ioral problems that require more care, and partly care responsibilities usually coincides with women’s because their caregivers are more likely to be their most productive middle-age years. This has conse- spouse and living together. Similarly, caregivers ex- quences for their present and future income as well perience more hypertension, stress related chronic as affecting their potential for qualifying for pen- illness, and physical injuries since very often they 10   National Alliance for Caregiving and AARP. (2009). Caregiving in the U.S.: https://www.aarp.org/pri/topics/ltss/family-caregiving/caregiv- ing_09.html [Accessed: December 8, 2022] 246 CHAPTER 8 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E are untrained themselves and may lack sufficient have specific reproductive health needs of their own. support networks. Family members who spend Although there are studies that focus on women as long hours caring for older relatives with advanced caregivers, there was limited information on meno- dementia, for example, are especially at risk. Care- pause related health needs and support for caregivers givers may lose income, Social Security and other for example. A 2019 study surveyed residents that retirement benefits, and career opportunities if they recognized important knowledge gaps with regards have to cut back on work hours or leave the work- to menopause, provision of comprehensive and com- force. This is especially true for caregivers with low- passionate care and definite room for improvement er education. They may also incur substantial out- (Berga and Garovic 2019). Usually, when and if stud- of-pocket expenses that may undermine their own ied, mature women are done so often in the context future financial security. Certainly, earlier evidence of informal care or in the realm of dementia and their from the US showed that single women caring for ability to deal with the stress of LTC of mentally ill their elderly parents are 2.5 times more likely than patients. The choices these caregivers face at pivotal non-caregivers to live in poverty in their own old moments of their lives and the type of support avail- age (Wakabayashi and Donato 2006). able to them or lack thereof are not documented. However, overall data on these issues appears to Women need cross-cutting support when they be limited. While studies found that female caregiv- step into the care giving role – support to not only pro- ers (working or not) may suffer a particularly high vide care to the elderly (e.g., training and awareness level of economic hardship due to caregiving, none about resources), but also to take care of themselves will address how these choices were made or wheth- (e.g., social protection). Thus, as also highlighted in er they benefited from a support system (whether literature, there is a necessity for policy makers and emotional, financial or network support). Women program planners to work together to provide acces- caregivers above 49 years fall in the intersectional- sible, affordable, and innovative support services and ity trap of menopause and caregiving. These women programs to support caregivers, especially women. DISCUSSION Men and women tend to experience aging differently, not just because of their biological differences, but also the socio-economic environment. This affects their health needs, demand, and access to health services. While there is a considerable body of literature that examines gender differences in health outcomes, few examine the structural and socio-economic factors behind these gender differences, particularly among aging populations. For instance, considerable attention is paid to certain reproductive health issues such as female and male cancers and menopause, but mainly in terms of clinical research. There is less focus on the socio-economic aspects or the quality of care for women as they exit their reproductive years. There is scope for further work in terms of access to healthcare services and barriers to health seeking behaviors among older women and men. While each country has its own story to tell, gender them, for example are necessary to ensure that old- differences exist in the prevalence of major non-com- er adults receive appropriate and timely care. At the municable diseases such as CVD and diabetes, as same time, such approaches can also support healthy well as in the burden of deaths. While the mortality decisions and health seeking behaviors across the life from CVD is higher among men overall, for older course delaying onset of NCDs as discussed in other age groups, the gap narrows among women and studies in the healthy longevity series. men, and the burden of deaths increases for wom- Women’s burden of depression is considerably en in the oldest age group (80 years and above). A higher than of men across the globe. On the other similar pattern can be observed for diabetes. On the hand, a higher number of men die of self-harm (e.g., other hand, as women live longer, they also live lon- suicide). Some studies suggest that women’s social ger with multimorbidity. It is important to recognize lives, which are more integrated within their com- these differences and the implications of prolonged munities compared to men’s, act as a buffer against prevalence of diseases among women in terms of self-harm and suicide among the older populations. policy and practice. Interventions that bring services On the other hand, studies also suggest that aging closer to women and make them more affordable for women face greater risks in terms of neglect or COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 8 247 abuse due to their dependence on others and epi- dressed. This can have a catastrophic effect their demiological studies suggest similar if not higher quality of life and contribute to early death. Howev- rates of attempted suicides as men. Addressing these er, it is hard to discern the true extent of the prob- issues necessitates not only availability of affordable lem, as highlighted by the OECD, since people who and accessible mental health services (e.g., through do not benefit from LTC are often outside formal different channels including online therapy) but systems, and may not have a voice because of their also social and behavior change interventions that vulnerable position (OECD 2011). promote healthy behaviors and normalize the use Moreover, there are gendered considerations of these services. This is particularly important for related to the supply of LTC, primarily due to the men, who often will avoid seeking help due, in part, large proportion of women who provide care ser- to norms around masculinity and strength. vices. Female caregivers are more likely than males Women’s vulnerabilities tend to exacerbate over to make alternate work arrangements, for instance; time, especially in lower income settings and where or taking up fewer demanding jobs, or even giving social and gender norms create barriers to women’s fi- up work entirely and losing job-related benefits.11 nancial and social empowerment. Many women tend Very little attention has been paid to the impact care to rely on their families for support including their giving has on their own health or their resources as residence, mobility, and accessing health services or they age, while caring for others. Although several other formal or informal care.  Although aging men studies have found that female caregivers (working also experience a growing dependence on others, they or not) may risk economic hardship due to care- are less financially vulnerable compared to women. giving (e.g., Albertini 2016), more work is needed In higher incomes countries, these differences are less to understand how these decisions are made, their magnified, especially in the presence of some form of impact on women’s overall health, and what type of health insurance or social protection that both wom- support systems (whether emotional, financial or en and men have access to. However, affordability networks) if any, women benefit from. issues remain forefront, with mixed evidence on gen- This is also important for future policy making der-based differences. This is likely due to the particu- in an era of global aging. As it stands, women rep- lar contexts of the countries being studied in terms of resent an invaluable underutilized reservoir of labor social norms, women’s employment, and their access and productivity for countries. Ensuring their con- to social protection and health insurance. tinued engagement in the labor markets, access to A similar case can be made for access and af- finance, and participation in social safety nets can fordability of health and long-term care (LTC) for improve their future access to health and long-term the elderly. While informal, family-based care may care. At the same time, women are an important hu- still be the norm in many low-income countries, man resource for LTC, both formal and informal. medical costs remain a barrier to appropriate care Policies and programs that support their roles as care for the aging, particularly women (for all the rea- givers and provide them with platforms for stan- sons discussed earlier). In middle- and higher- in- dardized training or other support, are important come countries with more options for formal LTC, for improving the quality of care as well as enhanc- affordability is still a critical issue. Since women ing the quality of life of care providers. While it is have lower labor force participation and employ- true that country contexts matter, and interventions ment rates or may drop out of the labor force early may differ based on the specific policy environment, to provide care for family members (children, elder- this is an important aspect to consider within these ly), they are less likely to have adequate savings or frameworks. Especially with the changes in family social protection coverage compared to men. This values and the international migration movements puts them at a disadvantage. that leave aged parents behind and alone, making When people do not have access to, or cannot family care more difficult and increasing the de- afford, formal care, and when their families cannot mand for care offered by the government. or will not support them, LTC needs remain unmet. At the same time, quality of care matters. This leaves the elderly without the support they While standardized protocols and regulations are need for daily tasks such as washing and getting at the crux of quality control, their implementa- 11   Ibid. 248 CHAPTER 8 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E tion varies. Unconscious bias may enter into stan- Second, the findings underscore the impor- dards and protocols of care for example because of tance of addressing gender gaps in social protec- a poor understanding of the age and sex differences tion and labor markets especially when considering in illness progression and treatment needs such as ‘retirement income’ or pensions i.e., social security. appropriate dosage. Similarly, implicit gender bias- Women’s limited access to and use of health ser- es may exist in the treatment of patients, which in vices is often to them being ‘locked out’ of formal turn, can contribute to delays in women receiving systems, have disrupted employment, and lack fi- appropriate care at the appropriate time. Elderly nancial resources, especially when they are old. The patients are also more likely to experience abuse at feminization of aging and poverty (i.e., there is a the hands of care givers, yet it can go unnoticed and greater concentration of women in older age groups unaddressed. While there has been some progress, due to their longer life spans and limited financial better representation of women in clinical trials and capacity) exacerbates their vulnerability. Continued in medical texts is still needed. On the latter, focus work on expanding labor market opportunities for on provider training to improve their understand- women and strengthening social protection systems ing of signs of abuse and reduce their own biases in will be one key area for reducing the gender gaps in service provision can go a long way in improving the healthcare sector and expanding women’s access the quality of care for older adults. to health services. Earlier analytical work on aging All these issues indicate the need for examining by the World Bank (Brodmann, Coll-Black, and gender differences further and making concerted ef- Von Lenthe 2023; Chawla, Betcherman, and Banerji forts to address any gaps at the policy and program- 2007; Miglena, Arribas-Banos and DeMarco 2023; matic levels in ensuring access to care, and provision World Bank 2016) and in this HLI series (DeMarco of services. The findings are relevant for the World et al. 2024; deSilva and Santos 2024) point to oppor- Bank and other organizations for strengthening tunities that can address these challenges such as country level work on human development. First, reducing gender gaps in retirement age, exploring health is an important consideration in building alternative pension schemes such as non-contribu- human capital. As the findings indicate, there are tary pensions, promoting pay equity, and expanding gender differences in prevalence and mortality from women’s opportunities in formal job markets. This NCDs especially in older age groups, as well as in also requires a paradigm shift on aging and older access to and use of health care. It is crucial to recog- populations that recognizes them as active partic- nize and address gender-based challenges, including ipants in the community and fosters their abilities for LGBTQ, when considering health policies and for continued contribution to society. programs. This requires a systematic approach to ex- Finally, good data is an important precursor to amine, define, and include the needs of health care policy and programmatic intervention. The HLI has provision for the aging populations, focusing not developed pilot dashboards that present key indica- only on currently aging societies but also younger tors relevant to measuring healthy longevity within countries such as in Africa where there is a dual bur- the context of a country’s priorities along the life den of communicable and non-communicable dis- course (Haldane et al. 2024; Wu et al. 2024). These eases. This is especially relevant for the World Bank indicators should be, to the extent possible, sex-dis- as it works with countries to implement Universal aggregated, and where necessary gender specific Health Coverage (UHC). The existence of UHC indicators should be added. However, the strength alone may not be a guarantee that people receive of these dashboards is only as good as the available the services they need as the above discussion high- data. Often data is incomplete and only disaggre- lights. UHC too needs to be implemented as part of gated for a handful of indicators, especially in lower larger reforms within the health sector that aims to income countries with weak systems. Therefore, it close gender gaps and reduce biases in service deliv- remains important that HMIS and CRVS systems ery as well as multi-sector interventions that reduce are strengthened and ensure wider coverage and social and other access barriers to healthcare. collection of sex-disaggregated data. COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 8 249 REFERENCES 1. [Editorial]. 2021. “It’s Time to Expand the Definition of ‘Wom- dreas Schulz, et al. 2018. “New Onset of Depression in Aging en’s Health.’” Nature 596 (7870): 7–7. https://doi.org/10.1038/ Women and Men: Contributions of Social, Psychological, Be- d41586-021-02085-6. havioral, and Somatic Predictors in the Community.” Psycho- 2. Abels, Miglena, Loli Arribas-Banos, and Gustavo DeMarco. logical Medicine 49 (07): 1148–55. https://doi.org/10.1017/ 2023. “The Gender Pension Gap: What Does It Tell Us and What s0033291718001848. Should Be Done About It?” Jobs and Development - World 15. Bhan, Nandita, Namratha Rao, and Anita Raj. 2020. “Gender Bank Blogs [Blog]. June 27, 2023. https://blogs.worldbank. Differences in the Associations between Informal Caregiv- org/jobs/gender-pension-gap-what-does-it-tell-us-and- ing and Wellbeing in Low- and Middle-Income Countries.” what-should-be-done-about-it. Journal of Women’s Health 29 (10). https://doi.org/10.1089/ 3. Adena, Maja, and Michal Myck. 2014. “Poverty and Transitions jwh.2019.7769. in Health in Later Life.” Social Science & Medicine 116 (Septem- 16. Black, Eleanor, Fran Hyslop, and Robyn Richmond. 2019. “Bar- ber): 202–10. https://doi.org/10.1016/j.socscimed.2014.06.045. riers and Facilitators to Uptake of Cervical Cancer Screening 4. Afroz, Afsana, Khurshid Alam, Liaquat Ali, Afsana Karim, Mo- among Women in Uganda: A Systematic Review.” BMC Wom- hammed J. Alramadan, Samira Humaira Habib, Dianna J. en’s Health 19 (1). https://doi.org/10.1186/s12905-019-0809-z. Magliano, and Baki Billah. 2019. “Type 2 Diabetes Mellitus in 17. Bone, Jessica K, Gemma Lewis, and Glyn Lewis. 2020. “The Role Bangladesh: A Prevalence Based Cost-of-Illness Study.” BMC of Gender Inequalities in Adolescent Depression.” The Lancet Health Services Research 19 (1). https://doi.org/10.1186/ Psychiatry 7 (6): 471–72. https://doi.org/10.1016/s2215- s12913-019-4440-3. 0366(20)30081-x. 5. Albertini, Marco. 2016. “Ageing and Family Solidarity in Eu- 18. Brodmann, Stefanie, Sarah Coll-Black, and Cornelius Von rope: Patterns and Driving Factors of Intergenerational Sup- Lenthe. 2023. “Breaking the Cycle of Poverty: Empowering port.” Social Science Research Network. Rochester, NY. May Women through Enhanced Social Protection in the Western 17, 2016. https://papers.ssrn.com/sol3/papers.cfm?abstract_ Balkans.” Investing in People - World Bank Blogs [blog]. Sep- id=2781299. tember 14, 2023. https://blogs.worldbank.org/investinpeople/ 6. Anagnostis, Panagiotis, and John C. Stevenson. 2023. “Cardio- breaking-cycle-poverty-empowering-women-through-en- vascular Health and the Menopause, Metabolic Health.” Best hanced-social-protection-western. Practice & Research Clinical Endocrinology & Metabolism, 19. Brodsky, Jenny, Jack Habib, Miriam J. Hirschfeld, and WHO Ini- April, 101781. https://doi.org/10.1016/j.beem.2023.101781. tiative on Home-Based Long-Term Care. 2003. Key Policy Issues 7. Applebaum, Robert., Anthony and Emily Robbins, E. 2013. in Long-Term Care. Geneva, Switzerland: World Health Organi- “International approaches to long-term services and sup- zation. https://iris.who.int/handle/10665/42604. ports.” Generations 37(1): 59–65. https://www.jstor.org/sta- 20. Brookfield, Kathleen F., Michael C. Cheung, Christopher Gomez, ble/26555977. Relin Yang, Alan M. Nieder, David J. Lee, and Leonidas G. Koni- 8. Argyrakopoulou, Georgia, Maria Dalamaga, Nikolaos Spyrou, aris. 2009. “Survival Disparities among African American Wom- and Alexander Kokkinos. 2021. “Gender Differences in Obe- en with Invasive Bladder Cancer in Florida.” Cancer 115 (18): sity-Related Cancers.” Current Obesity Reports 10 (February). 4196–4209. https://doi.org/10.1002/cncr.24497. https://doi.org/10.1007/s13679-021-00426-0. 21. Bussolo, Maurizio, Johannes Koettl, and Emily Sinnott. 2015. 9. Azad, Amee D., Anthony G. Charles, Qian Ding, Amber W. Trick- Golden Aging. Openknowledge.worldbank.org. Washing- ey, and Sherry M. Wren. 2020. “The Gender Gap and Health- ton, DC: World Bank. https://openknowledge.worldbank.org/ care: Associations between Gender Roles and Factors Affect- entities/publication/f77bc092-6cfa-5ace-9526-2c02f3b0221b. ing Healthcare Access in Central Malawi, June–August 2017.” 22. Cameron, Kenzie A., Jing Song, Larry M. Manheim, and Dorothy Archives of Public Health 78 (1). https://doi.org/10.1186/ D. Dunlop. 2010. “Gender Disparities in Health and Healthcare s13690-020-00497-w. Use among Older Adults.” Journal of Women’s Health 19 (9): 10. Barta, Julie A., Charles A. Powell, and Juan P. Wisnivesky. 2019. 1643–50. https://doi.org/10.1089/jwh.2009.1701. “Global Epidemiology of Lung Cancer.” Annals of Global 23. Carlo, Crestani, Masotti Vittoria, Corradi Natalia, Laura Schir- Health 85 (1). https://doi.org/10.5334/aogh.2419. ripa Maria, and Cecchi Rossana. 2019. “Suicide in the Elderly: 11. Berga, Sarah L., and Vesna D. Garovic. 2019. “Barriers to the Care A 37-Years Retrospective Study.” Acta Bio Medica: Atenei of Menopausal Women.” Mayo Clinic Proceedings 94 (2): 191– Parmensis 90 (1): 68–76. https://doi.org/10.23750/abm. 93. https://doi.org/10.1016/j.mayocp.2018.12.016. v90i1.6312. 12. Bergeron, Caroline D., Heather Honoré Goltz, Leigh E. Szucs, 24. Carmel, Sara. 2019. “Health and Well-Being in Late Life: Gender Jovanni V. Reyes, Kelly L. Wilson, Marcia G. Ory, and Matthew Differences Worldwide.” Frontiers in Medicine 6 (218). https:// Lee Smith. 2017. “Exploring Sexual Behaviors and Health Com- doi.org/10.3389/fmed.2019.00218. munication among Older Women.” Health Care for Women 25. Cené, Crystal W., Peggye Dilworth-Anderson, Iris Leng, Lore- International 38 (12): 1356–72. https://doi.org/10.1080/0739 na Garcia, Viola Benavente, Milagros Rosal, Leslie Vaughan, et 9332.2017.1329308. al. 2016. “Correlates of Successful Aging in Racial and Ethnic 13. Bertakis, K. D., R. Azari, L. J. Helms, E. J. Callahan, and J. A. Rob- Minority Women Age 80 Years and Older: Findings from the bins. 2000. “Gender Differences in the Utilization of Health Women’s Health Initiative.” The Journals of Gerontology Se- Care Services.” The Journal of Family Practice 49 (2): 147–52. ries A: Biological Sciences and Medical Sciences 71 (Suppl 1): https://pubmed.ncbi.nlm.nih.gov/10718692/. S87–99. https://doi.org/10.1093/gerona/glv099. 14. Beutel, Manfred E., Elmar Brähler, Joerg Wiltink, Jasmin 26. Chapman, Christina Hunter, Gerard Heath, Pamela Fairchild, Ghaemi Kerahrodi, Juliane Burghardt, Matthias Michal, An- Mitchell B. Berger, Daniela Wittmann, Shitanshu Uppal, Anagha 250 CHAPTER 8 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E Tolpadi, Katherine Maturen, and Shruti Jolly. 2018. “Gynecolog- org/10.1177/019251302236598. ic Radiation Oncology Patients Report Unmet Needs Regard- 40. Dhar, H. L. 2001. “Gender, Aging, Health and Society.” The Jour- ing Sexual Health Communication with Providers.” Journal of nal of the Association of Physicians of India 49 (October): Cancer Research and Clinical Oncology 145 (2): 495–502. 1012–20. https://pubmed.ncbi.nlm.nih.gov/11848308/. https://doi.org/10.1007/s00432-018-2813-3. 41. Dorak, M. Tevfik, and Ebru Karpuzoglu. 2012. “Gender Differ- 27. Chawla, Mukesh, Gordon Betcherman, and Arup Banerji. 2007. ences in Cancer Susceptibility: An Inadequately Addressed From Red to Gray: The “Third Transition” of Aging Popula- Issue.” Frontiers in Genetics 3. https://doi.org/10.3389/ tions in Eastern Europe and the Former Soviet Union. World fgene.2012.00268. Bank. https://doi.org/10.1596/978-0-8213-7129-9. 42. Dubal, Dena B. 2020. “Chapter 16 - Sex Difference in Alzheimer’s 28. Cheung, Ethan Siu Leung, and Ada C. Mui. 2021. “Gender Vari- Disease: An Updated, Balanced and Emerging Perspective on ation and Late-Life Depression: Findings from a National Sur- Differing Vulnerabilities.” Edited by Rupert Lanzenberger, Georg vey in the USA.” Ageing International 48 (November): 263–80. S. Kranz, and Ivanka Savic. ScienceDirect. Elsevier. January 1, https://doi.org/10.1007/s12126-021-09471-5. 2020. https://www.sciencedirect.com/science/article/abs/pii/ 29. Clemens, Kristin K., Mark Woodward, Bruce Neal, and Bernard B9780444641236000187. Zinman. 2020. “Sex Disparities in Cardiovascular Outcome 43. European Network of National Human Rights Institutions Trials of Populations with Diabetes: A Systematic Review and (ENNHRI). 2017. “Overview: long-term care in Europe.” January 30, Meta-Analysis.” Diabetes Care 43 (5): 1157–63. https://doi. 2017. Available at: https://ennhri.org/news-and-blog/overview- org/10.2337/dc19-2257. long-term-care-in-europe/. [Accessed: December 12, 2022]. 30. Cook, Nakela L, John Z Ayanian, E. John Orav, and LeRoi S Hicks. 44. Eke, C. B., I. B. Omotowo, O. M. Ukoha, and B. C. Ibe. 2015. “Hu- 2009. “Differences in Specialist Consultations for Cardiovascular man Rabies: Still a Neglected Preventable Disease in Nigeria.” Disease by Race, Ethnicity, Gender, Insurance Status, and Site Nigerian Journal of Clinical Practice 18 (2): 268–72. https:// of Primary Care.” Circulation 119 (18): 2463–70. https://doi. doi.org/10.4314/njcp.v18i2. org/10.1161/circulationaha.108.825133. 45. El Halabi, Sarah, Rawad El Hayek, Karine Kahil, Marwa Nofal, 31. Colombo, F., et al., F. 2011. Help Wanted? Providing and Paying and Samer El Hayek. 2020. “Characteristics of Attempted Sui- for Long-Term Care. Paris, France: OECD Health Policy Studies, cide in the Middle East and North Africa Region: The Mediat- Organisation for Economic Co-operation and Development. ing Role of Arab Culture and Religion.” Mediterranean Jour- https://doi.org/10.1787/9789264097759-en. nal of Emergency Medicine & Acute Care 1 (3). https://doi. 32. Cortez, Rafael, Seemeen Saadat, Sadia Chowdhury, and Intissar org/10.52544/2642-7184(1)3002. Sarker. 2014. “Maternal and Child Survival: Findings from Five 46. Etienne, Carissa F. 2018. “Addressing Masculinity and Men’s Countries’ Experience in Addressing Maternal and Child Health Health to Advance Universal Health and Gender Equality.” Re- Challenges.” Health, Nutrition, and Population (HNP) Discus- vista Panamericana de Salud Pública 42 (December): 1–2. sion Paper. Washington, D.C.: World Bank. http://hdl.handle. https://doi.org/10.26633/rpsp.2018.196. net/10986/20757. 47. Fabreau, G. E., A. A. Leung, D. A. Southern, M. L. Knudtson, J. 33. Cotlear, Daniel. 2011. Population Aging: Is Latin America M. McWilliams, J. Z. Ayanian, and W. A. Ghali. 2014. “Sex, So- Ready? Directions in Development: Human Development. The cioeconomic Status, Access to Cardiac Catheterization, and World Bank. https://doi.org/10.1596/978-0-8213-8487-9. Outcomes for Acute Coronary Syndromes in the Context of 34. Cui, Yiran, Sumaira Mubarik, Ruijia Li, Nawsherwan, and Ch- Universal Healthcare Coverage.” Circulation: Cardiovascular uanhua Yu. 2021. “Trend Dynamics of Thyroid Cancer Incidence Quality and Outcomes 7 (4): 540–49. https://doi.org/10.1161/ among China and the U.S. Adult Population from 1990 to 2017: circoutcomes.114.001021. A Joinpoint and Age-Period-Cohort Analysis.” BMC Public 48. Frank, Richard G. 2012. “Long-Term Care Financing in the Unit- Health 21 (1). https://doi.org/10.1186/s12889-021-10635-w. ed States: Sources and Institutions.” Applied Economic Per- 35. Dawson, Steven, and Rick Surpin. 2001. “Direct-care healthcare spectives and Policy 34 (2): 333–45. https://www.jstor.org/ workers: You get what you pay for.” Generations 25(1): 23-28. stable/23273828. https://www.jstor.org/stable/26555063. 49. Gálvez Galán, Isabel, Macarena Celina Cáceres León, Jorge 36. de Silva, Sara Johansson, and Indhira Santos. 2024. “Produc- Guerrero-Martín, Casimiro Fermín López Jurado, and Noelia tive longevity: what can work in low- and middle- income Durán-Gómez. 2021. “Health-Related Quality of Life in Diabe- countries?” in Unlocking the Power of Healthy Longevity: tes Mellitus Patients in Primary Health Care.” Enfermería Clíni- Compendium of Research for the Healthy Longevity Initia- ca (English Edition) 31 (5): 313–22. https://doi.org/10.1016/j. tive.  Washington D.C.: World Bank. enfcle.2021.03.003. 37. Demarco, Gustavo, Johannes Koettl, Miglena Abels, and Andrea 50. Gatti Roberta, Daniel Halim, Allen Hardiman and Shuqiao Sun. Petrelli. 2024. “Adequacy pensions and access to healthcare: 2024. “Gendered Responsibilities, Elderly Care, and Labor Sup- maintaining human capital during old age,” in Unlocking the ply: Evidence from Four Countries,” in Unlocking the Power of Power of Healthy Longevity: Compendium of Research for Healthy Longevity: Compendium of Research for the Healthy the Healthy Longevity Initiative.  Washington D.C.: World Bank. Longevity Initiative.  Washington D.C.: World Bank. 38. Deng, Yuanyuan, Katja Hanewald, Yafei Si, and Melantha Wang. 51. Getachew, Sefonias, Eyerusalem Getachew, Muluken Gizaw, 2022. “Gender Effect in Long-Term Care: Evidence from China.” Wondimu Ayele, Adamu Addissie, and Eva J. Kantelhardt. 2019. Sydney: ARC Centre of Excellence in Population Ageing Re- “Cervical Cancer Screening Knowledge and Barriers among search (CEPAR), UNSW Sydney. https://www.netspar.nl/assets/ Women in Addis Ababa, Ethiopia.” Edited by Violet Naanyu. uploads/39_Deng.pdf. PLOS ONE 14 (5): e0216522. https://doi.org/10.1371/journal. 39. Dentinger, Emma, and Marin Clarkberg. 2002. “Informal pone.0216522. Caregiving and Retirement Timing among Men and Wom- 52. Geyer, Johannes. 2019. “Notes about Comparing Long-Term en.” Journal of Family Issues 23 (7): 857–79. https://doi. Care Expenditures across Countries Comment on ‘Financing COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 8 251 Long-Term Care: Lessons from Japan.’” International Journal grant Home Care Workers Caring for Older People: Fictive Kin, of Health Policy and Management 9 (2): 80–82. https://doi. Substitute, and Complementary Family Caregivers in an Ethnically org/10.15171/ijhpm.2019.87. Diverse Environment.” International Journal of Ageing and Later 53. Gong, Ni, Ya Meng, Qin Hu, Qianqian Du, Xiaoyu Wu, Wenjie Zou, Life 5 (2): 7–16. https://doi.org/10.3384/ijal.1652-8670.10527. Mengyao Zhu, et al. 2022. “Obstacles to Access to Community 65. Holme, Francesca, Sharon Kapambwe, Ashrafun Nessa, Partha Care in Urban Senior-Only Households: A Qualitative Study.” BMC Basu, Raul Murillo, and Jose Jeronimo. 2017. “Scaling up Proven Geriatrics 22 (1). https://doi.org/10.1186/s12877-022-02816-y. Innovative Cervical Cancer Screening Strategies: Challenges 54. Griffin, Barbara, Vanessa Loh, and Beryl Hesketh. 2012. “Age, and Opportunities in Implementation at the Population Level Gender, and the Retirement Process.” In The Oxford Handbook in Low- and Lower-Middle-Income Countries.” International of Retirement, edited by Mo Wang, 202–14. New York: Oxford Journal of Gynecology & Obstetrics 138 (July): 63–68. https:// University Press. doi.org/10.1002/ijgo.12185. 55. Glinskaya, Elena, Xiaohui Hou, Zhanlian Feng, Marco Angrisani, 66. Hornby-Turner, Yvonne Claire, Nancye May Peel, and Ruth Elea- Guadalupe Suarez, Jigyasa Sharma, Drystan Phillips, et al. 2024. nor Hubbard. 2017. “Health Assets in Older Age: A Systematic “Demand for and Supply of Long-Term Care for Older persons Review.” BMJ Open 7 (5): e013226. https://doi.org/10.1136/ in Low- and Middle-Income Countries,” in Unlocking the Pow- bmjopen-2016-013226. er of Healthy Longevity: Compendium of Research for the 67. Hossain, Babul, K.S. James, Varsha P. Nagargoje, and Papai Bar- Healthy Longevity Initiative.  Washington D.C.: World Bank. man. 2021. “Differentials in Private and Public Healthcare Ser- 56. Guerra, Ricardo O., Beatriz Eugenia Alvarado, and Maria Victoria vice Utilization in Later Life: Do Gender and Marital Status Have Zunzunegui. 2008. “Life Course, Gender and Ethnic Inequalities Any Association?” Journal of Women & Aging 35 (2): 1–11. in Functional Disability in a Brazilian Urban Elderly Population.” https://doi.org/10.1080/08952841.2021.2011562. Aging Clinical and Experimental Research 20 (1): 53–61. 68. Hosseinpoor, Ahmad Reza, Jennifer Stewart Williams, Avni https://doi.org/10.1007/bf03324748. Amin, Islene Araujo de Carvalho, John Beard, Ties Boerma, Paul 57. Haldane, Victoria, Gisela M. Garcia, Tahir Bockarie, Daphne Kowal, Nirmala Naidoo, and Somnath Chatterji. 2012. “Social Wu, Cristian A Herrera, Maria Luisa Latorre Castro, Debapriya Determinants of Self-Reported Health in Women and Men: Chakraborty, Beverly Essue, Prabhat Jha, and Jeremy Veillard. Understanding the Role of Gender in Population Health.” Ed- 2024. “Healthy Longevity Initiative: A Performance Dashboard ited by Beverley J. Shea. PLoS ONE 7 (4): e34799. https://doi. for Decision-making in Low- and Middle-income Countries,” org/10.1371/journal.pone.0034799. in Unlocking the Power of Healthy Longevity: Compendium 69. Hossen, Abul, and Anne Westhues. 2010. “A Socially Excluded of Research for the Healthy Longevity Initiative.  Washington Space: Restrictions on Access to Health Care for Older Wom- D.C.: World Bank. en in Rural Bangladesh.” Qualitative Health Research 20 (9): 58. Hamada, O., Tsutsumi, T., Miki, A., Fukui, T., Shimokawa, T. and 1192–1201. https://doi.org/10.1177/1049732310370695. Imanaka, Y., 2020. Impact of the Hospitalist System in Japan on 70. Hou, Xiaohui, Jigyasa Sharma, and Feng Zhao, eds. 2023. Silver the Quality of Care and Healthcare Economics. Internal Medi- Opportunity: Building Integrated Services for Older Adults cine, pp.2872-19. around Primary Health Care. Washington, D.C.: World Bank. 59. Hamiduzzaman, Mohammad, Anita De Bellis, Wendy Abigail, http://hdl.handle.net/10986/39422. and Evdokia Kalaitzidis. 2017. “The Social Determinants of 71. Hughes, Anne K., and Terri D. W. Lewinson. 2014. “Facilitating Healthcare Access for Rural Elderly Women - a Systematic Re- Communication about Sexual Health between Aging Women view of Quantitative Studies.” The Open Public Health Journal and Their Health Care Providers.” Qualitative Health Research 10 (1): 244–66. https://doi.org/10.2174/1874944501710010244. 25 (4): 540–50. https://doi.org/10.1177/1049732314551062. 60. Hatano, Yutaka, Masanori Mori, Hiroaki Izumi, Koji Amano, Tet- 72. IHME (Institute for Health Metrics and Evaluation). 2020. Global suya Ito, Junko Nozato, Keisuke Kaneishi, Tomohiro Kawamura, Burden of Disease Study 2019 (GBD 2019) Results. Global Burden and Tatsuya Morita. 2022. “End-of-Life Experiences in Advanced of Disease Collaborative Network. Seattle, United States: Institute Cancer: Gender Differences.” BMJ Supportive & Palliative Care, for Health Metrics and Evaluation. Available from https://vizhub. June. https://doi.org/10.1136/bmjspcare-2022-003761. healthdata.org/gbd-results/. [Accessed: November 16, 2022]. 61. Heise, Lori, Margaret E Greene, Neisha Opper, Maria Stavropou- 73. ILO (International Labor Organization). 2018. “Women do 4 lou, Caroline Harper, Marcos Nascimento, Debrework Zewdie, times more unpaid care work than men in Asia and the Pacific.” et al. 2019. “Gender Inequality and Restrictive Gender Norms: https://www.ilo.org/asia/media-centre/news/WCMS_633284/ Framing the Challenges to Health.” The Lancet 393 (10189): lang--en/index.htm 2440–54. https://doi.org/10.1016/s0140-6736(19)30652-x. 74. ILO/WHO (International Labor Organization/World Health 62. Hess, Rachel, Rebecca C. Thurston, Ron D. Hays, Chung-Chou Organization). 2022. “The Gender Pay Gap in the Health and H. Chang, Stacey N. Dillon, Roberta B. Ness, Cindy L. Bryce, Care Sector a Global Analysis in the Time of COVID-19.” Gene- Wishwa N. Kapoor, and Karen A. Matthews. 2012. “The Impact va: World Health Organization and the International Labour of Menopause on Health-Related Quality of Life: Results from Organization; https://www.who.int/publications-detail-redi- the STRIDE Longitudinal Study.” Quality of Life Research 21 (3): rect/9789240052895. 535–44. https://doi.org/10.1007/s11136-011-9959-7. 75. Jamuna, D. 2003. “Issues of Elder Care and Elder Abuse in the 63. Hickey, Sebastian Martinez, Marokey Sawo, and Julia Wolfe. Indian Context.” Journal of Aging & Social Policy 15 (2-3): 2022. “The State of the Residential Long-Term Care Industry: 125–42. https://doi.org/10.1300/j031v15n02_08. A Comprehensive Look at Employment Levels, Demographics, 76. Jeon, G.-S., S.-N. Jang, D.-S. Kim, and S.-I. Cho. 2013. “Widow- Wages, Benefits, and Poverty Rates of Workers in the Industry.” hood and Depressive Symptoms among Korean Elders: The Washington, D.C.: Economic Policy Institute. https://www.epi. Role of Social Ties.” The Journals of Gerontology Series B: org/publication/residential-long-term-care-workers/. Psychological Sciences and Social Sciences 68 (6): 963–73. 64. Hoff, Andreas, Susan Feldman, and Lucie Vidovicova. 2011. “Mi- https://doi.org/10.1093/geronb/gbt084. 252 CHAPTER 8 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E 77. Jha, Prabhat. 2020. “The Hazards of Smoking and the Bene- 91. Lindahl, Anna, Miia Aro, Jere Reijula, Mika J. Mäkelä, Jukka fits of Cessation: A Critical Summation of the Epidemiological Ollgren, Mervi Puolanne, Asko Järvinen, and Tuula Vasankari. Evidence in High-Income Countries.” ELife 9 (1). https://doi. 2022b. “Women Report More Symptoms and Impaired Quality org/10.7554/elife.49979. of Life: A Survey of Finnish COVID-19 Survivors.” Infectious Dis- 78. Juutilainen, Auni, Saara Kortelainen, Seppo Lehto, Tapani eases (London, England) 54 (1): 53–62. https://doi.org/10.108 Ronnemaa, Kalevi Pyörälä, and Markku Laakso. 2004. “Gender 0/23744235.2021.1965210. Difference in the Impact of Type 2 Diabetes on Coronary Heart 92. Lipsky, Martin S., Sharon Su, Carlos J. Crespo, and Man Hung. 2021. Disease Risk.” Diabetes Care 27 (12): 2898–2904. https://doi. “Men and Oral Health: A Review of Sex and Gender Differences.” org/10.2337/diacare.27.12.2898. American Journal of Men’s Health 15 (3): 155798832110163. 79. Karpinska, K., Dykstra, P.A., van den Broek, T., Djundeva, M., https://doi.org/10.1177/15579883211016361. Abramowska-Kmon, A., Kotowska, I.I. and Marí-Klose, P. 2016. 93. Liu, Zuyun, Brian H. Chen, Themistocles L. Assimes, Luigi Ferrucci, “Intergenerational linkages in the family: The organization of Steve Horvath, and Morgan E. Levine. 2019. “The Role of Epigen- caring and financial responsibilities: Summary of results.” Fam- etic Aging in Education and Racial/Ethnic Mortality Disparities iliesAndSocieties, Working Paper Series, 61. among Older U.S. Women.” Psychoneuroendocrinology 104 80. Kautzky-Willer, Alexandra. 2021. “Does Diabetes Mellitus (June): 18–24. https://doi.org/10.1016/j.psyneuen.2019.01.028. Mitigate the Gender Gap in COVID-19 Mortality?” Europe- 94. Maas, A.H.E.M., and Y.E.A. Appelman. 2010. “Gender Differences an Journal of Endocrinology 185 (5): C13–17. https://doi. in Coronary Heart Disease.” Netherlands Heart Journal 18 (12): org/10.1530/eje-21-0721. 598–603. https://doi.org/10.1007/s12471-010-0841-y. 81. Keck, W. and Saraceno, C. 2009. “Balancing elderly care and em- 95. Madonis, Stephanie M, Kimberly A Skelding, and Madhur ployment in Germany.” (No. SP I 2009-401). WZB Discussion Paper. Roberts. 2017. “Management of Acute Coronary Syndromes: 82. Kiely, Kim M., Brooke Brady, and Julie Byles. 2019. “Gender, Special Considerations in Women.” Heart 103 (20): 1638–46. Mental Health and Ageing.” Maturitas 129 (November): 76–84. https://doi.org/10.1136/heartjnl-2016-309938. https://doi.org/10.1016/j.maturitas.2019.09.004. 96. Mascara, Mariah, and Constantina Constantinou. 2021. “Glob- 83. Krämer, Heike U, Gernot Rüter, Ben Schöttker, Dietrich Rothen- al Perceptions of Women on Breast Cancer and Barriers to bacher, Thomas Rosemann, Joachim Szecsenyi, Hermann Screening.” Current Oncology Reports 23 (7). https://doi. Brenner, and Elke Raum. 2012. “Gender Differences in Health- org/10.1007/s11912-021-01069-z. care Utilization of Patients with Diabetes.” The American Jour- 97. McGraw, Jacquie, Katherine M. White, and Rebekah Rus- nal of Managed Care 18 (7): 362–69. https://europepmc.org/ sell-Bennett. 2021. “Masculinity and Men’s Health Service Use article/med/22823530. across Four Social Generations: Findings from Australia’s Ten to 84. Kraus, Markus, Monika Riedel, Monika , Esther Mot, Peter Wil- Men Study.” SSM - Population Health 15 (September): https:// leme, Gerald Röhrling and Thomas Czypionka. 2010. A Typol- doi.org/10.1016/j.ssmph.2021.100838. ogy of Long-Term Care Systems in Europe. ENEPRI Research 98. McMaughan, Darcy Jones, Oluyomi Oloruntoba, and Matthew Report No. 91, August 2010. https://aei.pitt.edu/32248/. Lee Smith. 2020. “Socioeconomic Status and Access to Health- 85. Kuruvilla, Shyama, Julian Schweitzer, David Bishai, Sadia Chow- care: Interrelated Drivers for Healthy Aging.” Frontiers in Public dhury, Daniele Caramani, Laura Frost, Rafael Cortez, et al. 2014. Health 8 (231). https://doi.org/10.3389/fpubh.2020.00231. “Success Factors for Reducing Maternal and Child Mortality.” 99. Mergl, Roland, Nicole Koburger, Katherina Heinrichs, András Széke- Bulletin of the World Health Organization 92 (7): 533–44. ly, Mónika Ditta Tóth, James Coyne, Sónia Quintão, et al. 2015. https://doi.org/10.2471/blt.14.138131. “What Are Reasons for the Large Gender Differences in the Lethal- 86. Leeson, George and Sarah Harper. 2006. “The Status Quo of ity of Suicidal Acts? An Epidemiological Analysis in Four European Foreign Workers in the Health and Social Care Sector in the UK, Countries.” Edited by Thomas Niederkrotenthaler. PLOS ONE 10 (7): Germany and the Netherlands.” Japan National Council of So- e0129062. https://doi.org/10.1371/journal.pone.0129062. cial Welfare, Tokyo. ISBN4-9903255-0-8. 100. Miszkurka, Malgorzata, Slim Haddad, Étienne V Langlois, Ellen E 87. Lena, A, K Ashok, M Padma, V Kamath, and A Kamath. 2009. Freeman, Seni Kouanda, and Maria Victoria Zunzunegui. 2012. “Health and Social Problems of the Elderly: A Cross-Sectional “Heavy Burden of Non-Communicable Diseases at Early Age Study in Udupi Taluk, Karnataka.” Indian Journal of Community and Gender Disparities in an Adult Population of Burkina Faso: Medicine 34 (2): 131. https://doi.org/10.4103/0970-0218.51236. World Health Survey.” BMC Public Health 12 (1). https://doi. 88. Li, Wenshan, Douglas G. Manuel, Sarina R. Isenberg, and Pe- org/10.1186/1471-2458-12-24. ter Tanuseputro. 2022. “Caring for Older Men and Women: 101. Mobley, L.R., Root, E.D., Finkelstein, E.A., Khavjou, O., Farris, R.P. Whose Caregivers Are More Distressed? A Population-Based and Will, J.C. 2006. “Environment, Obesity, and Cardiovascu- Retrospective Cohort Study.” BMC Geriatrics 22 (1). https://doi. lar Disease Risk in Low-Income Women.” American Journal org/10.1186/s12877-022-03583-6. of Preventive Medicine, 30(4), pp. 327-332.e1. https://doi. 89. Lin, Hongyan, Mengdi Jin, Qian Liu, Yue Du, Jingzhu Fu, org/10.1016/j.amepre.2005.12.001. Changqing Sun, Fei Ma, et al. 2021. “Gender-Specific Preva�- 102. Moen, Phyllis, Julie Robison, and Vivian Fields. 1994. “Women’s lence and Influencing Factors of Depression in Elderly in Rural Work and Caregiving Roles: A Life Course Approach.” Journal China: A Cross-Sectional Study.” Journal of Affective Disorders of Gerontology 49 (4): S176–86. https://doi.org/10.1093/ 288 (June): 99–106. https://doi.org/10.1016/j.jad.2021.03.078. geronj/49.4.s176. 90. Lindahl, Anna, Miia Aro, Jere Reijula, Mika J. Mäkelä, Jukka 103. Montgomery, Rhonda J. V., Lyn Holley, Jerome Deichert, and Ollgren, Mervi Puolanne, Asko Järvinen, and Tuula Vasankari. Karl Kosloski. 2005. “A Profile of Home Care Workers from the 2022a. “Women Report More Symptoms and Impaired Quality 2000 Census: How It Changes What We Know.” The Gerontolo- of Life: A Survey of Finnish COVID-19 Survivors.” Infectious Dis- gist 45 (5): 593–600. https://doi.org/10.1093/geront/45.5.593. eases (London, England) 54 (1): 53–62. https://doi.org/10.108 104. Muir, Tim. 2017. “Measuring Social Protection for Long-Term 0/23744235.2021.1965210. Care.” OECD I Library. Paris. March 27, 2017. https://www. COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 8 253 oecd-ilibrary.org/social-issues-migration-health/measuring- ment). 2011. “Help Wanted?: Providing and Paying for Long- social-protection-for-long-term-care_a411500a-en. Term Care.” Paris: OECD Publishing. https://www.oecd.org/els/ 105. Muñoz Boudet, Ana María, Patti Petesch, Carolyn Turk, and health-systems/help-wanted-9789264097759-en.htm. Angélica Thumala. 2013. On Norms and Agency: Conversa- 117. _______. 2013. “Health at a Glance 2013: OECD Indicators.” tions about Gender Equality with Women and Men in 20 Paris: OECD Publishing. https://www.oecd.org/els/health-sys- Countries. Washington, D.C.: World Bank. http://hdl.handle. tems/Health-at-a-Glance-2013.pdf. net/10986/13818. 118. _______. 2019. “Women Are Well-Represented in Health 106. Murata, Chiyoe, Tetsuji Yamada, Chia-Ching Chen, Toshiyuki and Long-Term Care Professions, but Often in Jobs with Poor Ojima, Hiroshi Hirai, and Katsunori Kondo. 2010. “Barriers to Working Conditions,” OECD Gender Equality Data. March 2019. Health Care among the Elderly in Japan.” International Journal https://www.oecd.org/gender/data/women-are-well-repre- of Environmental Research and Public Health 7 (4): 1330–41. sented-in-health-and-long-term-care-professions-but-often- https://doi.org/10.3390/ijerph7041330. in-jobs-with-poor-working-conditions.htm. 107. Muñoz, Daniel Cobos, Carmen Sant, Rebeca Revenga Becedas, 119. _______. 2020. “Who Cares? Attracting and Retaining Care and Doris Ma Fat. 2020. “Dangers of Gender Bias in CRVS and Workers for the Elderly.” OECD Health Policy Studies. Par- Cause of Death Data : The Path to Health Inequality.”  Brief 3. is: OECD Publishing. https://www.oecd.org/publications/ Paper 3. Canada: International Development Research Centre.  who-cares-attracting-and-retaining-elderly-care-workers- http://hdl.handle.net/10625/60102. 92c0ef68-en.htm. 108. Naud, Daniel, Mélissa Généreux, Jean-François Bruneau, Aline 120. OECD (Organization for Economic Cooperation and Devel- - Alauzet, and Mélanie Levasseur. 2019. “Social Participation in Ol� opment)/European Union. 2013. “A Good Life in Old Age? : der Women and Men: Differences in Community Activities and Monitoring and Improving Quality in Long-Term Care.” OECD Barriers according to Region and Population Size in Canada.” BMC Health Policy Studies. Paris: OECD Publishing. https://doi. Public Health 19 (1). https://doi.org/10.1186/s12889-019-7462-1. org/10.1787/9789264194564-en. 109. Ngai, L. Rachel and Christopher A. Pissarides. 2009. “Welfare 121. Oertelt-Prigione, Sabine, Belle H. de Rooij, Floortje Mols, Simone policy and the distribution of hours of work.” CEP Discussion Oerlemans, Olga Husson, Dounya Schoormans, John B. Haanen, Paper (962). London School of Economics and Political Science. and Lonneke V. van de Poll-Franse. 2021. “Sex-Differences in Centre for Economic Performance, London, UK. https://eprints. Symptoms and Functioning in >5000 Cancer Survivors: Results lse.ac.uk/id/eprint/28698. from the PROFILES Registry.” European Journal of Cancer 156 110. Nilsson, Anna G., Daniel Sundh, Lisa Johansson, Martin Nils- (October): 24–34. https://doi.org/10.1016/j.ejca.2021.07.019. son, Dan Mellström, Robert Rudäng, Michail Zoulakis, Märit 122. Östlin, Piroska. 2009. “Transforming Health Systems and Wallander, Anna Darelid, and Mattias Lorentzon. 2017. “Type 2 Services for Women and Girls”. Paper prepared for the Expert Diabetes Mellitus Is Associated with Better Bone Microarchi- Group Meeting on The Impact of the Implementation of the tecture but Lower Bone Material Strength and Poorer Physical Beijing Declaration and Platform for Action on the Achieve- Function in Elderly Women: A Population-Based Study.” Jour- ment of the Millennium Development Goals, 11-13 Novem- nal of Bone and Mineral Research: The Official Journal of ber 2009, United Nations, New York. the American Society for Bone and Mineral Research 32 (5): 123. Pagidipati, Neha J., and Eric D. Peterson. 2016. “Acute Coronary 1062–71. https://doi.org/10.1002/jbmr.3057. Syndromes in Women and Men.” Nature Reviews. Cardiology 111. Novak, Josh R., Terry Peak, Julie Gast, and Melinda Arnell. 2019. 13 (8): 471–80. https://doi.org/10.1038/nrcardio.2016.89. “Associations between Masculine Norms and Health-Care 124. Palta, Priya, Sarah L. Szanton, Richard D. Semba, Roland J. Thor- Utilization in Highly Religious, Heterosexual Men.” American pe, Ravi Varadhan, and Linda P. Fried. 2015. “Financial Strain Is Journal of Men’s Health 13 (3): 155798831985673. https://doi. Associated with Increased Oxidative Stress Levels: The Wom- org/10.1177/1557988319856739. en’s Health and Aging Studies.” Geriatric Nursing 36 (2): S33– 112. Nummela, Olli P., Tommi T. Sulander, Heikki S. Heinonen, and 37. https://doi.org/10.1016/j.gerinurse.2015.02.020. Antti K. Uutela. 2007. “Self-Rated Health and Indicators of SES 125. Pan, An, Michel Lucas, Qi Sun, Rob M. van Dam, Oscar H. Fran- among the Ageing in Three Types of Communities.” Scandi- co, Walter C. Willett, JoAnn E. Manson, Kathryn M. Rexrode, Al- navian Journal of Public Health 35 (1): 39–47. https://doi. berto Ascherio, and Frank B. Hu. 2011. “Increased Mortality Risk org/10.1080/14034940600813206. in Women with Depression and Diabetes Mellitus.” Archives of 113. O’Hanlan, Katherine A., Suzanne L. Dibble, H. Jennifer J. Ha- General Psychiatry 68 (1): 42. https://doi.org/10.1001/arch- gan, and Rachel Davids. 2004. “Advocacy for Women’s Health genpsychiatry.2010.176. Should Include Lesbian Health.” Journal of Women’s Health 13 126. Papadopoulou, Sofia A., and Juan Carlos Kaski. 2013. “Isch- (2): 227–34. https://doi.org/10.1089/154099904322966218. aemic Heart Disease in the Ageing Woman.” Best Practice & 114. O’Keefe, Philip, and Victoria Haldane. 2024. “Towards a frame- Research Clinical Obstetrics & Gynaecology 27 (5): 689–97. work for impact pathways between non-communicable dis- https://doi.org/10.1016/j.bpobgyn.2013.03.003. eases, human capital and healthy longevity, economic and 127. Perkins, Jessica M., Hwa-young Lee, K. S. James, Juhwan Oh, wellbeing outcomes,” in Unlocking the Power of Healthy Lon- Aditi Krishna, Jongho Heo, Jong-koo Lee, and S. V. Subrama- gevity: Compendium of Research for the Healthy Longevity nian. 2016. “Marital Status, Widowhood Duration, Gender and Initiative.  Washington D.C.: World Bank. Health Outcomes: A Cross-Sectional Study among Older Adults 115. Obuobi-Donkor, Gloria, Nnamdi Nkire, and Vincent I. O. Agyapong. in India.” BMC Public Health 16 (1). https://doi.org/10.1186/ 2021. “Prevalence of Major Depressive Disorder and Correlates of s12889-016-3682-9. Thoughts of Death, Suicidal Behaviour, and Death by Suicide in 128. Peters, Sanne A E, and Mark Woodward. 2022. “Sex and Gen- the Geriatric Population—a General Review of Literature.” Behav- der Matter in Cardiovascular Disease and Beyond.” Heart 108 ioral Sciences 11 (11): 142. https://doi.org/10.3390/bs11110142. (13): heartjnl-2021-320719. https://doi.org/10.1136/heart- 116. OECD (Organization for Economic Cooperation and Develop- jnl-2021-320719. 254 CHAPTER 8 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E 129. Pilania, Manju, Vikas Yadav, Mohan Bairwa, Priyamadhaba Behe- al Labour Organization. https://ideas.repec.org/p/ilo/ilo- ra, Shiv Dutt Gupta, Hitesh Khurana, Viswanathan Mohan, Gir- wps/994886493402676.html. ish Baniya, and S. Poongothai. 2019. “Prevalence of Depression 143. Schmitz, Hendrik, and Matthias Westphal. 2017. “Informal Care among the Elderly (60  Years and Above) Population in India, and Long-Term Labor Market Outcomes.” Journal of Health 1997–2016: A Systematic Review and Meta-Analysis.” BMC Pub- Economics, 56 (December): 1–18. https://doi.org/10.1016/j. lic Health 19 (1). https://doi.org/10.1186/s12889-019-7136-z. jhealeco.2017.09.002. 130. Pirkis, Jane, Matthew J Spittal, Louise Keogh, Tass Mousaferi- 144. Song, Yan, and Ying Bian. 2014. “Gender Differences in the adis, and Dianne Currier. 2017. “Masculinity and Suicidal Think- Use of Health Care in China: Cross-Sectional Analysis.” Inter- ing.” Social Psychiatry and Psychiatric Epidemiology 52 (3): national Journal for Equity in Health 13 (1): 8. https://doi. 319–27. https://doi.org/10.1007/s00127-016-1324-2. org/10.1186/1475-9276-13-8. 131. Polverini, Francesca and Giovanni Lamura. 2005. “Labour Sup- 145. Srivastava, Shobhit, Paramita Debnath, Neha Shri, and T. Mu- ply in Care Services” National Report on Italy. https://api.se- hammad. 2021. “The Association of Widowhood and Living manticscholar.org/CorpusID:201697054. Alone with Depression among Older Adults in India.” Scientific 132. Potter, Andrew J. 2017. “Care Configurations and Un- Reports 11 (1). https://doi.org/10.1038/s41598-021-01238-x. met Care Needs in Older Men and Women.” Journal 146. Stankuniene, Aurima, Mindaugas Stankunas, Mark Avery, Jutta of Applied Gerontology 38 (10): 1351–70. https://doi. Lindert, Rita Mikalauskiene, Maria Gabriella Melchiorre, Francis- org/10.1177/0733464817733239. co Torres-Gonzalez, et al. 2015. “The Prevalence of Self-Report- 133. Prus, Steven G., and Ellen Gee. 2003. “Gender Differences in the ed Underuse of Medications due to Cost for the Elderly: Results Influence of Economic, Lifestyle, and Psychosocial Factors on from Seven European Urban Communities.” BMC Health Ser- Later-Life Health.” Canadian Journal of Public Health 94 (4): vices Research 15 (1). https://doi.org/10.1186/s12913-015- 306–9. https://doi.org/10.1007/bf03403611. 1089-4. 134. Quick, Jonathan, Jonathan Jay, and Ana Langer. 2014. “Improv- 147. Stone, Robyn I., and Steven L. Dawson. 2008. “The Origins of ing Women’s Health through Universal Health Coverage.” PLoS Better Jobs Better Care.” The Gerontologist 48 (suppl 1): 5–13. Medicine 11 (1): e1001580. https://doi.org/10.1371/journal. https://doi.org/10.1093/geront/48.supplement_1.5. pmed.1001580. 148. Sun, Tao, Shu-E Zhang, Meng-yao Yan, Ting-hui Lian, Yi-qi Yu, 135. Rahbari, Reza, Lisa Zhang, and Electron Kebebew. 2010. “Thy- Hong-yan Yin, Chen-xi Zhao, et al. 2022. “Association between roid Cancer Gender Disparity.” Future Oncology 6 (11): 1771– Self-Perceived Stigma and Quality of Life among Urban Chi- 79. https://doi.org/10.2217/fon.10.127. nese Older Adults: The Moderating Role of Attitude toward 136. Rivera, Ashley, and Jonas Scholar. 2020. “Traditional Masculin- Own Aging and Traditionality.” Frontiers in Public Health 10 ity: A Review of Toxicity Rooted in Social Norms and Gender (February). https://doi.org/10.3389/fpubh.2022.767255. Socialization.” Advances in Nursing Science 43 (1): E1–10. 149. Sunwoo, Duk. 2017. “Public Long-Term Care Insurance Program https://doi.org/10.1097/ans.0000000000000284. for the Elderly (LTCI) - Performance Evaluation and Policy Impli- 137. Rivera-Franco, Monica M, and Eucario Leon-Rodriguez. cations.” Sejong city, Korea: Korea Institute for Health and Social 2018. “Delays in Breast Cancer Detection and Treatment Affairs (KIHASA). in Developing Countries.” Breast Cancer: Basic and Clini- 150. Suteau, Valentine, Mathilde Munier, Claire Briet, and Patrice Ro- cal Research 12 (January): 117822341775267. https://doi. dien. 2021. “Sex Bias in Differentiated Thyroid Cancer.” Interna- org/10.1177/1178223417752677. tional Journal of Molecular Sciences 22 (23): 12992. https:// 138. Robertson, Josefina. 2014. “Waiting Time at the Emergency doi.org/10.3390/ijms222312992. Department from a Gender Equality Perspective.” MSc. Thesis, 151. Syed, Samina T., Ben S. Gerber, and Lisa K. Sharp. 2013. “Trav- Institute of Medicine at the Sahlgrenska Academy, Univer- eling towards Disease: Transportation Barriers to Health Care sity of Gothenburg. https://gupea.ub.gu.se/bitstream/han- Access.” Journal of Community Health 38 (5): 976–93. https:// dle/2077/39196/gupea_2077_39196_1.pdf;jsessionid=005E- doi.org/10.1007/s10900-013-9681-1. 6B1A7CAB3E2C8DADD1F49AE79ADB?sequence=1. 152. Szanton, Sarah L., Jerelyn K. Allen, Roland J. Thorpe, Teresa See- 139. Rocks, Stephen, Daniela Berntson, Alejandro Gil-Salmerón, man, Karen Bandeen-Roche, and Linda P. Fried. 2008. “Effect Mudathira Kadu, Nieves Ehrenberg, Viktoria Stein, and Apos- of Financial Strain on Mortality in Community-Dwelling Older tolos Tsiachristas. 2020. “Cost and Effects of Integrated Care: A Women.” The Journals of Gerontology Series B: Psychological Systematic Literature Review and Meta-Analysis.” The Europe- Sciences and Social Sciences 63 (6): S369–74. https://doi. an Journal of Health Economics 21 (8): 1211–21. https://doi. org/10.1093/geronb/63.6.s369. org/10.1007/s10198-020-01217-5. 153. Thielke, Stephen, and Paula Diehr. 2012. “Transitions among 140. Rungreangkulkij, Somporn, Ingkata Kotnara, Nilubol Ruji- Health States Using 12 Measures of Successful Aging in raprasert, and Napaphat Khuandee. 2019. “Gender Inequality Men and Women: Results from the Cardiovascular Health Identified as an Underlying Cause of Depression in Thai Wom- Study.” Journal of Aging Research 2012: 1–9. https://doi. en.” Journal of International Women’s Studies 20 (7): 395– org/10.1155/2012/243263. 408. https://vc.bridgew.edu/jiws/vol20/iss7/25/. 154. Thorpe, Roland James, Judith D. Kasper, Sarah L. Szanton, 141. Sanjose, Silvia de, and Vivien Tsu. 2019. “Prevention of Cervi- Kevin D. Frick, Linda P. Fried, and Eleanor M. Simonsick. 2008. cal and Breast Cancer Mortality in Low- and Middle-Income “Relationship of Race and Poverty to Lower Extremity Function Countries: A Window of Opportunity.” International Journal and Decline: Findings from the Women’s Health and Aging of Women’s Health Volume 11 (July): 381–86. https://doi. Study.” Social Science & Medicine 66 (4): 811–21. https://doi. org/10.2147/ijwh.s197115. org/10.1016/j.socscimed.2007.11.005. 142. Scheil-Adlung, Xenia. 2015. “Long-term Care Protection for 155. U.S. Department of Health and Human Services, Health Re- Older Persons: A Review of Coverage Deficits in 46 coun- sources and Services Administration, National Center for tries.” ILO Working Papers. Geneva, Switzerland: Internation- Health Workforce Analysis. 2017. Long-Term Services and COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 8 255 Supports: Direct Care Worker Demand Projections, 2015-2030. E. Egede. 2017. “Sex Differences in Healthcare Expenditures Rockville, Maryland. among Adults with Diabetes: Evidence from the Medical Ex- 156. United Nations, Department of Economic and Social Affairs, penditure Panel Survey, 2002–2011.” BMC Health Services Population Division (2020). World Population Ageing 2019 (ST/ Research 17 (1). https://doi.org/10.1186/s12913-017-2178-3. ESA/SER.A/444). 169. Witters, Dan. 2011. “Caregiving Costs U.S. Economy $25.2 Billion 157. Vaccarino, Viola, Lori Parsons, Eric D. Peterson, William J. Rog- in Lost Productivity.” Gallup. July 27, 2011. https://news.gallup. ers, Catarina I. Kiefe, and John Canto. 2009. “Sex Differences com/poll/148670/caregiving-costs-economy-billion-lost-pro- in Mortality after Acute Myocardial Infarction.” Archives of ductivity.aspx. Internal Medicine 169 (19). https://doi.org/10.1001/archin- 170. World Bank. 2012. “World Development Report 2012: Gender ternmed.2009.332. Equality and Development.” Washington, D.C.: World Bank. 158. Van Houtven, Courtney Harold, Nicole DePasquale, and Norma http://hdl.handle.net/10986/4391. B. Coe. 2020. “Essential Long‐Term Care Workers Common- 171. _______. 2015. “World Bank Group Gender Strategy 2016- ly Hold Second Jobs and Double‐ or Triple‐Duty Caregiving 2023: Gender Equality, Poverty Reduction, and Inclusive Roles.” Journal of the American Geriatrics Society 68 (8): Growth.” Washington, D.C.: World Bank. http://hdl.handle. 1657–60. https://doi.org/10.1111/jgs.16509. net/10986/23425 159. Vogel, Birgit, Monica Acevedo, Yolande Appelman, C Noel 172. _______. 2016. “Live Long and Prosper: Aging in East Asia and Bairey Merz, Alaide Chieffo, Gemma A Figtree, Mayra Guerre- Pacific.”World Bank East Asia and Pacific Regional Report. Wash- ro, et al. 2021. “The Lancet Women and Cardiovascular Disease ington, D.C.: World Bank. http://hdl.handle.net/10986/23133. Commission: Reducing the Global Burden by 2030.” The Lancet 173. _______. 2017. “Women, Business, and the Law: Saving for 397 (10292). https://doi.org/10.1016/s0140-6736(21)00684-x. Old Age.” Washington, D.C.: World Bank. https://thedocs.world- 160. Wakabayashi, Chizuko, and Katharine M. Donato. 2006. “Does bank.org/en/doc/121231541445506749-0050022018/origi- Caregiving Increase Poverty among Women in Later Life? nal/WBLSavingForOldAgeFINALWEB.pdf. Evidence from the Health and Retirement Survey.” Journal 174. WHO (World Health Organization). 2007. Women, Ageing and of Health and Social Behavior 47 (3): 258–74. https://doi. Health: A Framework for Action. Geneva, Switzerland: World org/10.1177/002214650604700305. Health Organization. https://www.who.int/publications/i/ 161. Wang, Yingqi, and Tao Liu. 2020. “The ‘Silent Reserves’ of the Pa- item/9789241563529. triarchal Chinese Welfare System: Women as ‘Hidden’ Contribu- 175. _______. 2015. World Report on Ageing and Health. Gene- tors to Chinese Social Policy.” International Journal of Environ- va, Switzerland: World Health Organization. https://www.who. mental Research and Public Health 17 (15): 5267. https://doi. int/publications/i/item/9789241565042. org/10.3390/ijerph17155267. 176. _______. 2017a. Global Strategy and Action Plan on Ageing 162. Wannamethee, S. G., O. Papacosta, D. A. Lawlor, P. H. Whincup, and Health. Geneva, Switzerland: World Health Organization. G. D. Lowe, S. Ebrahim, and N. Sattar. 2011. “Do Women Exhibit https://apps.who.int/iris/handle/10665/329960. Greater Differences in Established and Novel Risk Factors be- 177. _______. 2017b. Health Employment and Economic tween Diabetes and Non-Diabetes than Men? The British Re- Growth: An Evidence Base. Geneva, Switzerland: World Health gional Heart Study and British Women’s Heart Health Study.” Organization. https://apps.who.int/iris/handle/10665/326411. Diabetologia 55 (1): 80–87. https://doi.org/10.1007/s00125- 178. _______. 2018. “World Elder Abuse Awareness Day 2018.” 011-2284-4. World Health Organization. June 15, 2018. https://www.who. 163. Washington, Donna L., Bevanne Bean-Mayberry, Deborah int/campaigns/world-elder-abuse-awareness-day/2018. Riopelle, and Elizabeth M. Yano. 2011. “Access to Care for Wom- 179. _______. 2021. “Cardiovascular Diseases (CVDs).” World Health en Veterans: Delayed Healthcare and Unmet Need.” Journal Organization, June 11, 2021. https://www.who.int/news- of General Internal Medicine 26 (S2): 655–61. https://doi. room/fact-sheets/detail/cardiovascular-diseases-(cvds). org/10.1007/s11606-011-1772-z. 180. _______. 2022. “Cancer.” World Health Organization. Feb- 164. Weber, Ann M, Beniamino Cislaghi, Valerie Meausoone, Safa ruary 3, 2022. https://www.who.int/news-room/fact-sheets/ Abdalla, Iván Mejía-Guevara, Pooja Loftus, Emma Hallgren, detail/cancer. et al. 2019a. “Gender Norms and Health: Insights from Global 181. Wu, Daphne C., Jeremy Veillard, Victoria Haldane, Seemeen Survey Data.” The Lancet 393 (10189): 2455–68. https://doi. Saadat, and Prabhat Jha. 2024. “Assessing Healthy Longevity org/10.1016/s0140-6736(19)30765-2. in Low- and Middle-Income Countries: A Proposed Healthy 165. Weber, Ann M, Beniamino Cislaghi, Valerie Meausoone, Safa Longevity Performance Dashboard for India.” in Unlocking the Abdalla, Iván Mejía-Guevara, Pooja Loftus, Emma Hallgren, Power of Healthy Longevity: Compendium of Research for the et al. 2019b. “Gender Norms and Health: Insights from Global Healthy Longevity Initiative.  Washington D.C.: World Bank. Survey Data.” The Lancet 393 (10189): 2455–68. https://doi. 182. Yamada, Y. (2002) ‘Profile of Home Care Aides, Nursing Home org/10.1016/s0140-6736(19)30765-2. Aides, and Hospital Aides: Historical Changes and Data Recom- 166. Williams, Allison M. 2018. “Education, Training, and Mentorship mendations’, The Gerontologist 42(2), pp. 199–206. https:// of Caregivers of Canadians Experiencing a Life-Limiting Illness.” doi.org/10.1093/geront/42.2.199. Journal of Palliative Medicine 21 (S1): S-45-S-49. https://doi. 183. Yeh, Shu-Chuan, Shwu-Feng Tsay, Wen Chun Wang, Ying-Ying org/10.1089/jpm.2017.0393. Lo, and Hon-Yi Shi. 2021. “Determinants of Successful Nursing 167. Williams, Fiona. 2011. “Towards a transnational analysis of the Home Accreditation.” Inquiry: A Journal of Medical Care Orga- political economy of care. Feminist ethics and social policy: To- nization, Provision and Financing 58 (November). https://doi. wards a new global political economy of care.” The Stockholm org/10.1177/00469580211059998. University Linnaeus Center for Integration Studies (SULCIS). 184. Zeng, Qingjun, Qingqing Wang, Lu Zhang, and Xiaocang Xu. Working Paper 2011:6. ISSN 1654-1189. 2020. “Comparison of the Measurement of Long-Term Care 168. Williams, Joni S., Kinfe Bishu, Clara E. Dismuke, and Leonard Costs between China and Other Countries: A Systematic Re- 256 CHAPTER 8 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E view of the Last Decade.” Healthcare 8 (2): 117. https://doi. org/10.1080/07399330500177196. org/10.3390/healthcare8020117. 186. Zucker, Irving, and Brian J. Prendergast. 2020. “Sex Differences 185. Zhan, Heying Jenny. 2005. “Aging, Health Care, and Elder in Pharmacokinetics Predict Adverse Drug Reactions in Wom- Care: Perpetuation of Gender Inequalities in China.” Health en.” Biology of Sex Differences 11 (1). https://doi.org/10.1186/ Care for Women International 26 (8): 693–712. https://doi. s13293-020-00308-5. COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 8 257 ANNEX 8.1 Cancer mortality and prevalence by type for older women and men FIGURE 8A.1 Cancer Mortality by Type, Ages 45+ years Source: IHME 2020 258 CHAPTER 8 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E FIGURE 8A.2 Prevalence of Cancer by Type, Ages 45+ years Source: IHME 2020 9 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 9 259 Gendered Responsibilities, Elderly Care, and Labor Supply Evidence from four middle-income countries Roberta Gatti a, Daniel Halim b,c, Allen Hardiman d, and Shuqiao Sun e a Office of the Chief Economist, Middle East and North Africa, World Bank b Global Gender Unit, World Bank c Poverty and Equity, Middle East and North Africa, World Bank d Department of Economics, University of Illinois e Human Development Practice Group, World Bank 260 CHAPTER 9 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E INTRODUCTION The world’s population is aging. The World Health Organization (WHO) predicts that, in 2030, 1 in 6 persons worldwide will be aged 60 years or over (WHO 2021). The developing world is aging especially fast. Fifty-seven percent of the world’s population ages 75 and above—more than 138 million people—lived in developing regions in 2015. By 2050, these regions are expected to be home to nearly 8 in 10 older adults globally (UN Women 2017). This trend suggests that there will be an increase in the demand for long-term care (LTC) for older persons. The International Labor Organization (ILO) estimates that the number of older persons needing care worldwide will increase by 50 percent—from 0.2 billion in 2015 to 0.3 billion in 2030. This is faster than the rise in care needs among young children (ILO 2018). Care for older parents has traditionally been provid- countries in different regions. Due to data limita- ed by adult children through an “implicit social con- tions in Egypt, our investigation there looks at the tract,” whereby adult children return the care they re- relationship between an individual’s parents’ disabil- ceived in their childhood (Folbre 2014). This implicit ity status and labor supply. To conduct our analysis, contract continues to operate in many societies. The we utilize a combination of rich time-use surveys marketization of care work can be perceived as de- and longitudinal household survey data among indi- valuing the “sacredness” of care and diminishing the viduals ages 40 to 59. Overall, for older working-age quality and authenticity of the care provided (En- adults in all four countries, we find that there is a gland and Folbre 1999; Himmelweit 1999). Howev- significant negative correlation between labor sup- er, traditional family-based care arrangements may ply and providing care to parents. This negative im- not be sustainable as populations age: falling fertility pact spans both the extensive and intensive margins rates imply that older parents will depend on few- of employment for both men and women. However, er adult children. Meanwhile, the market for formal the negative effect on labor supply is stronger among LTC services is underdeveloped in many countries, women than men, consistent with other findings in with limited supply and/or unaffordable costs (Spill- the literature. Also consistent with the literature, our man et al. 2014; Agree & Glaser 2009). results show a stronger labor supply effect among Due to gender stereotypes, much of the respon- those who provide intensive caregiving of more than sibility for informal care is shouldered by women 10 hours per week, compared to those who provide (UN Women 2017). Time-use surveys from 64 less than 10 hours of care per week. countries suggest that women bear 76.2 percent of It is useful to contextualize our findings against the total time spent on unpaid care work, 3.2 times the backdrop of different labor market landscapes more than men (ILO 2018). Often in their prime in the four countries. As depicted in Figure 9.1, working years, care providers can experience vari- there is a negative correlation between a country’s ous negative effects on their labor supply caused by gender gap in labor force participation and its per the care burden (Connelly et al. 2018; Fahle & Mc- capita GDP. Looking in detail at each country, we Garry 2017). One global estimate suggests that 647 find that, in Colombia and Poland, both men and million working-age individuals are outside of the women who provide parental care1 are signifi- labor force due to family responsibilities, 606 mil- cantly less likely to be employed. Specifically, in lion of them women (ILO 2018). Because of this, Poland we see a 17 percentage-point decline from it is imperative to examine whether the growing the mean in employment among women and a 16 older-adult care burden imposes a disproportionate percentage-point decline among men, associated toll on labor supply among women. with having to provide care to parents. Colombia This study explores the relationship between shows a 20 percentage-point decline among wom- providing care to older parents/parents-in-law and en and a 16 percentage-point decline among men. labor supply for men and women in Colombia, In Poland, the reduction in employment is driven Egypt, Indonesia, and Poland, four middle-income by a fall in formal wage employment, whereas in 1   Throughout this chapter, the term “parental care” refers to care provided to older parents/parents-in-law within a household. COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 9 261 FIGURE 9.1  Gross domestic product (GDP) per capita and male-female labor force participation gap, four coun- tries, 2013-2018 Note: This graph visualizes the relationship between gross domestic product (GDP) per capita and the male-female labor force participation (LFP) gap in the four study countries between 2013 and 2018, among individuals ages 15+. Data are from the World Bank Gender Data Portal. COL= Colombia, EGY=Egypt, IDN=Indonesia, POL=Poland. Colombia the decline is driven by both wage em- that there is a negative relationship between care ployment and casual employment. In Indonesia, hours per week and labor supply, conditional on while there is no overall change in the extensive providing care. We observe a slightly larger estimat- margin of labor supply for either gender, we see a ed change among men than women, which may be significant decline in formal employment among driven by the higher propensity for men than wom- both men and women who provide care. However, en to be employed in the formal sector, where work while men transition from formal to informal em- schedules are less flexible. ployment when they provide care for their parents, To look at how parental care interacts with we do not see this transition among female workers. individual characteristics, we also investigate how When looking at the intensive margin, we observe our results change among individuals with different a reduction in weekly work hours and annual earn- educational levels. Among men, we find that there ings among men and women in Poland, whereas a is no differential impact on employment based on similar decline is seen only among men in Indo- educational level. On the other hand, we observe nesia. In Colombia, the reduction in work hours is a larger negative impact among women who are only observed among female workers. On the other highly educated. This may be driven by the signifi- hand, our evidence from Egypt shows that, in that cantly higher base employment rate among highly country, there is no significant correlation between educated women. Because of this, care responsibili- an individual’s parents’ disability status and overall ties toward parents would likely have more effect on labor supply for either men or women. this particular group of female individuals. In addition to the extensive margin of provid- Responsibilities for childcare and parental ing parental care, we also investigate how the inten- care are often intertwined. When investigating the sity of parental care (as measured by the number of broader family structure, the presence of children hours spent providing care per week) relates to la- can either help alleviate the burden of caring for bor supply. Similar to the extensive margin, we find parents or, on the contrary, exacerbate the employ- 262 CHAPTER 9 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E ment impacts of care responsibilities toward older LTC has yet to be recognized as an urgent policy adults in the household. We find evidence for the issue in most LMICs (UN Women 2017). Moreover, former among men. The presence of older children evidence from LMICs is more limited than from ages 13-18 and above seems to act as a facilitator high-income settings, and much of the available for men in the household to provide care to older LMIC evidence stems from a single country, China. parents. Among women, we find similar evidence Among Chinese studies, Liu et al. (2010) find that, concerning the presence of young children (ages while caring for her own parents does not typically 0-6) in the household. affect a woman’s labor market outcomes, caring for This study builds on a growing literature on a parent-in-law harms a woman’s labor supply in the relationship between caring for older parents terms of paid work and work hours. Chai et al. and labor supply and how gender may be associ- (2021) report an inverse relationship between care- ated with this relationship. While a vast literature giving and labor force participation among Chinese shows a negative relationship between parental care men, while the results are mixed for women. Huang and labor force participation, most of the analysis et al. (2021) find that providing care to parents/ has been done in developed countries, particularly parents-in-law at home exerts a negative effect on the United States and European countries. Ettner female caregivers’ well-being. Chen et al. (2017) also (1995) was one of the first papers to find that co-res- show that married women who spend more than 15 idence with disabled parents leads to a significant hours per week providing elder care are less likely to reduction in work hours among female individu- participate in the labor force. They find that in- als ages 35 to 64 in the United States. More recent tensive caregivers (defined as those providing more literature also finds reduced labor supply due to than 15 hours of older-adult care per week) work parental care among individuals aged 50+ (John- significantly fewer hours of paid work each week son & Lo Sasso 2006; Johnson & Lo Sasso 2000; when they remain employed. Papers from China Van Houtven et al. 2013; Fahle & McGarry 2017; that find no negative impact of elder care on labor Heger & Korfhage 2020). In Europe, similar results supply include Mao et al. (2018) and Wang & Zhang are reported among elderly individuals (Bolin et al. (2018). Meanwhile, in Türkiye, Terkoglu & Memis 2008; Carmichael & Charles 2003; Ciccarelli & Van (2022) find that providing care to older adults has a Soest 2018; Kotsadam 2011). In Australia, Nguyen statistically significantly negative impact on the labor & Connely (2014) and Bittman et al. (2007) find a force participation of women ages 30 to 49. In similar negative relationship between caregiving Nepal, Sinha & Sedai (2022) find that both men and and labor supply. Some findings indicate a larger women caregivers experience a decline in labor impact on women’s labor supply compared to men. supply, with a larger impact in women. While Carmichael & Charles (2003) find that both The current study is most closely related men and women carers are less likely to be in paid to Stampini et al. (2020), who focus their analysis work, conditional on working, women earn sig- on Chile, Colombia, Costa Rica, and Mexico. nificantly less than men. Fevang et al. (2012) find Across these countries, they find that women who an almost four times larger impact on a daughter’s provide LTC are less likely to work. Additionally, labor supply, compared to a son’s, during the years condition-al on working, women who provide before a parent’s death. Van Houtven et al. (2013) parental care work fewer hours per week than likewise find reduced hours worked among caregiv- those who do not provide such care. ing women, conditional on working, while they find The rest of this chapter is structured as follows. no such effect on men. Our analysis contributes to The next section describes the data and the two this literature by exploring the relationship between treatment variables used. The study’s methodology parental care and labor supply in middle-income is explained in the following section. The following countries in regions (East Asia and Pacific, Middle section discusses the results from our main analysis East and North Africa, Latin America and Carib- using long-term parental care as our variable of in- bean, and Europe and Central Asia) that are aging terest. The next section follows up with heterogene- faster than other parts of the world. ity analysis, while the proceeding section provides a While the majority of the aging population sensitivity analysis with various control groups. The lives in low- and middle-income countries (LMICs), final section concludes. COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 9 263 DATA Our analysis focuses on four countries: Colombia, Egypt, Indonesia, and Poland. For Colombia and Poland, we take advantage of rich time-use data from the 2017 National Survey of Time Use (Colombia) and the 2013 Poland Time Budget Survey. Due to the lack of time-use surveys in Indonesia, we utilize a rich panel household survey from the 2014 wave of the Indonesia Family Life Survey (IFLS). For Egypt, we utilize the 2018 Egypt Labor Market Panel Survey (ELMPS) to construct our proxy treatment variable, which is defined below. Since the Egypt data does not have information on parental care, we use the individual’s parents’ disability status as a proxy for the provision of long-term care. Due to the differences in the sources of data used, Egypt. Details on the construction of the variables there is a slight difference in how our long-term for all four countries are presented in Table 9.1. care variable is defined across countries. The time- Our main outcome variable is individuals’ use surveys from Poland and Colombia directly ask labor supply. For the extensive margin, we look survey respondents (adult children) whether they into the association between our main treatment provide care for their parents. In Indonesia, IFLS variables and whether an individual is currently instead asks each older parent whether their chil- working. We also disaggregate our results based dren often provide help in conducting their daily on informal/formal sector. The formal sector in- activities. In all three countries, in addition to the cludes wage workers, such as private/government extensive margin, we also extract information on employees. The informal sector includes all those the number of hours spent per week on providing who are self-employed or who work as casual fam- parental LTC. Since Egypt does not have informa- ily workers. We also investigate the impact of pa- tion on providing long-term care, we utilize the in- rental care on the intensive margin, including con- formation on whether an individual parent reports ditional hours worked and annual earnings among having some degree of difficulty in conducting some those who continue working. For annual earnings, aspect of their daily activities. We use this informa- we transform earnings from each individual in each tion as a proxy for our main treatment variable for country to the equivalent in 2018 US dollars. TABLE 9.1  Definition of long-term care by country Country (Year) Definition of Treatment Variables Poland (2013)1 Provide care to older parents who are ill/disabled Indonesia (2014)2 Older parents report receiving help with activities of daily living (ADL) or instrumental activities of daily living (IADL) Colombia (2016)3 Provide help to older parents with activities such as: 1) Feeding/bathing 2) Medical appointments 3) Taking medication Egypt (2018)4 Individual’s parent report having some difficulty with or being unable to: 1) Walk/climb stairs 2) Perform self-care such as washing all over or dressing 1 Poland Population Time Budget Survey 2013 2 Indonesia Family Life Survey 2014-2015 3 Colombia Encuesta Nacional De Uso Del Tiempo (ENUT) 2016-2017 4 Egypt Labor Market Panel Survey 2018 264 CHAPTER 9 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E METHODOLOGY To estimate the relationship between provision fixed effects in our harmonized analysis. of parental long-term care/presence of a disabled Additionally, in order to observe whether the parent and labor supply in Colombia, Indonesia, differential impact of LTC on men’s and women’s Poland, and Egypt, we estimate the following equa- labor supply is significant, we also run our analysis tions for each individual i: interacting LTC dummy and gender dummy. Specif- ically, we will run:   [1] [2] To fully understand how gender interacts with the provision of long-term care and labor supply, we separate our estimation results for males and fe- In our main analysis, we focus on midlife individ- males. In addition to running the harmonized esti- uals ages 40 to 59, because these individuals are mates, we also run our estimates separately for each the subpopulation most likely to provide care to country. L represents our dependent variables: a aging parents/parents-in-law (Moussa et al. 2019). dummy variable that indicates whether an individu- Our data also confirm that the majority of persons al is currently working, types of employment, condi- providing care for elderly parents are aged 40 to 59 tional hours worked per week, and annual earnings (Annex Figure 9A.1). Meanwhile, individuals ages conditional on working. LTC is our main treatment 60 and above are excluded, because there is a high variable and indicates whether an individual is pro- probability that the parents of these individuals are viding care to their parents/parents-in-law who are deceased and that they themselves may be in need 60 years old or above. In the case of Egypt, LTC re- of care. Since the decision to provide parental LTC fers to an individual’s parent’s disability status. X is is endogenous, we do not claim causality in our a vector of individual and household characteristics analysis. For example, there is a potential for reverse such as individual age, age-squared, highest educa- causality, where the lack of labor market opportu- tional level, number of children ages 0-6, 7-12, and nities may imply lower opportunity costs for those 13-18 in the household, marital status, and country becoming caregivers. RESULTS Co-residency and care needs Table 9.2 shows the basic summary statistics for co-residency and care needs among men and women. It also shows whether there is a significant difference in these statistics between men and women. Table 9.2 shows that, in all four countries, the share of co-residency with parents is generally higher among men than women. How- ever, we also see a higher proportion of women who are spouses of older adults. We observe a higher proportion of men who live with a disabled parent in Poland, while the proportion is equal between genders in Indonesia, Egypt, and Colombia. Despite this, women bear the majority of the responsibility to provide care for their par- ents. In general, women are at least twice as likely as men to provide care to their older parents. We also observe that, among the three countries, Indonesia has the largest proportion of women who provide LTC to their par- ents, while Poland has the smallest proportion of women providing such care. In addition, in Indonesia, women spend a significantly longer time caring for their parents, conditional on providing care, while in Colombia men spend a slightly longer time than women. Figure 9.2 shows how the basic co-residency and both men and women. We also observe that men care needs analyzed in Table 9.2 change as indi- experience the peak of biological reproduction at a viduals age in all four countries. In these figures, much older age than women. We see a downward we include individuals ages 25 to 59 to see a more trend in co-residency with age, which is not surpris- complete picture of this evolution. Among the four ing given the increasing number of deceased older countries, Egypt has the highest fertility rates for parents as individuals age. COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 9 265 TABLE 9.2  Co-residency and long-term care provision among individuals ages 40 to 59 Poland (2013) Indonesia (2014) Colombia (2016) Egypt (2018) Male Female Male Female Male Female Male Female (1) (2) (3) (4) (5) (6) (7) (8) Coreside with an older parent 0.134 0.108*** 0.142 0.124** 0.164 0.141*** 0.083 0.081 (0.341) (0.311) (0.349) (0.330) (0.370) (0.348) (0.277) (0.274) Relationship to the older adult: Spouse 0.047 0.145*** 0.005 0.186*** 0.030 0.212*** 0.007 0.269*** (0.211) (0.352) (0.067) (0.389) (0.171) (0.409) (0.086) (0.443) Child 0.101 0.080*** 0.076 0.097*** 0.129 0.131 0.079 0.044*** (0.301) (0.271) (0.265) (0.295) (0.335) (0.337) (0.270) (0.205) Child-in-law 0.034 0.028 0.066 0.028*** 0.035 0.011*** 0.004 0.037*** (0.181) (0.166) (0.249) (0.165) (0.184) (0.105) (0.066) (0.190) # of coresident parents in 1.200 1.158* 1.224 1.145*** 1.233 1.224 1.124 1.174* the household (0.400) (0.364) (0.417) (0.353) (0.423) (0.420) (0.356) (0.380) # of elderly parents in the 1.201 1.158* 1.174 1.141 1.222 1.219 1.101 1.141 household (0.401) (0.365) (0.379) (0.349) (0.415) (0.413) (0.302) (0.349) Older parent(s) have a 0.068 0.069 0.120 0.108*** 0.052 0.038** chronic illness (0.252) (0.253) (0.325) (0.311) (0.221) (0.191) Share of disabled older 0.023 0.018* 0.076 0.079 0.029 0.031 0.048 0.050 parent (0.149) (0.132) (0.265) (0.270) (0.167) (0.173) (0.214) (0.218) Provide parental care to 0.008 0.014** 0.020 0.076*** 0.004 0.015*** older parent (0.090) (0.119) (0.142) (0.265) (0.065) (0.123) Time spent providing care to 9.079 9.567 14.666 19.221 11.268 8.194* older parents (hours/week) (8.273) (8.159) (20.880) (26.760) (17.151) (9.197) # of children aged … in the household: 0-6 1.230 1.197 1.208 1.226 1.226 1.259* 1.429 1.409 (0.470) (0.451) (0.460) (0.508) (0.520) (0.573) (0.692) (0.750) 7-12 1.215 1.172* 1.224 1.214 1.247 1.213** 1.497 1.352*** (0.473) (0.419) (0.492) (0.476) (0.534) (0.500) (0.656) (0.599) 13-18 1.265 1.250 1.221 1.214 1.281 1.250** 1.501 1.432*** (0.530) (0.515) (0.469) (0.453) (0.538) (0.517) (0.628) (0.627) Observations 4,769 8,656 4,906 5,077 15,925 19,453 4,841 4,860 Notes: The columns show means and standard deviations for each variable in different countries, separately for men and women. Standard deviations are shown in parentheses. *, **, *** in “Female” columns indicate that the null hypothesis for similarity of means between male and female samples within each country is rejected at the 10%, 5%, and 1% level, respectively. 266 CHAPTER 9 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E FIGURE 9.2  Share of individuals (ages 25 to 59) with a coresident parent and young children in the household COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 9 267 Determinants of parental care Before investigating the relationship between labor This implies that there is a selection of parental care supply and parental care, we try to investigate fac- providers among women, whereby less educated tors that determine parental care. To do this, we re- women are more likely to provide care to an elderly gress our explanatory variable (parental care) on in- parent. We also observe that being married reduces dividual and household characteristics. The results the likelihood of providing parental care. As women are presented in Table 9.3. We observe that more age, they are more likely to provide care to their par- highly educated women are less likely to provide ents, while this relationship is not significant among parental care. This negative relationship might be men. When looking at household composition, we driven by increasing opportunity costs of providing observe that the number of children ages 0 to 18 is care among women who enjoy higher education and associated with an increase in the propensity for who are more likely to be in a higher-paying job. providing parental care among men. TABLE 9.3  Determinants of parental care, three-country pooled sample Parental care Male Female (1) (2) Primary or less 0.000 0.000 (.) (.) Secondary 0.000 -0.009*** (0.002) (0.002) Tertiary 0.001 -0.010*** (0.002) (0.002) Married -0.018*** -0.021*** (0.002) (0.002) Age 0.003 0.008*** (0.002) (0.003) Age-sq -0.000 -0.000*** (0.000) (0.000) # children 0-6 0.004** -0.002 (0.002) (0.002) # children 7-12 0.004*** -0.001 (0.001) (0.001) # children 13-18 0.002** -0.002* (0.001) (0.001) Constant -0.054 -0.125* (0.055) (0.069) Observations 29727 36520 Mean 0.014 0.026 Notes: Robust standard errors in parentheses. *, **, *** indicate statistical significance at 10%, 5%, and 1%, respectively. Each regression includes individuals ages 40-59 from Poland, Indonesia, and Colombia. 268 CHAPTER 9 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E Labor supply Next, we investigate the difference in labor sup- provider that plays a role in muting the gender dis- ply between men and women. We also observe parity in labor supply. the gender difference between LTC providers and It is also interesting to note the different labor non-providers (or persons with disabled parents market landscapes in the four countries. Looking versus non-disabled parents for Egypt). As is well at non-carers as the baseline, around 90 percent or documented in the literature, we see a large dispar- more of males ages 40 to 59 in Indonesia, Colom- ity in labor supply between men and women in all bia, and Egypt are employed; the share is lowest in four countries. In general, men are more likely to be Poland, at 76 percent. There is more variation in the employed, work longer hours conditional on work- employment of females ages 40 to 59; the highest ing, and receive higher annual earnings than wom- share is in Indonesia, with 73 percent of women in en. Men are in general more likely to be employed this age group employed, compared with around 60 in wage employment and self-employment but are percent in Poland and Colombia and just 24 percent less likely to be employed as casual/family workers. in Egypt. There are also different degrees of formal- Unsurprisingly, this gender disparity in labor ity in the labor market. Roughly half of the employ- supply is even more pronounced among persons ment in Poland (for both women and men) is in who do not provide care for parents (or, in Egypt, wage work. Meanwhile, employment in Colombia who do not have disabled parents) across all four and Indonesia is dominated by casual and self-em- countries. For example, the difference in hours ployment. While there is a greater share of wage worked between men and women is smaller and not workers among males in Egypt (67 percent), many significant among persons providing care to par- fewer women work as wage workers (16 percent). ents in Poland and Indonesia. We observe a similar These are important contextual factors, considering trend in the gender difference in wage employment, that there are different social norms around wom- self-employment, and casual/family work. Since the en’s work in these four countries and that casual and decision to provide care to older parents is endog- self-employment may offer more flexibility to com- enous, there will be selection into becoming a care bine paid market work and unpaid care work. TABLE 9.4  Labor supply among individuals ages 40 to 59 Carer Non-carer Male Female M–F Male Female M–F (3) – (6) (1) (2) (3) (4) (5) (6) (7) Panel A: Poland Employed 0.462 0.432 0.0295 0.759 0.636 0.123*** -0.094 (0.505) (0.497) (0.428) (0.481) Hours worked / week 36.111 35.000 1.111 43.125 38.594 4.531*** -3.42 (11.146) (12.232) (10.038) (9.992) Wage worker 0.179 0.296 -0.117 0.528 0.492 0.036*** -0.153* (0.389) (0.458) (0.499) (0.500) Self-employed 0.282 0.120 0.162* 0.227 0.122 0.105*** 0.057 (0.456) (0.326) (0.419) (0.327) Casual/family worker 0.000 0.016 -0.016 0.004 0.022 -0.018*** 0.002 (0.000) (0.126) (0.063) (0.148) Annual earnings (USD) 3893.03 4331.42 -438.4 5431.29 4758.31 673.0*** -1111.3* (2513.70) (1769.45) (2078.07) (1875.48) Primary educated or less 0.154 0.128 0.026 0.109 0.106 0.003 0.022 (0.366) (0.335) (0.312) (0.307) Secondary educated 0.769 0.672 0.097 0.719 0.657 0.061*** 0.036 (0.427) (0.471) (0.450) (0.475) COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 9 269 Carer Non-carer Male Female M–F Male Female M–F (3) – (6) (1) (2) (3) (4) (5) (6) (7) Panel A: Poland (continued) Tertiary educated 0.077 0.200 -0.123 0.172 0.237 -0.065*** -0.058 (0.270) (0.402) (0.378) (0.425) Observations 39 125 4730 8531 Panel B: Indonesia Employed 0.945 0.713 0.232*** 0.956 0.726 0.231*** 0.002 (0.229) (0.453) (0.204) (0.446) Hours worked / week 33.372 33.306 0.0659 39.195 34.230 4.965*** -4.899* (21.009) (22.820) (21.265) (23.811) Wage worker 0.319 0.169 0.15*** 0.369 0.196 0.173*** -0.023 (0.469) (0.375) (0.483) (0.397) Self-employed 0.462 0.321 0.141* 0.445 0.317 0.128*** 0.013 (0.501) (0.468) (0.497) (0.465) Casual/family worker 0.165 0.220 -0.0548 0.140 0.208 -0.068*** -0.013 (0.373) (0.415) (0.347) (0.406) Annual earnings (USD) 1278.01 803.90 474.1** 1394.50 799.05 595.5*** -121.35 (1374.15) (1264.89) (1288.81) (1165.54) Primary educated or less 0.330 0.557 -0.228*** 0.427 0.570 -0.143*** -0.085 (0.473) (0.498) (0.495) (0.495) Secondary educated 0.462 0.334 0.127* 0.438 0.332 0.107*** 0.02 (0.501) (0.473) (0.496) (0.471) Tertiary educated 0.209 0.108 0.101* 0.135 0.099 -0.036*** 0.065 (0.409) (0.311) (0.341) (0.298) Observations 91 296 4,351 3,601 Panel C: Colombia Employed 0.687 0.416 0.270*** 0.885 0.569 0.316*** -0.046 (0.467) (0.494) (0.319) (0.495) Hours worked / week 45.457 36.258 9.198*** 49.696 41.748 7.948*** 1.250 (14.725) (16.310) (14.384) (16.537) Wage worker 0.254 0.141 0.113* 0.328 0.230 0.098*** 0.015 (0.438) (0.349) (0.469) (0.421) Self-employed 0.418 0.245 0.173** 0.520 0.285 0.235*** -0.062 (0.497) (0.431) (0.500) (0.451) Casual/family worker 0.015 0.030 -0.015 0.038 0.054 -0.016*** 0.001 (0.122) (0.171) (0.191) (0.227) Annual earnings (USD) 4606.89 3488.75 1118.1 3733.00 3330.56 403.4*** 714.7 (2872.13) (2828.99) (2489.28) (2564.28) Primary educated or less 0.284 0.333 -0.05 0.377 0.358 0.019*** -0.069 (0.454) (0.472) (0.485) (0.480) Secondary educated 0.403 0.404 -0.001 0.411 0.409 0.002 -0.003 (0.494) (0.492) (0.492) (0.492) 270 CHAPTER 9 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E Carer Non-carer Male Female M–F Male Female M–F (3) – (6) (1) (2) (3) (4) (5) (6) (7) Panel C: Colombia (continued) Tertiary educated 0.313 0.263 0.051 0.211 0.232 -0.021*** 0.072 (0.467) (0.441) (0.408) (0.422) Observations 67 298 15,856 19,139 Panel D: Egypt Employed 0.854 0.306 0.548*** 0.896 0.244 0.652*** -0.103*** (0.321) (0.470) (0.281) (0.434) Hours worked / week 45.143 36.605 8.538*** 47.011 36.260 10.75*** -2.212 (18.701) (13.151) (17.746) (14.980) Wage worker 0.644 0.207 0.437*** 0.665 0.162 0.503*** -0.066 (0.480) (0.406) (0.472) (0.369) Self-employed 0.202 0.041 0.16*** 0.225 0.041 0.184*** -0.024 (0.402) (0.199) (0.418) (0.198) Casual/family worker 0.009 0.074 -0.065* 0.008 0.046 -0.038*** -0.028* (0.092) (0.263) (0.091) (0.209) Annual earnings (USD) 1468.32 1292.61 175.7 1588.65 1483.99 104.7** 71.057 (937.08) (649.23) (908.68) (785.07) Primary educated or less 0.412 0.541 -0.129** 0.406 0.601 -0.195*** 0.065 (0.493) (0.499) (0.491) (0.490) Secondary educated 0.403 0.281 0.122** 0.396 0.275 0.121*** 0.002 (0.492) (0.450) (0.489) (0.447) Tertiary educated 0.185 0.178 0.007 0.198 0.124 0.074*** -0.067* (0.389) (0.383) (0.398) (0.330) Observations 233 242 4,567 4,592 Notes: Columns 1, 2, 4, and 5 show means and standard deviations for each variable for male and female samples, as indicated. Standard deviations shown in parentheses. Columns 3, 6, and 7 test the statistical significance of difference in means. *, **, *** indicate statistical significance at 10%, 5%, and 1%, respectively. How does parental care differentially affect male and female labor supply? We begin by looking at the relationship between pa- hours worked, the point estimates for women are in rental care and labor supply for men and women ages general higher than for men. While men are 5.6 per- 40 to 59 in all four countries combined. Panel A in centage points (p.p.) less likely to be employed when Table 9.5 presents the results for males, while panel B they provide care to their parents, the decline for presents the results for females. We find that providing women is almost twice as large, with a 10 p.p. decline care to older parents is associated with a significant in the employment rate among female carers. We reduction in labor supply for both men and women. observe a similarly sharper decline in women’s prob- This reduction applies not only to the probability of ability of wage employment (7.9 p.p. versus 5.3 p.p. employment but also to intensive margins such as for men) and annual earnings (22 percent versus 17 conditional hours worked and log annual earnings. percent for men). In contrast, we observe zero impact Both men and women carers work on average about on the probability of self/casual employment for men three hours less per week and earn less, condition- and women. This result highlights the importance of al on working. In almost all cases except conditional flexible work hours in the informal sector. COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 9 271 TABLE 9.5  Correlational effect on employment of providing long-term care to parents, by gender, four-country pooled sample Wage Casual/Self Weekly Hours Log Earnings Employed Employment Employment Worked (USD) (1) (2) (3) (4) (5) Panel A: Male Parental Care -0.056*** -0.053** -0.001 -3.108*** -0.171** (0.018) (0.022) (0.022) (0.996) (0.084) Observations 29725 29725 29725 25908 14758 Dep. Var Mean 0.876 0.418 0.458 46.627 7.319 Share of Parent Carer 0.014 0.014 0.014 0.013 0.017 Panel B: Female Parental Care -0.099*** -0.079*** -0.020 -2.709*** -0.228* (0.016) (0.012) (0.014) (0.890) (0.130) Observations 36519 36519 36519 20284 12067 Dep. Var Mean 0.556 0.276 0.279 39.501 6.994 Share of Parent Carer 0.026 0.026 0.026 0.023 0.028 Notes: Robust standard errors in parentheses. *, **, *** indicate statistical significance at 10%, 5%, and 1%, respectively. Each regression includes individual and house- hold-level characteristics: individual age, age-squared, country fixed effect, highest educational level, number of children aged 0-6, 7-12, and 13-18 in the household, and marital status. Sample includes all individuals aged 40-59 in Poland, Egypt, Indonesia, and Colombia. Log earnings indicate log annual earnings (in USD) conditional on being employed. Before diving into the results for each specific coun- men and women, where the negative effect is less try, we investigate whether the worse outcomes we pronounced among males. LTC is associated with a observe among women in Table 9.5 are significant. 5.7 percentage points gender employment gap and a Table 9.6 provides the estimates from equation (2) 31.5 percent gender earnings gap to women’s disad- for all four countries combined where we interact vantage. In contrast, we do not observe meaningful our main explanatory variable (LTC) with the gen- gender differences in the impact on hours worked der indicator to investigate whether there is a signif- and self-employment. These results underscore icant impact difference between men and women. women’s disproportionate responsibility to provide Columns (1) through (5) confirm what we already LTC and their associated disadvantage in paid em- observe in Table 9.5: there is a larger negative im- ployment relative to men. Men’s advantage in wage pact on labor supply among women than among employment also highlights the well-document- men who provide long-term care for their parents. ed observation that women often turn to casual/ More notably, we observe a statistically significant self-employment, which offers more time flexibility differential effect of long-term care on employment, to shoulder the double responsibilities of unpaid wage employment, and annual earnings between care and paid market work. 272 CHAPTER 9 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E TABLE 9.6  Gender difference in the effect on employment of providing long-term care to parents, four-country pooled sample Wage Casual/Self Weekly Hours Log Earnings Employed Employment Employment Worked (USD) (1) (2) (3) (4) (5) Parental Care -0.087*** -0.087*** 0.000 -2.076** -0.476*** (0.015) (0.012) (0.014) (0.888) (0.131) Male 0.317*** 0.150*** 0.167*** 7.030*** 0.676*** (0.003) (0.004) (0.004) (0.161) (0.027) Parental Care * Male 0.057** 0.082*** -0.024 -1.324 0.315** (0.025) (0.026) (0.025) (1.332) (0.155) Observations 66244 66244 66244 46192 26825 Dep. Var Mean 0.700 0.340 0.359 43.498 7.173 Share of Parent Carer 0.021 0.021 0.021 0.018 0.022 Notes: Robust standard errors in parentheses. *, **, *** indicate statistical significance at 10%, 5%, and 1%, respectively. Each regression includes individual and house- hold-level characteristics: individual age, age-squared, country fixed effect, highest educational level, number of children aged 0-6, 7-12, and 13-18 in the household, and marital status. Sample includes all individuals aged 40-59 in Poland, Egypt, Indonesia, and Colombia. Log earnings indicate log annual earnings (in USD) conditional on being employed. While the harmonized results are informative, (2) separately for each country. Column (1) to (3) in we recognize that there are vast differences among Table 9.7 present the results for men for each coun- the study countries in culture, social norms, and try, while columns (4) to (6) present the results for gender gaps in labor force participation, as shown women. LTC has a negative relationship with labor in Figure 9.1. To investigate how the relationship market outcomes for both men and women in Po- between parental LTC and labor market outcomes land and Colombia. However, the difference in im- differs across countries, we run equations (1) and pact on women and men is unclear. TABLE 9.7  Correlational effect of providing long-term care to parents on employment, by country Male Female Wage Casual/Self Wage Casual/Self Employed Employment Employment Employed Employment Employment (1) (2) (3) (4) (5) (6) Panel A: Poland Long-term care -0.158** -0.223*** 0.065 -0.174*** -0.169*** -0.005 (0.076) (0.062) (0.067) (0.041) (0.038) (0.030) Observations 4768 4768 4768 8653 8653 8653 Dep. Var. mean 0.757 0.525 0.232 0.633 0.489 0.144 Share of parent carer 0.008 0.008 0.008 0.014 0.014 0.014 Panel B: Indonesia Long-term care 0.010 -0.077 0.089* -0.032 -0.038* 0.008 (0.023) (0.048) (0.050) (0.027) (0.020) (0.029) Observations 4438 4438 4438 3894 3894 3894 Dep. Var. mean 0.956 0.368 0.586 0.725 0.194 0.527 Share of parent carer 0.021 0.021 0.021 0.076 0.076 0.076 Panel C: Colombia Long-term care -0.156*** -0.064 -0.092 -0.195*** -0.110*** -0.085*** (0.055) (0.052) (0.060) (0.029) (0.020) (0.026) COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 9 273 Male Female Wage Casual/Self Wage Casual/Self Employed Employment Employment Employed Employment Employment (1) (2) (3) (4) (5) (6) Panel C: Colombia (continued) Observations 15727 15727 15727 19153 19153 19153 Dep. Var. mean 0.885 0.324 0.560 0.565 0.225 0.339 Share of parent carer 0.004 0.004 0.004 0.016 0.016 0.016 Panel D: Egypt Long-term care -0.017 -0.005 -0.010 0.035 0.013 0.023 (0.023) (0.032) (0.028) (0.028) (0.022) (0.020) Observations 4792 4792 4792 4819 4819 4819 Dep. Var. mean 0.894 0.664 0.228 0.247 0.165 0.082 Share of parent carer 0.049 0.049 0.049 0.050 0.050 0.050 Notes: Robust standard errors in parentheses. *, **, *** indicate statistical significance at 10%, 5%, and 1%, respectively. Each regression includes individual and household-level characteristics: individual age, age-squared, country fixed effect, highest educational level, number of children aged 0-6, 7-12, and 13-18 in the household, and marital status. Sample includes all individuals aged 40-59 in Poland, Egypt, Indonesia, and Colombia. In Poland, providing LTC is associated with of women live with disabled parents but less than 15.8 percentage points and 17.6 percentage points half of them provide care to the parents (Table 9.2). declines in employment for men and women, re- We also provide evidence of correlations in terms of spectively. This decline seems to be driven entirely hours worked and earnings for each country. Those by the drop in wage employment, which in general results are presented in Annex Table 9A.1. has more rigid hours and schedules. The decline in wage employment is also significantly larger among How does the intensity of parental care men. In Colombia, both men and women drop out affect labor supply among men and women? of employment when they provide LTC to their par- ents, and, in contrast to Poland, we observe a larger In the previous section, we investigated how the decline in all types of employment among women decision to provide long-term parental care affects compared to men. This larger decline in Colombia labor supply among men and women. However, a than Poland might be attributed to the more egal- substantial heterogeneity exists across individuals itarian society in Poland, as shown by the smaller in the intensity/number of hours spent caring for gap in labor force participation from Figure 9.1. parents per week. In Table 9.8, we take a closer look In Indonesia, we observe a transition from wage into the relationship between labor supply and the employment to informal employment among men hours spent per week on providing parental care. when they provide LTC, suggesting the role of more We exclude Egypt in this harmonized sample, as the flexibility in the casual employment sector. Howev- Egyptian data do not provide information on the in- er, this transition is not observed among Indonesian tensity of parental care. In general, we observe that, women who drop out of the wage employment sec- conditional on providing parental care, women on tor. Meanwhile, the result from Egypt shows that the average devote 0.5 to 1 hour more per week to care mere presence of a disabled parent in the household compared to men. Similar to the previous findings, does not have any meaningful impact on labor mar- both men and women experienced a significant de- ket outcomes for men and women in that country. cline in both the extensive margin of work and week- The null effect may partly be explained by the fact ly hours worked. In general, a 10 percent increase in that having parents who need help with daily activ- hours of providing parental care is associated with ities does not necessarily translate to care provided a 0.4 percentage points and a 0.6 percentage points by adult children. For example, in Indonesia, 7.6 decline in the probability of employment for men percent of males living with older parents need help and women, respectively. Both men and women also with daily activities, but only 2 percent provide care experience fewer hours worked, conditional on be- to the parents. Similarly, in Colombia, 3.1 percent ing employed. In Panel C, we test whether the differ- 274 CHAPTER 9 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E ence between men and women in the impact of care for this is the higher proportion of men who are in intensity on labor supply is significant. Contrary to formal employment, where there is less flexibility in our hypothesis, we find that the impact on men is work schedules. The added responsibility of parental slightly higher, indicating that additional time for care might drive men in this situation to be more providing parental care reduces the labor supply of likely to leave their employment. men more than women. One possible explanation TABLE 9.8  Correlational effect on employment of care hours devoted to parents, pooled sample from three countries Wage Casual/Self Weekly Hours Log Earnings Employed Employment Employment Worked (USD) (1) (2) (3) (4) (5) Panel A: Male Log Care Hours -0.039*** -0.050*** 0.012 -1.754*** -0.122* (0.012) (0.013) (0.015) (0.673) (0.070) Observations 25396 25396 25396 22161 12468 Dep. Var Mean 0.875 0.369 0.505 46.356 7.311 Mean Hours Cared 13.496 13.496 13.496 13.386 14.671 Panel B: Female Care Hours -0.055*** -0.041*** -0.014* -1.348*** -0.112* (0.008) (0.006) (0.007) (0.453) (0.061) Observations 32890 32890 32890 19955 12265 Dep. Var Mean 0.608 0.289 0.318 39.397 6.845 Mean Hours Cared 14.086 14.086 14.086 14.474 15.886 Panel C: Interaction Care Hours -0.044*** -0.038*** -0.005 -0.965** -0.217*** (0.008) (0.006) (0.007) (0.454) (0.061) Male 0.265*** 0.102*** 0.164*** 6.679*** 0.764*** (0.003) (0.004) (0.004) (0.167) (0.028) Care Hours*Male -0.026* -0.025* -0.000 -1.300 0.132 (0.015) (0.014) (0.016) (0.815) (0.092) Observations 58286 58286 58286 42116 24733 Dep. Var Mean 0.724 0.324 0.399 43.059 7.080 Mean Hours Cared 13.959 13.959 13.959 14.169 15.537 Notes: Robust standard errors in parentheses. *, **, *** indicate statistical significance at 10%, 5%, and 1%, respectively. Each regression includes individual and household-level characteristics: individual age, age-squared, highest educational level, number of children aged 0-6, 7-12, and 13-18 in the household, country fixed effects, and marital status. The regression includes individuals aged 40-59 from Poland, Colombia, and Indonesia. Individuals from Egypt are excluded because there is no information on amount of time devoted to parental care. Log earnings indicate log annual earnings (in USD) conditional on being employed. COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 9 275 HETEROGENEITY Educational level We proceed to observe how educational level interacts with the relationship between labor supply and older parental care. The feminization-U hypothesis suggests that the relationship between female labor supply and education is U-shaped, where it is higher for low- and high-skilled women and lower for middle-skilled women (Goldin 1995). Women with less education are driven by the necessity to be employed. But as women gain more education, the income effect starts to dominate the substitution effect. Women then tend to demand more leisure and drop out of the labor force. As women become highly educated, however, the substitution effect begins to dominate again, with increased opportunity costs of staying out of the labor force. In Table 9.9, we observe how an individual’s high- (secondary) than among those with low (primary est educational level interacts with the relationship or less) or high (tertiary) levels. Meanwhile, there is between long-term care provision/disabled parent no differential impact on employment based on ed- and labor supply. The omitted educational level ucation level among men. Conditional on working, category here includes those whose highest level of we do not see a differential effect of LTC on hours education is primary school or less. Indeed, we do worked by education level for either men or women. see that the coefficients make a U-shape for wom- For women, the results suggest that the impact of en, where the negative effect is greater in magnitude LTC on earnings is alleviated among women with among women with a “middle” educational level tertiary education. TABLE 9.9  Heterogeneous effect of long-term care provision by education level, pooled sample from four countries Male Female Hours Log earnings Hours Log earnings Employed worked/week (USD) Employed worked/week (USD) (1) (2) (3) (4) (5) (6) Parental Care -0.044 -3.569* -0.142 -0.040* -2.582* -0.582*** (0.029) (1.919) (0.112) (0.023) (1.516) (0.220) Parental Care * -0.026 2.107 -0.239 -0.116*** -0.166 0.285 Secondary educated (0.040) (2.365) (0.177) (0.034) (2.218) (0.302) Parental Care * -0.005 -2.104 0.345 -0.077* -0.272 0.950*** Tertiary educated (0.046) (2.590) (0.217) (0.043) (1.903) (0.315) Observations 29725 25908 14758 36519 20284 12067 Dep. var. mean 0.876 46.627 7.319 0.556 39.501 6.994 Share of parent carer 0.014 0.013 0.017 0.026 0.023 0.028 Notes: Robust standard errors in parentheses. *, **, *** indicate statistical significance at 10%, 5%, and 1%, respectively. Each regression includes individual and house- hold-level characteristics: individual age, age-squared, highest educational level, number of children aged 0-6, 7-12, and 13-18 in the household, country fixed effects, and marital status. The regression includes individuals from Poland, Colombia, Indonesia, and Egypt. Log earnings indicate log annual earnings (in USD) conditional on being employed. Presence of children This section tests how the presence of children in and older parents in the household. On the one the household may affect the relationship between hand, the double care responsibility can impose an labor supply and providing parental care. Specifi- extra burden on employed adults and encourage a cally, we test whether there is a differential impact decline in labor supply. On the other hand, older on labor supply in the “sandwich” generation, where children (ages 13 to 18) may help employed adults individuals have to take care of both young children provide parental care in the household. 276 CHAPTER 9 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E Table 9.10 shows that the presence of old- care provision for aged parents and the number of er children in the household, especially children young children ages 0 to 6. While it is not definitive, ages 13 to 18, helps male adults stay in their jobs one possible explanation for this is that older adults when there is a parental care responsibility. How- might be able to help out with supervising young ever, we also observe that this presence of older children. Another possible explanation is that the children exacerbates the negative impact on work care burden may exhibit economies of scale, mean- hours among male adults, conditional on working. ing that additional dependents affect women’s labor In contrast, the presence of older children does not force participation decision to a lesser degree than affect women’s already-heavy parental care burden. the first dependent in need of care. Further, we observe a positive interaction between TABLE 9.10  Heterogeneous effect of long-term care provision by presence of children in household, pooled sam- ple from four countries Male Female Hours Log earnings Hours Log earnings Employed worked/week (USD) Employed worked/week (USD) (1) (2) (3) (4) (5) (6) Parental Care -0.092*** -2.027 -0.215 -0.120*** -2.900*** -0.106 (0.026) (1.360) (0.142) (0.020) (1.081) (0.157) Parental Care * # -0.009 -0.792 0.056 0.077** 0.019 -0.282 children age 0-6 (0.018) (1.403) (0.090) (0.030) (1.466) (0.232) Parental Care * # 0.028 1.539 -0.032 0.019 1.028 0.059 children age 7-12 (0.018) (1.307) (0.093) (0.027) (1.753) (0.251) Parental Care * # 0.046** -2.687** 0.063 -0.003 -0.271 -0.167 children age 13-18 (0.021) (1.342) (0.094) (0.024) (1.538) (0.225) Observations 29725 25908 14758 36519 20284 12067 Dep. var. mean 0.876 46.627 7.319 0.556 39.501 6.994 Share of parent carer 0.014 0.013 0.017 0.026 0.023 0.028 Notes: Robust standard errors in parentheses. *, **, *** indicate statistical significance at 10%, 5%, and 1%, respectively. Each regression includes individual and house- hold-level characteristics: individual age, age-squared, highest educational level, number of children aged 0-6, 7-12, and 13-18 in the household, country fixed effects, and marital status. The regression includes individuals aged 40-59 from Poland, Colombia, Indonesia, and Egypt. Log earnings indicate log annual earnings (in USD) conditional on being employed. ROBUSTNESS CHECK The presence of adult household members In this section, we test sensitivity of our results by examining whether and how the relationship between LTC and labor supply will change with the presence of adult men/women in the household. Specifically, the hypothesis here is that the presence of adult women in the household may help other women participate in the labor market by mitigating the care burden, while the presence of adult men may reduce that probability due to gender stereotypes whereby men are expected to preferentially join the labor force. In this section, we test whether including the number of adult males and adult females in the household would change the relationship between LTC and labor supply for men and women. Table 9.11 presents our results. Columns (1) and of adult males and females in the household in the (3) show results from the original specification for main specification control variables. The results males and females, respectively. Columns (2) and show that, in general, controlling for the number (4) show the results when including the number of adult males and adult females does not signifi- COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 9 277 cantly change our results. However, this table also male adults seems to diminish the probability of confirms our previous hypothesis. Among female employment among female individuals. This result individuals, the presence of other adult women in may be driven by gender-cultural norms, where the household seems to increase the probability of male adults are more likely than women to enter the employment. We also observe that the presence of labor force. TABLE 9.11  Sensitivity check: including the number of adult females/males in households Employed Male Female (1) (2) (3) (4) Parental Care -0.056*** -0.056*** -0.099*** -0.099*** (0.018) (0.018) (0.016) (0.016) Number of Adult Male in HH -0.006** -0.010*** (0.003) (0.003) Number of Adult Female in HH -0.003 0.007** (0.003) (0.004) Observations 29725 29725 36519 36519 Mean 0.876 0.876 0.556 0.556 Share of Parent Carer 0.014 0.014 0.026 0.026 Notes: Robust standard errors in parentheses. *, **, *** indicate statistical significance at 10%, 5%, and 1%, respectively. Each regression includes individual and house- hold-level characteristics: individual age, age-squared, highest educational level, number of children aged 0-6, 7-12, and 13-18 in the household, country fixed effects, and marital status. The regression includes individuals aged 40-59 from Poland, Colombia, Indonesia, and Egypt. HH=household. Intensity of parental care Past literature has indicated that the length of care- care of more than 10 hours per week. giving that adults provide has a non-trivial impact on Table 9.12 presents our results. In general, these labor supply. Chai et al. (2021) find that only women align with findings in the previous literature. We find who provide more than eight hours of care per week that, while there is a negative effect on employment experience negative associated changes in labor force among men and women who provide less than 10 participation. Walsh & Murphy (2018) provide simi- hours of care per week, the effect is substantially lar evidence among those who provide more than 15 larger among those who provide more than 10 hours hours of care per week in Ireland. To investigate fur- of care per week. Moreover, individuals who provide ther, we conduct additional analysis looking into how more than 10 hours of care per week also work fewer labor supply is differentially affected by the intensity hours and earn significantly less, while this reduction of parental care. To do this, we divide our sample into is not observed among those who are not intensive three categories. First, some individuals do not pro- caregivers. These results indicate that the relation- vide care for their older parents. The second category ship between parental care and labor supply might consists of individuals who provide parental care, but not be linear; as the responsibility of caregiving gets do so for less than 10 hours per week. The third cat- larger, the effect on labor supply will be amplified. egory includes those who provide lengthy parental 278 CHAPTER 9 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E TABLE 9.12  Correlational effect on employment of the intensity of long-term care provision to parents, pooled sample from three countries Wage Casual/Self Weekly Hours Log Earnings Employed Employment Employment Worked (USD) (1) (2) (3) (4) (5) Panel A: Male Care Hours (1-10 Hrs) -0.086** -0.066 -0.019 -3.642 -0.136 (0.039) (0.047) (0.049) (2.359) (0.253) Care Hours (>10 Hrs) -0.117** -0.190*** 0.075 -5.634** -0.500* (0.047) (0.045) (0.056) (2.199) (0.265) Observations 25368 25368 25368 22137 12450 Dep. Var Mean 0.875 0.369 0.505 46.367 7.312 Mean Hours Cared 15.506 15.506 15.506 15.684 17.085 Panel B: Female Care Hours (1-10 Hrs) -0.127*** -0.092*** -0.033 -1.503 -0.100 (0.025) (0.018) (0.023) (1.329) (0.202) Care Hours (>10 Hrs) -0.151*** -0.118*** -0.034 -4.835*** -0.532** (0.029) (0.022) (0.027) (1.720) (0.230) Observations 32890 32890 32890 19955 12265 Dep. Var Mean 0.608 0.289 0.318 39.397 6.845 Mean Hours Cared 14.086 14.086 14.086 14.474 15.886 Notes: Robust standard errors in parentheses. *, **, *** indicate statistical significance at 10%, 5%, and 1%, respectively. Each regression includes individual and household-level characteristics: individual age, age-squared, highest educational level, number of children aged 0-6, 7-12, and 13-18 in the household, country fixed effects, and marital status. The regression includes individuals aged 40-59 from Poland, Colombia, and Indonesia. Individuals from Egypt are excluded because there is no information on the intensity of parental care. Log earnings indicate log annual earnings (in USD) conditional on being employed. CONCLUSION The population of older adults worldwide is projected to increase significantly in the coming years. This surge is expected to be accompanied by an increase in demand for informal care, the bulk of which will be borne by women. Because of this, analyzing how parental care differentially affects the labor supply between men and women is imperative in designing policies that may alleviate the forgone labor market opportunities that come with additional care responsibility. Our main analysis from all four countries confirms ment change, as in Indonesia, men may transition that women provide the large majority of care for ag- from formal work to informal work, while the same ing parents. We also observe a negative relationship transition is not observed in women. In Egypt, this between parental caregiving and labor supply, with study shows that the mere presence of a disabled a larger correlation among women. Further analysis parent in the household is not accompanied by a also shows a significantly larger effect among in- significant change in labor supply for either gender. tensive caregivers: those who provide parental care We observe some degree of heterogeneity in for more than 10 hours per week. When looking the relationship between long-term care and labor at specific countries, the results from Poland and supply, based on individuals’ educational levels and Colombia suggest that both men and women who the presence of young children in the household. provide parental care are significantly less likely to This provides us with a glimpse into the evidence be employed. In cases where there is no employ- of the increasing opportunity cost of parental care COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 9 279 among those who are highly educated and suggests of which has foregrounded developed countries. how the presence of children affects the relationship While the evidence presented in this study is not between care provision and labor supply. causal, our results can contribute to future policy With this study’s focus on developing coun- dialogue to reduce the gender imbalance in LTC tries, we add to the vast literature on the relation- burdens, particularly in rapidly aging regions where ship between informal care and labor supply, much the gender gap in labor market participation is large. REFERENCES 1. Agree, Emily, and Karen Glaser. 2009. “Demography of Informal “Labor Supply in the Terminal Stages of Lone Parents’ Lives.” Caregiving.” In International Handbook of Population Aging, Journal of Population Economics 25 (4): 1399–1422. https://doi. edited by Peter Uhlenberg. Springer Dordrecht. https://link. org/10.1007/s00148-012-0402-3. springer.com/book/10.1007/978-1-4020-8356-3. 14. Folbre, Nancy. 2014. “Who Cares? A Feminist Critique of the 2. Bittman, Michael, Trish Hill, and Cathy Thomson. 2007. “The Care Economy.” New York, NY: Rosa Luxemburg Stiftung. https:// Impact of Caring on Informal Carers’ Employment, Income and rosalux.nyc/a-feminist-critique-of-the-care-economy/. Earnings: A Longitudinal Approach.” Australian Journal of Social 15. Goldin, Claudia. 2004. “The U-Shaped Female Labor Force Issues 42 (2): 255–72. https://doi.org/10.1002/j.1839-4655.2007. Function in Economic Development and Economic History.” In tb00053.x. Investment in Women’s Human Capital and Economic Develop- 3. Bolin, Kristian, Bjorn Lindgren, and Petter Lundborg. 2008. “Your ment, edited by T. Paul Schultz. University of Chicago Press. next of Kin or Your Own Career?” Journal of Health Economics 27 https://scholar.harvard.edu/goldin/publications/u-shaped-fe- (3): 718–38. https://doi.org/10.1016/j.jhealeco.2007.10.004. male-labor-force-function-economic-development-and-eco- 4. Carmichael, Fiona, and Susan Charles. 2003. “The Opportunity nomic-history. Costs of Informal Care: Does Gender Matter?” Journal of Health 16. Heger, Dörte, and Thorben Korfhage. 2020. “Short- and Economics 22 (5): 781–803. https://doi.org/10.1016/s0167- Medium-Term Effects of Informal Eldercare on Labor Market 6296(03)00044-4. Outcomes.” Feminist Economics 26 (4): 205–27. https://doi.org/1 5. Chai, Huamin, Rui Fu, and Peter C. Coyte. 2021. “Unpaid Caregiv- 0.1080/13545701.2020.1786594. ing and Labor Force Participation among Chinese Middle-Aged 17. Himmelweit, Susan. 1999. “Caring Labor.” The Annals of the Adults.” International Journal of Environmental Research and Pub- American Academy of Political and Social Science 561 (1): 27–38. lic Health 18 (2): 641. https://doi.org/10.3390/ijerph18020641. https://doi.org/10.1177/000271629956100102. 6. Chen, Lu, Na Zhao, Hongli Fan, and Peter C. Coyte. 2016. “Infor- 18. Huang, Guogui, Fei Guo, and Gong Chen. 2021. “The Role mal Care and Labor Market Outcomes: Evidence from Chinese and Wellbeing of Female Family Caregivers in the Provision of Married Women.” Research on Aging 39 (2): 345–71. https://doi. Aged Care in China.” Social Indicators Research 159 (2): 707–31. org/10.1177/0164027515611184. https://doi.org/10.1007/s11205-021-02769-6. 7. Ciccarelli, Nicola, and Arthur Van Soest. 2018. “Informal 19. International Labour Organisation (ILO). 2018. “Care Work and Caregiving, Employment Status and Work Hours of the 50+ Care Jobs for the Future of Decent Work.” Geneva, Switzerland: Population in Europe.” De Economist 166 (3): 363–96. https://doi. International Labour Office. https://www.ilo.org/global/publi- org/10.1007/s10645-018-9323-1. cations/books/WCMS_633135/lang--en/index.htm. 8. Connelly, Rachel, Xiao-yuan Dong, Joyce Jacobsen, and Yaohui 20. Johnson, Richard W., and Anthony T. Lo Sasso. 2000. “The Trade- Zhao. 2018. “The Care Economy in Post-Reform China: Feminist off between Hours of Paid Employment and Time Assistance to Research on Unpaid and Paid Work and Well-Being.” Feminist Elderly Parents at Midlife.” Washington, D.C. : The Urban Institute. Economics 24 (2): 1–30. https://doi.org/10.1080/13545701.201 21. ———. 2006. “The Impact of Elder Care on Women’s Labor 8.1441534. Supply.” INQUIRY: The Journal of Health Care Organization, Pro- 9. Crespo, Laura, and Pedro Mira. 2014. “Caregiving to Elderly vision, and Financing 43 (3): 195–210. https://doi.org/10.5034/ Parents and Employment Status of European Mature Women.” inquiryjrnl_43.3.195. Review of Economics and Statistics 96 (4): 693–709. https://doi. 22. Kotsadam, Andreas. 2011. “Does Informal Eldercare Impede org/10.1162/rest_a_00426. Women’s Employment? The Case of European Welfare States.” 10. England, Paula, and Nancy Folbre. 1999. “The Cost of Caring.” The Feminist Economics 17 (2): 121–44. https://doi.org/10.1080/135 Annals of the American Academy of Political and Social Science 45701.2010.543384. 561 (1): 39–51. https://doi.org/10.1177/000271629956100103. 23. Liu, Huan, and Meng Wang. 2022. “Socioeconomic Status 11. Ettner, Susan L. 1995. “The Impact of ‘Parent Care’ on Female and ADL Disability of the Older Adults: Cumulative Health Labor Supply Decisions.” Demography 32 (1): 63. https://doi. Effects, Social Outcomes and Impact Mechanisms.” Edited org/10.2307/2061897. by Petri Böckerman. PLOS ONE 17 (2): e0262808. https://doi. 12. Fahle, Sean, and Kathleen McGarry. 2018. “Caregiving and org/10.1371/journal.pone.0262808. Work: The Relationship between Labor Market Attachment and 24. Liu, Lan, Xiao-yuan Dong, and Xiaoying Zheng. 2010. “Parental Parental Caregiving.” Innovation in Aging 2 (suppl_1): 580–80. Care and Married Women’s Labor Supply in Urban China.” https://doi.org/10.1093/geroni/igy023.2150. Feminist Economics 16 (3): 169–92. https://doi.org/10.1080/135 13. Fevang, Elisabeth, Snorre Kverndokk, and Knut Røed. 2012. 45701.2010.493717. 280 CHAPTER 9 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E 25. Mao, Shangyi, Rachel Connelly, and Xinxin Chen. 2017. “Stuck in Family Long-Term Care in Latin America.” IDB Working Paper Se- the Middle: Off-Farm Employment and Caregiving among Mid- ries No. 1105. Inter-American Development Bank (IDB). https:// dle-Aged Rural Chinese.” Feminist Economics 24 (2): 100–121. doi.org/10.18235/0002738. https://doi.org/10.1080/13545701.2017.1387670. 30. Terkoğlu, Özge İzdeş, and Emel Memiş. 2022. “Impact of Elderly 26. Moussa, Margaret Malke. 2018. “The Relationship between Care on ‘Sandwiched-Generation’ Women in Turkey.” New Per- Elder Care-Giving and Labour Force Participation in the spectives on Turkey 66 (May): 88–121. https://doi.org/10.1017/ Context of Policies Addressing Population Ageing: A Review of npt.2022.12. Empirical Studies Published between 2006 and 2016.” Ageing 31. UN Women. 2017. “Long-Term Care for Older People: A New and Society 39 (06): 1281–1310. https://doi.org/10.1017/ Global Gender Priority.” Policy Brief No. 9. UN Women. https:// s0144686x18000053. www.unwomen.org/en/digital-library/publications/2017/12/ 27. Nguyen, Ha Trong, and Luke Brian Connelly. 2014. “The Effect long-term-care-for-older-people. of Unpaid Caregiving Intensity on Labour Force Participation: 32. Van Houtven, Courtney H., Norma B. Coe, and Meghan M. Skira. Results from a Multinomial Endogenous Treatment Model.” 2013. “The Effect of Informal Care on Work and Wages.” Journal Social Science & Medicine 100 (January): 115–22. https://doi. of Health Economics 32 (1): 240–52. https://doi.org/10.1016/j. org/10.1016/j.socscimed.2013.10.031. jhealeco.2012.10.006. 28. Sinha, Aashima, and Ashish Kumar Sedai. 2022. “Why Care for 33. Wang, Yafeng, and Chuanchuan Zhang. 2017. “Gender the Care Economy: Empirical Evidence from Nepal.” Centre Inequalities in Labor Market Outcomes of Informal Caregivers for Applied Macroeconomic Analysis (CAMA) Working Paper near Retirement Age in Urban China.” Feminist Economics 24 (2): 31/2022. The Australian National University. https://doi. 147–70. https://doi.org/10.1080/13545701.2017.1383618. org/10.2139/ssrn.4078353. 34. World Health Organization (WHO). 2022. “Ageing and Health.” 29. Stampini, Marco, María Laura Oliveri, Pablo Ibarrarán, Diana Fact Sheet. World Health Organization. October 1, 2022. Londoño, Ho June Sean) Rhee, and Gillinda M. James. 2020. https://www.who.int/news-room/fact-sheets/detail/ageing- “Working Less to Take Care of Parents?: Labor Market Effects of and-health. ANNEX 9.1 TABLE 9A.1  Correlational effect on employment of providing long-term care to parents Male Female Hours worked/ Log earnings Hours worked/ Log earnings week (USD) week (USD) (1) (2) (3) (4) Panel A: Poland Long-term care -6.589*** -0.351** -3.436** -0.113* (2.541) (0.144) (1.637) (0.064) Observations 3608 2164 5477 3521 Dep. var. mean 43.091 8.514 38.558 8.381 Share of parent carer 0.005 0.007 0.010 0.010 Panel B: Indonesia Long-term care -5.209** -0.379** -0.572 -0.265 (2.267) (0.190) (1.625) (0.181) Observations 4194 4207 2799 2792 Dep. var. mean 39.070 6.580 34.158 4.878 Share of parent carer 0.017 0.017 0.055 0.055 Panel C: Colombia Long-term care -3.317 0.394 -5.838*** -0.017 (2.181) (0.489) (1.460) (0.348) Observations 13912 5650 10816 5094 Dep. var. mean 49.733 7.469 41.714 7.174 Share of parent carer 0.003 0.003 0.011 0.010 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 9 281 Male Female Hours worked/ Log earnings Hours worked/ Log earnings week (USD) week (USD) (1) (2) (3) (4) Panel D: Egypt Long-term care -2.045 -0.084 -0.843 -0.071 (1.345) (0.058) (1.581) (0.082) Observations 4194 2737 1192 660 Dep. var. mean 46.921 7.201 36.301 7.147 Share of parent carer 0.047 0.047 0.073 0.078 Notes: Robust standard errors in parentheses. *, **, *** indicate statistical significance at 10%, 5%, and 1%, respectively. Each regression includes individual and house- hold-level characteristics: individual age, age-squared, country fixed effect, highest educational level, number of children aged 0-6, 7-12, and 13-18 in the household, and marital status. Sample includes all individuals aged 40-59 in Poland, Egypt, Indonesia, and Colombia. TABLE 9A.2  Gender difference in the effect on employment of providing long-term care to parents Wage Casual/Self Weekly Hours Log Earnings Employed Employment Employment Worked (USD) (1) (2) (3) (4) (5) Panel A: Poland Parental Care -0.159*** -0.150*** -0.009 -3.544** -0.099 (0.041) (0.037) (0.030) (1.652) (0.065) Male 0.140*** 0.059*** 0.081*** 4.530*** 0.192*** (0.008) (0.009) (0.007) (0.220) (0.011) Parental Care * Male -0.065 -0.150** 0.086 -3.495 -0.306* (0.086) (0.073) (0.073) (3.054) (0.161) Observations 13421 13421 13421 9085 5685 Dep. Var Mean 0.677 0.502 0.175 40.358 8.432 Share of Parent Carer 0.012 0.012 0.012 0.008 0.009 Panel B: Indonesia Parental Care -0.022 -0.034* 0.013 -0.599 -0.246 (0.027) (0.021) (0.029) (1.621) (0.184) Male 0.229*** 0.151*** 0.081*** 4.113*** 1.589*** (0.009) (0.009) (0.011) (0.595) (0.058) Parental Care * Male 0.005 -0.060 0.066 -4.876* -0.302 (0.036) (0.052) (0.058) (2.774) (0.273) Observations 8332 8332 8332 6993 6999 Dep. Var Mean 0.848 0.286 0.558 37.104 5.901 Share of Parent Carer 0.046 0.046 0.046 0.042 0.042 Panel C: Colombia Parental Care -0.167*** -0.093*** -0.073*** -5.289*** 0.011 (0.029) (0.020) (0.026) (1.475) (0.351) Male 0.329*** 0.106*** 0.223*** 8.179*** 0.353*** (0.004) (0.005) (0.005) (0.212) (0.049) Parental Care * Male -0.068 -0.014 -0.054 0.819 0.296 (0.064) (0.056) (0.065) (2.620) (0.598) Observations 34880 34880 34880 24728 10744 282 CHAPTER 9 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E Panel C: Colombia (continued) Mean 0.709 0.270 0.439 46.226 7.329 Share of Parent Carer 0.010 0.010 0.010 0.007 0.006 Panel D: Egypt Parental Care 0.049* 0.020 0.030 -0.529 -0.109 (0.029) (0.023) (0.020) (1.566) (0.077) Male 0.610*** 0.450*** 0.158*** 10.902*** 0.186*** (0.009) (0.010) (0.008) (0.574) (0.027) Parental Care * Male -0.087** -0.039 -0.047 -1.810 0.024 (0.037) (0.039) (0.033) (2.051) (0.096) Observations 9611 9611 9611 5386 3397 Dep. Var Mean 0.570 0.414 0.155 44.571 7.191 Share of Parent Carer 0.049 0.049 0.049 0.051 0.050 Notes: Robust standard errors in parentheses. *, **, *** indicate statistical significance at 10%, 5%, and 1%, respectively. Each regression includes individual and house- hold-level characteristics: individual age, age-squared, highest educational level, number of children aged 0-6, 7-12, and 13-18 in the household, country fixed effects, and marital status. The regression includes individuals aged 40-59 from Poland, Colombia, and Indonesia. FIGURE 9A.1  Kernel density of long-term care provision by age and country COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 9 283 FIGURE 9A.2  Work hours and care hours gradient FIGURE 9A.3  Correlation between time spent on long-term care and employment 284 CHAPTER 9 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E FIGURE 9A.4  Correlational effect of providing long-term care on individual employment by choice of control group FIGURE 9A.5  Correlational effect of providing long-term care on annual earnings by choice of control group COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 9 285 FIGURE 9A.6  Correlational effect of providing long-term care on work hours by choice of control group 286 CHAPTER 10 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E Long-term Care Chapter 10 Health and Long-term Care Needs in a Context of Rapid Population Aging Chapter 11 Demand for and Supply of Long-Term Care for Older Persons in Low- and Middle-Income Countries 10 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 10 287 Health and Long-Term Care Needs in a Context of Rapid Population Aging Natalia Aranco Araújo a and Gisela M. Garcia b a Independent Consultant b IEG Human Development and Corporate Program, World Bank 288 CHAPTER 10 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E INTRODUCTION Population aging is a global phenomenon: people all over the world are living longer. Undoubtedly, this is good news. Extended longevity reflects the social and economic progress of societies; it is the result of innumerable advancements in medicine, nutrition, education, and technology, bringing increased opportunities for both in- dividuals and countries alike. One of the consequences of this aging process has sponding. The study was developed as a background to do with the challenges that an aging population report for the Independent Evaluation Group evalu- poses for countries’ health care and long-term care ation “World Bank Support to Aging Countries.” It arrangements. A longer life increases the probabil- focuses on developing countries that are aging fast, ity of disease, disability, and the loss of functional where anticipation and action are important, high- autonomy (WHO 2015). Thus, an increase in the lighting steps and policies that some countries have share of older individuals put pressures on health been adopting to address the challenges. care and long-term care systems, which need to A comprehensive review of all health and adapt to a new pattern of demand, characterized long-term care policies and actions taken by coun- by growth in the prevalence of chronic diseases tries is outside the scope of the study. Instead, we and in the rates of functional dependency. aim to highlight selected examples that illustrate These pressures are even more critical in de- diverse country responses to the different chal- veloping countries, given the unprecedented speed lenges posed by population aging. of their demographic transition and the resource The chapter is organized as follows. The Main constraints, limited-service coverage, barriers to Trends in Longevity section provides a brief over- access to services, and low-quality services that view of the trends in increasing longevity observed characterize many of these countries. Yet the pre- in countries in different stages of the demograph- dictability of the demographic and epidemiologi- ic transition over the last decades; it also shows cal changes provides governments with an unprec- where the gains in life expectancy have been trans- edented opportunity to intervene proactively to lated into gains in healthy life expectancy. The Ag- find effective and innovative solutions to address ing, Health, and the Challenges for Health Care the needs of the elderly (UN 2015a, 2017a). Systems section focuses on the consequences of Ensuring that the elderly enjoy the best pos- the demographic transition for the epidemiologi- sible health is not only an aim in its own right but cal profile of aging countries and the implications is also justified on economic grounds. Older peo- for health care systems; this section also provides ple’s contributions to societies, in the form of fi- some examples of how countries are responding nancial support to younger generations, help with to the challenges. Increasing Risk of Functional childcare, participation in labor markets, and vol- Dependency and Challenges for Long-Term Care unteerism, among others, can only materialize if Systems section discusses the implications of an gains in longevity are translated into years lived in aging population for long-term care services; this good health (UN 2017a).1 section also presents examples of countries’ efforts This literature review identifies key challenges in to address rising long-term care needs in a context health care and long-term care as populations age and of shrinking families. The Final Remarks and Key provides relevant examples of how countries are re- Policy Consideration section concludes. 1   The need to promote the health and wellbeing of older people, and to make sure that they are considered in the development process, ensuring supportive environments that allow them to contribute fully to society, was recognized in the 2002 Madrid International Plan of Action on Aging (MIPAA), adopted at the Second World Assembly on Aging (UN 2002), and was further reinforced in the United Nations 2030 Sustainable Development Goals (Zaidi 2016). COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 10 289 MAIN TRENDS IN LONGEVITY All over the world, people are living longer. As a result, the proportion of older people in the global population is increasing. Global average life expectancy at birth has risen from 47 years in 1950 to 72.6 years in 2019 (UN- WPP 2019). According to the United Nations medium variant projections,2 average life expectancy at birth is expected to reach 77.1 years by 2050 and the share of people aged 65 years or more worldwide to rise from 9 percent in 2019 to 12 percent in 2030 and 16 percent in 2050. In 2018, for the first time in history, the number of individuals 65 years old or more surpassed that of children younger than five years old (UN 2019) (Figure 10.1). FIGURE 10.1 Population by broad age group (thousands), estimates 1950-2020 and projections 2025-2100 1950 1970 1990 2010 2030 2050 2070 2090 Source: United Nations Population Highlights 2019, Figure 12. Note: 2050 projections are based on the United Nations “medium variant” assumptions. Life expectancy is also increasing at older ages. an aging country). Overall, the number of people 65 Thanks to progress made in disease management years old or more is projected to increase by almost and control, global average life expectancy at age 60 2 billion from 2020 to 2050; more than 60 percent of increased almost three years between 1990-1995 and this increase will happen in middle-income countries 2015-2020, rising from 18.1 to 20.7 (UNWPP 2019). (particularly lower-middle income). The popula- Thus, the share of the “oldest old” is also increasing, tion share of people older than 65 in middle-income and individuals aged 80 years or more went from countries is expected to double (from 8 percent to 16 representing 1 percent of the world’s total popula- percent). Countries in Latin America and the Carib- tion in 1990, to 1.9 percent in 2019, and an estimated bean, South Asia, and East Asia are expected to see 4.4 percent in 2050 (UNWPP 2019). the largest increases, more than doubling the percent- This trend is seen in all regions of the world, and age of people older than 65 in their populations from it is expected to continue, albeit at different levels and 2020 to 2050. In these three regions, that population rates, depending on the stage of the demographic tran- share will rise from 9 percent to 19 percent, from 6.2 sition in each country (see Box 10.1 for a definition of percent to 13.2 percent, and from 13.4 percent to 27.2 2   The medium variant projection is a probabilistic method that considers not only the historical fertility rates of each country but also the uncertainty regarding future trajectories, based on countries with similar past rates. For more details, see: “Definition of projection variants. World Population Prospects 2019,” United Nations Department of Economic and Social Affairs, Population Dynamics. Available at: https:// population.un.org/wpp/DefinitionOfProjectionVariants/#:~:text=Medium%2Dvariant%20projection%3A%20in%20projecting,of%20 changes%20in%20each%20variable. 290 CHAPTER 10 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E BOX 10.1 When is a country “aging”? A country is usually defined as “aging” when the share of people ages 65+ is above 7 percent, “aged” when it is 14 percent or more, and “super-aged” when it exceeds 20 percent. Other definitions are also used, such as the median age, or the old-age dependency ratio, which is the ratio of older dependents (people older than 64) to the working-age population (people ages 15–64). Cut-offs are arbitrary, however, because they do not consider how healthy and functional the “elderly” under such definitions are. The Pro- spective Old Age Dependency Ratio (Sanderson and Scherbov, 2005, 2007, 2010) measures population aging based on remaining life expectancy instead of the number of years lived. This measure reflects improvements in life expectancy over time, instead of an- choring old age to a fixed threshold, and suggests a slower increase in dependency than the traditional old-age dependency ratio. The 2015/16 Global Monitoring Report (GMR) (World Bank and IMF 2016) proposes a definition that combines trends in fertility and in the size of the working-age population. Based on these trends, countries are classified into four stages of the “demographic transition”: pre-dividend countries (where fertility is greater than 4 births/woman); early-dividend countries (where fertility is lower than 4 births/woman, but the working-age population is still increasing); late-dividend countries (with a shrinking working-age population, but where fertility fell only recently); and post-dividend countries (with a shrinking working-age population and where fertility fell below replacement level, or 2.1 births/woman, three decades earlier). The last two stages characterize aging countries. This chapter adopts the GMR definition to select aging countries, although it may refer to other definitions when appropriate. Source: World Bank (2021). percent, respectively (UNWPP 2019). 69 years, respectively, for the share of people aged In fact, the demographic change in developing 65+ to rise from 7 to 14 percent, in some late-divi- countries is taking place at an unprecedented speed. dend countries, the same change is taking place at a While in some developed countries like France, considerably faster pace (in some cases, less than 20 Sweden, and the United States, it took 115, 85, and years) (Figure 10.2). FIGURE 10.2 Years required or expected for the population share aged 65+ to rise from 7 percent to 14 percent Source: United Nations World Population Prospects 2019 Note: The graph shows the number of years needed for a country to transition from “young” to “old” (as measured by a rise in the share of its 65+ population from 7 percent to 14 percent of the total). For late-dividend countries, calculations are based on the “medium-variant” UN projections; it should be noted that UN estima- tions are published at five-year intervals. COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 10 291 The figures and trends shown so far lead At older ages, the gap between life expectancy to questions about the quality of the years of life and healthy life expectancy widens, indicating that expectancy gained in the past decades. In other most of the years lived in bad health or disability words, are we living healthier or just longer? The are concentrated later in life. On average, life ex- extent to which gains in life expectancy translate pectancy at 65 years old is 17.6 years; 73 percent of into new opportunities for both individuals and those years are expected to be lived in good health. countries depends largely on the quality of these This proportion has remained stable since 1990, additional years. The concept of healthy life ex- and does not show much variation across country pectancy, which reflects the number of years that a groups, as depicted in Figure 10.3. person can expect to live in good health, free of dis- Women live longer than men in all regions of ease or disability, can shed some light on this ques- the world, but they also spend more years in poor tion. From 1990 to 2019, globally, 83 percent of the health due to differences in risk factors and disease years of life gained were years free of disability and patterns. Women are more vulnerable than men to disease, with the largest increases seen in countries disabling non-fatal afflictions, while men are more that are less advanced in the demographic transi- affected by conditions that have high death rates tion, as seen in Table 10.1 (GBD 2019). (Luy and Minagawa 2014). Gender differences, TABLE 10.1 Years gained in life expectancy and healthy life expectancy at birth, 1990-2017. Both sexes, by demographic group 1990 2019 Variation 1990-2019 LE0 HALE0 LE0 HALE0 Years gained in LE0 Years gained in HALE0 (%) Years gained in good health Pre-dividend 53.0 45.5 64.5 56.3 11.5 10.8 94.1% Early-dividend 65.1 56.5 71.2 61.9 6.1 5.3 88.1% Late-dividend 70.8 61.7 75.8 66.1 5.0 4.4 88.2% Post-dividend 75.0 65.1 80.4 69.4 5.4 4.3 79.9% Source: Authors’ analysis based on GBD and collaborators Results Tool 2019. Note: LE0: life expectancy at birth; HALE0: healthy life expectancy at birth. Values were computed as the unweighted average for the countries in each group. FIGURE 10.3 Healthy and unhealthy life expectancy at age 65 years, 1990 and 2017 Source: Original analysis based on the Global Burden of DIsease (GBD) Results Tool 2019. 292 CHAPTER 10 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E both in life expectancy and healthy life expectancy, pectancy, so that people are living longer with mild/ are narrower in younger countries. This discrepan- non-fatal conditions, but rates from fatal diseases cy could be explained by a higher risk of mortality remain stable (Manton 1982). in both sexes, a high prevalence of infectious dis- Evidence in this regard is not conclusive. Re- eases at old ages (such as tuberculosis), a recent in- search in high-income countries seems to favor the crease in the prevalence of behavioral risk factors hypothesis of a compression of morbidity when (especially tobacco consumption), and low levels of looking at severe disability rates, but results are health care coverage (GBD Collaborators 2017; Ma- contradictory when looking at mild disability and thers et al. 2015). Over time, gender differences are chronic disease prevalence among older people, expected to widen in developing countries as well with some studies pointing to an expansion of mor- (He, Goodkind, and Kowal 2016). bidity due to the improvements in secondary and Some researchers argue that the definition of tertiary prevention treatments that allow people “old age” should not be based on a universal thresh- with disabilities to live longer (Howse 2006; Chat- old, but instead should consider health status mea- terji et al. 2015; Rechel et al. 2009). Research for sures and, more generally, whether the aging of the low- and middle-income countries is much scarcer population is achieved because of health improve- and difficult to reconcile, as results vary depending ments or not. When indicators of life expectan- on the methods and data used (Chatterji et al. 2015). cy and indicators of morbidity and mortality are Last but not least, policy responses to popu- combined, it is shown that, although older people’s lation aging should consider the important distri- health is improving worldwide, as the fatality and butional issues that come with population aging severity of age-related diseases decreases, the onset (World Bank 2021). These include gender gaps; in- of age-related diseases occurs at younger ages in rel- tergenerational disparities; spatial (e.g., rural versus atively less-developed countries (Chang et al. 2019). urban) differences; and socioeconomic inequalities. This is reflected in the fact that, despite having low- Gender inequalities that can be exacerbated by pop- er levels of life expectancy, less-developed countries ulation aging are rooted not only in gender differ- still experience a high burden of age-related diseas- ences in life expectancy, but in differences in care es. For example, a 76-year-old individual in Japan responsibilities and their implications for women’s is equivalent to a 46-year-old individual in Papua own health and wellbeing, women’s participation in New Guinea, and in both cases, they show the same the labor market and, ultimately, their income secu- age-related burden of disease as the global average rity in old age. As care demands increase in aging 65-year-old individual (Chang et al. 2019). societies, care responsibilities usually fall to women. The discussion as to whether increasing lon- When unpaid, such demands further limit women’s gevity has been accompanied by increases in labor market participation or impose a double work healthy life expectancy is closely linked to the de- burden, when women are responsible for both paid bate on the compression or expansion of morbidity. and domestic labor (World Bank 2015). Similarly, In this sense, if, on the one hand, increments in life policies should consider that younger generations expectancy are led by a delay in the onset of illness- will face greater risks of inequality in old age than es, then, unhealthy years would be concentrated at current retirees (OECD 2017). Spatial disparities the end of the life cycle, and we would be witnessing are already reflected in large inequalities in access a compression of morbidity and an improvement in to health and social care between old people living the health indicators of older people (Fries 1980). in urban and rural areas. These gaps are expected If, on the other hand, the rise in life expectancy is to widen, as younger generations leave rural areas mainly led by a drop in death rates from fatal dis- looking for better opportunities. Important so- eases, but the age-specific pattern of incidence re- cioeconomic inequalities are found among adults mains stable, then we may witness an expansion aged 50+ years in low- and middle-income coun- of morbidity, as the average duration of disability tries when it comes to medical frailty. This has been increases (Gruenberg 1977; Olshansky 1991). A measured using the Study on global AGEing and third alternative, the so-called dynamic equilibrium adult health (SAGE) data from China, Ghana, In- theory, posits that it is neither the postponement of dia, Mexico, Russia, and South Africa (Hoogendijk the onset of disease nor the postponement of death 2018). Such inequalities may widen further with from diseases, but the delay in the progression of changing gender norms and cohabitation patterns disease, that characterizes the increases in life ex- and increasing informality in these countries. COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 10 293 AGING, HEALTH, AND THE CHALLENGES FOR HEALTH CARE SYSTEMS Highlights • Population aging inevitably comes with rising prevalence of chronic diseases, as well as geriatric neurological conditions, such as dementia and Alzheimer’s disease. Depression and musculoskeletal disorders in older people are also major causes of disability. • Several characteristics of the epidemiological transition in developing countries complicate the picture, in- cluding the double burden of communicable and non-communicable diseases; the increased risk of multimor- bidity; under-diagnosis and lack of affordable, timely, and effective treatments; and the increased adoption of harmful habits such as smoking, alcohol use, lack of physical activity, and unhealthy diets. • Thus, population aging comes with at least three main challenges for health care systems in developing coun- tries: (i) guaranteeing coverage, access, and affordability of services in a context of increasing demand; (ii) shifting the service model from a curative-based approach towards a person-centered, holistic, and integrated approach that encourages a continuum of care through the life course, with greater emphasis on primary health care, including health promotion and prevention activities; and (iii) keeping costs under control with- out losing attention to quality. • Health care systems need to adapt and prepare. How people age is, to a large extent, determined by their health earlier in life, and the choices they made when young. Countries should therefore promote healthy lifestyles throughout the entire life course and work towards better management of chronic diseases. Universal health coverage in countries with aging populations should consider that older adults not only have different health care needs than the younger population but are often also less able to afford care. Changing patterns of disease: the increase ly-dividend countries. In late-dividend countries, in chronic conditions and the importance chronic diseases already represent more than 80 of multimorbidity percent of the total burden of disease, almost as As people live longer, the main causes of mortality much as in post-dividend countries. and morbidity shift from communicable or infec- Following current demographic projections, tious diseases to chronic diseases,3 in a process that these trends are expected to accelerate during the is known as the epidemiological transition (Omran coming decades. Among people aged 70 years or 1971). It is estimated that 15 percent of the increase more, chronic diseases already explain more than in the global burden of mortality and disability 90 percent of the burden of disease in post- and from 1990 to 2016 could be explained by popula- late-dividend countries. The World Health Organi- tion aging (Gakidou et al. 2017). In 2019, chronic zation (WHO) estimates that, by 2060, 83 percent diseases explained 64 percent of the burden of dis- of global deaths will be caused by chronic diseases, ease, measured by Disability Adjusted Life Years an increase of more than 10 percentage points com- (DALYs),4 and are the first cause of disability glob- pared to 71 percent in 2016. This means an increase ally (GBD 2019). As seen in Figure 10.4, while the of almost 44 million people dying from chronic burden of chronic diseases in 1990 was already high diseases in 44 years. More than 75 percent of that in post-dividend countries, in the last two decades increase will take place in middle-income countries the increase has been pronounced in late- and ear- (WHO Global Health Estimates 2018). 3   In this chapter, the term chronic disease refers to conditions that are non-communicable, long in duration, and generally progress slowly. 4   The Disability Adjusted Life Years metric is used to measure the current health status of a population with respect to an ideal health situ- ation, where everyone lives free of illness and disability to an old age. It is calculated as the sum of the Years of Life Lost due to Premature Mortality and the Years of Life Lost Due to Disability. A more detailed explanation can be found at: https://www.who.int/healthinfo/ global_burden_disease/metrics_daly/en/ 294 CHAPTER 10 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E FIGURE 10.4 Burden of disease, by disease type, 1990-2017 (% of total DALYs) Source: Original analysis based on GBD Results Tool 2019. Note: Country classification is based on the World Bank typology presented in Box 10.1. The prevalence of behavioral risk factors for obesity (Prince et al. 2015). In Middle Eastern and chronic conditions increases with age. Evidence North African countries, chronic diseases account- shows that this is true for obesity, hypertension, ed for 69 percent of the total burden of disease and insufficient vegetable and fruit intake, and insuffi- 78 percent of total deaths in 2019, due to the combi- cient physical activity (Wu et al. 2015; Rampal et nation of rapid population aging and the evolution al. 2007). The only risk factors that show a negative of chronic disease risk factors, such as rising obe- gradient with age are alcohol and tobacco consump- sity and the persistence of elevated rates of hyper- tion (Wu et al. 2015). In part, this might be a con- tension, which are among the highest in the world sequence of higher mortality among smokers (Storr (GBD 2019; Sibai et al. 2017). The prevalence of hy- et al. 2010). However, tobacco use affects dispropor- pertension among older adults increased 20 percent tionately more older adults in developing countries, between 1995 and 2004 in Tunisia and almost tri- with prevalence higher among middle-aged persons pled in Lebanon during the period 2000-2010 (Sibai than among the youngest (Storr et al. 2010). et al. 2017). Similar trends are seen in other aging The link between population aging and the ris- countries; in Malaysia, for example, the prevalence ing prevalence of chronic diseases becomes more of hypertension among individuals aged 30 or more evident in those countries that are aging rapidly, increased from 32.9 percent in 1996 to 40.5 percent as evidenced by the observed trends in conditions in 2004 (Ramli and Taher 2008). that are known to be strongly age-dependent, such Rising prevalence rates of diabetes and chronic as cardiovascular diseases, diabetes, dementia, or obstructive pulmonary disease are also associated chronic obstructive pulmonary disease (Prince et with population aging. In the case of diabetes, the al. 2015). Mortality and morbidity due to cardio- number of people living with the condition is pro- vascular diseases has increased in China, India, and jected to increase 51 percent from 2019 to 2045. Al- Russia in the latest years, due to an increase in the though in absolute terms most people with diabetes prevalence of elevated cholesterol, diabetes, and are concentrated in middle-income countries, the COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 10 295 largest relative increase is expected to take place in understood (Byers and Yaffe 2011). low-income countries, where the number of people Older people are more likely to experience living with the disease is expected to increase by 150 multiple health conditions at once (Afshar et al. 2015; percent (International Diabetes Federation 2019). He, Goodkind, and Kowal 2016). The epidemiological In the case of chronic obstructive pulmonary transition not only leads to an increasing burden of disease, it is projected that the burden of disease will chronic diseases as causes of death and disability, but increase by 89 percent from 2004 to 2030 (Prince also to an increased prevalence of multimorbidity. In et al. 2015). The problem is particularly serious in Colombia, a 2018 study found multimorbidity South Asian countries, where chronic respiratory present in approximately half of patients of all ages diseases rank second as cause of morbidity and pre- with chronic diseases. This means that between 4. 5 mature mortality among older people, representing and 6 million of the approximately 11 million 17 percent of total DALYs. Population aging, deteri- patients per year who accessed health services in oration of environmental conditions, and increased Colombia for at least one chronic disease were pollution due to rapid urbanization in some coun- patients with multimorbidity. Yet programs for an tries, as well as high rates of smoking, are behind integrated approach to multimorbidity are incipient these large figures (Bishwajit et al. 2017). or exist only as pilots (Alfonso-Sierra et al. 2018). As the share of the older population rises, the Comparable patterns appear to hold, irrespective importance of geriatric neurological conditions, of the region and the stage of demographic transition such as dementia and Alzheimer’s disease, also of the country. In Uruguay, 43 percent of adults 60 increases. While in 2017, neurological conditions years old or more suffer from more than one chronic explained only 4 percent of DALYs for the world condition (Aranco and Sorio 2019); in Mexico the population of all ages, the figure reaches 6 percent figure is 33 percent (Lopez-Ortega and Aranco 2019). among those older than 70 years, and 9 percent Data for Asia show rates of multi-morbidity of more among those older than 80 (GBD 2019). With the than 30 percent in India (for those aged 70 years or projected increases in the share of older adults in more), 40 percent in Viet Nam, and 58 percent in the population, the burden of dementia is expected Bangladesh (for those aged 60 or more) (Khanam et to rise. In 2015, 58 percent of those with dementia al. 2011; Ha et al. 2015; Pati et al. 2014). Data for lived in low- and middle-income countries, and it China show that the so-called metabolic risk is expected that by 2050 this figure will reach syndrome, which is a combination of cardiovascular almost 70 percent (ADI 2015). risk factors including abdominal obesity, high blood Due to their debilitating effects and their pressure, increased glucose level, and dyslipidemia contribution to disability, musculoskeletal dis- rises from 14 percent among middle-aged adults to 46 orders deserve special consideration. Even though percent among older people (He et al. 2009). their mortality rates are low, musculoskeletal Evidence for Kuwait also shows an increase in multi- diseases are one of the main causes of chronic morbidity with age: the proportion of individuals pain and functional impairment, hence having who suffer from diabetes, hypertension, and heart important detrimental effects on quality of life disease simultaneously rises from 3.6 percent among among the elderly (Woolf et al. 2012). In fact, those aged 50–59 to 9.4 percent among those 60–69, although they represent only 6 percent of DALYs, then surges to 20.9 percent among persons older than these diseases account for 17 percent of Years Lost 70 (Boutayeb, Boutayeb, and Boutayeb 2013). due to Disability (YLDs) (GBD 2019). Common clinical practice and guidelines that Depression in older adults is a major cause of focus on managing a single disease can negatively disability, although highly underdiagnosed and, as impact the quality of life and health of people a result, left untreated. In fact, evidence shows suffering from multiple chronic conditions. The that depression symptoms are more difficult to interactions, not only between the biological impacts recognize in older adults, as they usually manifest of different conditions, but also among their in concurrence with other physical and recommended treatments and/or medications, can neurological factors that are thought to be “part represent important health risks for individuals of the normal aging process”, such as loss of suffering them (WHO 2015). Different medications mobility, chronic pain, or dementia (Rodda, prescribed by different providers could trigger Walker, and Carter 2011). In particular, the link adverse drug reactions that are harmful for patients between dementia and depression is well (van der Heide et al. 2015). The fragmented approach recognized, although the direction of causality of multiple specialists treating single diseases and the mechanisms behind it are not yet fully separately not only increases the risk of drug inter- 296 CHAPTER 10 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E actions but also reduces adherence to crucial from 38.3 percent in high-income countries to 52.6 treatments and impairs the percep-tion of safety. percent in middle-income countries and reaching 66.8 percent in low-income countries (International The epidemiological transition in Diabetes Federation 2019). Data for Argentina, Bra- developing countries zil, Chile, Colombia, and Mexico show that 40 per- cent of hypertensive individuals are not diagnosed Some characteristics of the epidemiological (OPS 2017; Arredondo and Recaman 2018). Simi- transition in developing countries could worsen lar figures have been found for Egypt and Viet the aging challenges ahead. First, many of today’s Nam (37.5 percent and 48.4 percent, respectively) older adults in developing countries were (Ibrahim and Damasceno 2012). International exposed to infectious diseases in their early evidence shows that, even among those who are childhood and survived thanks to medical diagnosed with key chronic diseases, a very small advances that were not necessarily related to an fraction of people follows an appropriate treatment improvement in their socio-economic conditions, and have the condition under control (Ibrahim and a circumstance that may have unknown effects Damasceno 2012). Data from China show that on older people’s health during the coming years more than half of diabetes cases are undiagnosed (Palloni, Pinto-Aguirre, and Pelaez 2002). Indeed, and that, among those who are diagnosed, less than poor health conditions in early life have been 40 percent have the recommended glycemic shown to be linked to an increased risk of controls (Chen et al. 2018). In the case of high diabetes, obesity, and heart disease in adulthood blood pressure, figures are even more worrisome: (He, Goodkind, and Kowal 2016). This is only 8 percent of diagnosed individuals in Egypt especially true for older adults who were born in and 10.7 percent in Viet Nam have the disease the 1930s and 1940s in countries that under control with adequate theraapies (Ibrahim experienced rapid mortality reductions at that and Damasceno 2012). time (such as Brazil, Chile, Costa Rica, and Fourth, globalization and modernization, which Mexico, among others) (McEniry 2014; McEniry have prompted widespread adoption of harmful and McDermott 2015). habits such as smoking, alcohol use, lack of physical Second, although chronic diseases are the activity, and unhealthy diets, have unfolded faster in leading causes of mortality and disability among developing than in developed countries. This has older people, the importance of infectious diseases fueled the increase in chronic diseases and should not be underestimated. This double burden epidemiological change (Miranda et al. 2009). of disease increases the risk of multimorbidity, with Obesity, for instance, is an increasing global consequences for the morbidity and mortality phenomenon, with 39 percent of adults aged 18 and patterns of older adults that are not yet fully over being overweight or obese in 2016, an almost understood (WHO 2015). Older people suffering three-fold increase from 1975. The World Bank from chronic diseases are more exposed to estimates that more than three-quarters of obese or infectious diseases; type 2 diabetes, for example, is overweight individuals live in middle-income a known risk factor for the onset of tuberculosis, countries (Shekar and Popkin 2020). The percentage leading to serious health consequences (WHO of deaths from all causes attributable to a high body 2011a). The recent pandemic generated by the mass index rose from 4.7 percent in 1990 to 8.9 spread of the Severe Acute Respiratory Syndrome percent in 2019 (GBD 2019). In upper-middle- Coronavirus 2 (SARS-CoV-2) has highlighted the income countries the reported increase was from 5.8 vulnerability of older adults, particularly those with percent to 10.4 percent, and from 2.4 percent to 7.9 previous chronic conditions, in the face of the percent in lower-middle-income countries. The surge of new infectious diseases (or the upsurge of problem is particularly acute in Asia Pacific and in older ones). countries in the Middle East and North African Third, under-diagnosis of key medical condi- region (Prentice 2006), where public policies have tions and the lack of affordable, timely, and effec- stimulated the consumption of highly energy-dense tive treatments are still common problems in many foods. This has led to a change in dietary habits that, developing countries (He, Goodkind, and Kowal in conjunction with increasingly sedentary lifestyles, 2016). Around half of the people with type 2 diabetes has contributed to worrying levels of overweight and are unaware of their condition (International Diabe- obesity in populations (Kilpi et al. 2013). Papua New tes Federation 2019). Rates of underdiagnosis show Guinea, for instance, has an overweight/obesity a clear income gradient across countries, ranging prevalence of 60 percent (Shekar and Popkin 2020). COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 10 297 Prevalence rates of obesity among adults older than to deal mainly with acute and infectious diseases, 18 years reach almost 30 percent in Egypt and Libya, towards a holistic and integrated approach that en- 24 percent in Tunisia, and more than 20 percent in courages a continuum of care throughout the life Algeria and Morocco (WHO 2018a). course, with a greater emphasis on primary health Tobacco consumption, the single most import- care and health promotion and prevention activi- ant avoidable risk for chronic diseases, had been ties; and (iii) last, but not least, keeping costs under increasing in developing countries, particularly control without losing attention to quality. among men, until 2018, in stark contrast to the ob- Effective health care coverage for older served trends in developed countries. However, the individuals in low- and middle-income countries is far most recent report on tobacco use estimates that from universal, signaling the existence of critical tobacco consumption rates reached their peak in unmet needs and important gaps in access to and 2018, and projects a decreasing trend up to 2025 availability of health care services (Goeppel et al. through all country income groups (WHO 2019). 2014). Despite significant progress made during recent Only five countries globally are currently experi- years, important inequalities still exist both between encing increasing trends in tobacco use: Congo, Le- and within nations when it comes to timely and sotho, and Niger, in Africa, and Egypt and Oman in effective access to care (Bloom et al. 2019) (Box 10.2). the Middle Eastern region (WHO 2019). Older people belonging to the lowest socioeconomic Yet disaggregating the data by sex shows quite a strata in Latin America and the Caribbean, for different picture, as in some places women are instance, are more likely to have difficulties accessing increasingly taking up smoking. For example, in the health care services they need (Huenchan 2010). Croatia, the share of women who smoke has increased Inequalities in coverage and access, and poor by 9 percentage points in the past 20 years, from 27 performance in terms of effectiveness and quality of percent in year 2000 to 36 percent in 2020; an care, are critically salient when looking at chronic increase of almost 5 percentage points was seen in the disease management. Effective coverage rates among Slovak Republic and in Türkiye.5 In Latin Ame- older people with chronic conditions range from rica, tobacco use prevalence was very similar among 21 percent in Mexico to 48 percent in South Africa male and female adolescents aged between 13 and 15 (Atun et al. 2015; Glassman et al. 2010). Similar years old (15 percent and 12 percent for young men results have been found elsewhere (see for example, and women respectively), with Chile showing the Tham et al. 2 018 for Asia-Pacific; Atun et al. 2015; highest tobacco use among young women (26 and Glassman et al. 2010 for Latin America; percent), followed by Argentina (25 percent) and and Kujawska 2017 and Rechel et al. 2012 for Mexico (18 percent) (OECD and WHO 2020). Europe and Central Asia). Barriers to access in the form of unaffordability of services or lack of Challenges for health care systems and transportation are often cited by older adults in countries’ responses low- and middle-in-come countries as the main reasons for not being able to access health care The increasing number of people with chronic when needed (Table 10.2). health conditions exerts pressures on health systems, The lack of formal coverage leads to the neglect of which will need to adapt to satisfy not only a health or to high out-of-pocket expenditures when care growing demand but also a different type of is needed, imposing a heavy financial burden on demand. In this section, we discuss what these individuals and households. In low- and lower- challenges mean for health systems in rapidly aging middle-income countries in Asia-Pacific, out-of- countries and present some of the steps taken by pocket expenditures in 2015 accounted for almost 50 developing countries to overcome them. per-cent of total health expenditures, increasing the Population aging brings about at least three risk of household impoverishment in the face of main challenges for health care systems in deve- chronic and long-term illness (WHO 2018b; OECD loping countries: (i) first, guaranteeing coverage, and WHO 2018a). For example, in Viet Nam, more access, and affordability of services in a context of than one-quarter of older people are not insured, increasing demand; (ii) second, shifting their care and more than half refrain from seeking medical model from a curative-based approach, designed treatment due to high costs (Hoang and Duong 2018). 5   See World Bank Open Data website, retrieved from: https://data.worldbank.org/ 298 CHAPTER 10 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E BOX 10.2 Health and socioeconomic status The patterns of diseases and risk factors vary, not only across countries at different stages of development, but also across individuals in different socioeconomic strata within countries. Individuals from lower socioeconomic levels are usually exposed to a larger number of risk factors and have limited access to health care and to the resources that could allow them to follow a healthier lifestyle (Blas and Kurup 2010). These disadvantages develop in early childhood (sometimes, even in utero) and accumulate throughout the life course, widening the health gap between individuals from different socioeconomic strata (Blas and Kurup 2010; Wilkinson and Marmot 2003). This positive association between socioeconomic status and individual health has been widely documented in the literature and has been proved to hold for different countries and periods, regardless of the measures of health and socioeconomic status used (Goldman 2001).* However, given differences in the access to and quality of health care, as well as in social, cultural, and economic environments, the degree (and in some cases even the sign) of the association may differ between developed and developing countries (Blas and Kurup 2010). In developed countries, individuals in higher socioeconomic groups were the first to adopt harmful behaviors such as smoking or unhealthy eating; soon lower socioeconomic groups followed their behavior, while prevalence among the higher socioeconomic groups started to drop as educated individuals learned the health consequences of these habits. The evidence for low- and middle-income countries is less clear, although recent evidence suggests that tobacco and alcohol con- sumption decrease with socioeconomic status while the reverse is true for physical inactivity. Results for dietary pattern are more mixed: individuals in higher socioeconomic groups tend to consume more fruits, vegetables, fish, and fiber but also more fats, salts, and processed foods (Allen et al. 2017). Yet the evidence suggests that, as country income increases, the burden of over- weight/obesity shifts from the wealthy to the poor (Jones-Smith et al. 2011, 2012b; Shekar and Popkin 2020). Differences in risky behaviors lead to differences in disease prevalence. Various studies in developing countries have shown a higher prevalence of cardiovascular diseases among lower socioeconomic groups (Vathesatogkit, Batty and Woodward 2014; Lustigova et al. 2018), and a study for Latin America shows that individuals with no education are twice as likely to suffer from Alzheimer’s disease as individuals with some education (Nitrini et al. 2009). Similar results have been shown for Brazil, China, and India (Scazufca et al. 2008; Libre Rodriguez et al. 2008). The evidence for type 2 diabetes is not as clear: some studies find that, in developing countries, higher prevalence is associated with higher socioeconomic status, in stark contrast to the evidence in developed regions (Xu et al. 2017); however, other studies point in the direction of a negative socioeconomic gradient (see for example Sandoya 2016 for Uruguay and Suwannaphant et al. 2017 for Thailand). Multimorbidity prevalence is also shown to be negatively associated with education in the multi-country study from Afshar et al. (2015), even when adjusted for age and gender. Note: Grossman (2015) provides an extensive review of recent studies carried out in this field. Table 10.2 Reasons given by adults aged 60 years or older for not accessing health care services, by country income category Upper-middle Lower-middle High income Lower income income income Could not afford the visit 15.7 30.9 60.9 60.2 No transport 12.1 19.3 20.7 29.1 Could not afford transport 8.7 12.9 28.1 33.0 Health care provider’s equipment inadequate 11.2 10.5 14.1 16.7 Health care provider’s skills inadequate 19.0 8.3 7.8 13.1 Previously treated badly 23.8 8.7 7.9 8.3 Did not know where to go 12.2 9.7 9.8 7.8 Was not sick enough 21.5 31.8 27.3 25.8 Tried but was denied health care 20.0 16.2 8.3 8.5 Other 43.8 22.5 23.5 13.9 Source: WHO 2015, based on WHO World Health Surveys 2002-2004 (http://www.who.int/healthinfo/survey/en/) COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 10 299 Chronic diseases, common in old age, can re- well as adequate financial protection for older adults, sult in catastrophic health expenditures. Older especially the poorest and most vulnerable, without people with diagnosed chronic diseases face cat- compromising the quality of services. astrophic health expenditure even in some of the Within this context, countries have taken steps wealthiest countries in Europe (Arsenijevic et al to improve the coverage, accessibility, and afford- 2016). In developing countries, data from WHO’s ability of health care services for the older popula- SAGE surveys shows that households with at least tion. Some examples are presented below. one member aged 50 years or more have a higher probability than those with only younger members of facing catastrophic health care expenditures, hav- Improving coverage and effective access ing to borrow money from relatives to afford health care services, or being impoverished by health care a. Extended coverage or subsidized health costs (He, Goodkind, and Kowal 2016). Similar re- care contributions sults were found in Latin American and Caribbean countries (Knaul et al. 2011), in Asia (Somkotra and In most countries, health care coverage has been Lagrada 2009), and in a more recent global stock- traditionally linked to formal employment, leaving take (Hoang-Vu Eozenou et al. 2021). those outside the formal workforce and who cannot The path towards universal health care cover- afford private insurance (including informal work- age should take distinctive characteristics in coun- ers but also pensioners and retirees) unprotected. tries with aging populations, as older adults not only During recent decades, countries have taken ac- have different health care needs than the younger tions to correct this situation, improving population population but are often less able to afford care (Sad- coverage, either as part of overall system reforms, or ana, Soucat, and Beard 2018). The Universal Health focusing on specific vulnerable groups. Coverage Political Declaration from 2019 includes Health care reforms in Latin American and Ca- clear commitments to promoting healthy ageing. ribbean countries in the early 2000s were motivated Specifically, it commits governments to “scale up ef- by the need to address the institutional and financial forts to promote healthy and active ageing, maintain fragmentation of systems that led to important gaps and improve quality of life of older persons and to re- in coverage (Atun et al. 2015). For example, in Uru- spond to the needs of the rapidly ageing population, guay, the health care reform of 2007 extended insur- especially the need for promotive, preventive, curative, ance coverage to previously uncovered population rehabilitative and palliative care as well as specialized groups, including retirees. By 2018, 83.5 percent of care and the sustainable provision of long-term care.” 6 pensioners and retirees were covered by the system.7 This implies adapting health care systems to a differ- The reform also increased coverage to other popula- ent and increasing demand through: (i) improving tion groups, including the children and spouses of timely access to and affordability of health care ser- formal workers, who were previously excluded from vices and medicines for conditions that are com- insurance. In Bolivia, special provisions to improve mon in old age; (ii) delivering person-centered ser- older people’s coverage were adopted: in 2006, the vices that address older people’s health needs, with country created the Health Insurance for the Older an emphasis on primary and community care; (iii) Adult (SSPAM), which covers adults 60 years old improving the management of chronic conditions or more who have no other health insurance. Simi- (including cognitive conditions, such as dementia) larly, in Costa Rica, health care provision is free for and multimorbidity; and (iv) the development of a older individuals (among other vulnerable groups), trained workforce that responds to the needs of ag- regardless of formal coverage (PHCPI Costa Rica ing populations. Additionally, it will require efforts n.d.; Montenegro Torres 2013). to ensure the financial sustainability of services, as In Asia, China has also taken steps to increase 6   United Nations General Assembly (UNGA), Political Declaration of the High-level Meeting on Universal Health Coverage “Universal health coverage: moving together to build a healthier world”, September 2019. Retrieved from: https://www.un.org/pga/73/wp-content/up- loads/sites/53/2019/07/FINAL-draft-UHC-Political-Declaration.pdf 7   Original calculations based on data from the Ministry of Public Health (2019) (retrieved from: https://www.gub.uy/ministerio-salud-pu- blica/datos-y-estadisticas/datos/evolucion-afiliados-fonasa-2007-mayo-2019), and statistics from the Uruguayan Social Security Bank (retrieved from: https://www.bps.gub.uy/2692/principales-indicadores.html). 300 CHAPTER 10 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E coverage for older people, with the establishment of Salvador 2018). Increasing coverage is not a guarantee two new health insurance schemes for urban and rural of effective access to services, particularly when populations (Smith and Majmundar 2012). health care systems are fragmented (Guanais et al. Subsidized or exempted contributions for the el-derly 2018). have been another instrument to facilitate their Aiming at improving older people’s access to access to health care in the region. In Thailand, for health care services and medicines, some countries example, before the introduction of free health care have established free or subsidized consultation coverage for the entire population in 2006, and medicines for older adults, giving them priori- vulnerable groups, including older individuals, were ty in access and treatments, and even providing free exempted from copayments for health services transportation to health care facilities. Free or sub- (Hughes and Leethongdee 2007). Viet Nam, a sidized consultation and medicines for older people country that has made important efforts to are provided both in countries that are advanced in achieve universal health care coverage through the the demographic transition, such as in Azerbaijan, creation of a social health insurance program, has Bermuda, the Czech Republic, Hungary, Jamaica, compulsory affiliation but fully subsidized premiums Kazakhstan, Oman, Tunisia, and Uruguay, but also for vulnerable groups, including older populations in countries where the transition is less advanced, (Somanathan et al. 2014). for example Iraq, Sudan, or Uzbekistan (UN ECE Examples in Eastern Europe can be found in 2017; UN ECWA 2017; Bonilla-Chacín et al. 2018). Bulgaria, where older adults - among other In Uruguay, the 2007 reform sought to improve older vulne-rable groups - are covered for free by the people’s access to services and medications through national mandatory health insurance, or in the subsidization of service fees and medications Serbia, where individuals aged 65 years or above and the right to one or two free checkups per year have free access to the public health care services (depending on the age of the older person) (UN and (Pitheckoff 2017; Stokic and Bajec 2018). HelpAge International 2012). In Bermuda, hospital- In spite of these efforts, the lack of public re- ization services are publicly covered for individuals sources can be a barrier for implementation of well- older than 75, and in Jamaica user fees were elimi- intended policies, as in the case of Lebanon, which nated at public health facilities (UN ECLAC 2017). had to block enrolment in its optional health care Programs that provide free medications, checkups, plan for older people (launched in the year 2000), and other benefits for senior citizens living with due to underfunding (UN 2017b). In most Eastern chronic disease are in place in many Latin Ameri- European countries, health care systems are can and Caribbean countries, such as Argentina, characterized by constrained resources, inadequate The Bahamas, Barbados, Belize, Bermuda, Brazil, infrastructure, large out-of-pocket expenditures, and Costa Rica, Cuba, Mexico, Paraguay, Saint Maarten, insufficient coordination between local and na-tional and Trinidad and Tobago (WHO 2015; UN ECLAC authorities, resulting in fragmented systems 2017; UN and HelpAge International 2012). (Kujawska 2017; Rechel et al. 2012). Since 2005, Chile has in place the Universal Access to Explicit Guaranteed Entitlements plan b. Improving affordability and accessibility (AUGE as per its name in Spanish), that guaran- tees timely and quality access to treatment for 85 Regardless of coverage, evidence shows that a large conditions and prevents beneficiaries from incur- share of older individuals faces affordability and ac- ring catastrophic expenses (Aguilera et al. 2015; cess restrictions when it comes to the actual use of Superintendencia de Salud de Chile 2019). Many of services. For example, in Peru, data from 2012 these conditions relate directly or indirectly to old- show that almost 14 percent of people over 80 years er adults. The Seguro Popular in Mexico also has a old say they do not have access to the health care special fund that covers catastrophic expenditures they need due to cost or transport barriers for complex illness (Atun et al. 2015). Similarly, the (Encuesta de Salud y Bienestar del Adulto Mayor de Catastrophic Medical Insurance scheme in China, Perú 2012); in El Salvador, 44 percent of implemented at the national level in 2015, provides individuals aged 80 or more face barriers related to extra reimbursement for individuals who incur ex- the quality and timing of services, such as not cessive out-of-pocket medical expenses, providing getting an appointment in due time, lack of extra financial protection to those who are at risk personnel, or being poorly treated at the institution of catastrophic expenditure, including older people (Encuesta de Hogares de Propositos Múltiples de El (Fang et al. 2019). The National Resources Fund COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 10 301 (Fondo Nacional de Recursos) in Uruguay has a sim- that is timely, equitable, integrated, and efficient (Kie- ilar aim, although its application is more restricted, ni et al 2018). In most developing countries, health covering health care expenses for high-complexity care systems were designed to deal with acute and procedures and very expensive medicines.8 infectious diseases and hence focus on “finding the In many Asian countries, such as aging Viet problem and fixing it” (WHO 2015, p.93). With the Nam, but also relatively younger countries, such as upsurge of chronic diseases, this approach is rapidly Bhutan or Nepal, senior citizens are given priority in becoming obsolete. For people with chronic diseas- treatment, which allows them to shorten waiting es, the ultimate goal of care should be to improve the times, and they have access to subsidized care quality of life and to enhance functional status through (Williamson 2015). In 2012, Kuwait issued a “priority secondary prevention (Grumbach 2003). pass” card that gives preference to older citizens and Many health care systems around the world reduces waiting periods for them at primary health have failed to incorporate into their model of care centers (WHO 2015; UN ESCWA 2017). Bangladesh, the typical problems that matter to older people, Cambodia, China, India, Malaysia, Maldives, such as frailty, chronic pain, and the loss of func- Mongolia, Nepal, Solomon Islands, and Tuvalu are tional and sensory capacity. This leads to the onset among the countries that have provisions to ensure of many health problems that could be prevented or that older people in remote areas can access health reversed if they were identified in a timely manner care through, for example, providing free transpor- (WHO 2018a). This is closely linked to the quality tation to facilities (Williamson 2015). of services, which is defined by the WHO as “the However, and despite the efforts described above, degree to which health services for individuals and older people in many settings still face barriers in populations increase the likelihood of desired health seeking to access timely and quality health care. For outcomes and are consistent with evidence-based instance, Armenia has offered free primary health care professional knowledge.”9 Although this problem is since 2006, and vulnerable groups in the country - common worldwide, it is more pronounced in mid- including the elderly - are entitled to a package of dle- and lower-income countries (WHO 2015). For basic health care benefits free of charge. However, the example, it is estimated that the low quality of ser- lack of resources means that sometimes these benefits vices (and not the lack of access) is responsible for cannot be provided; this has led to the development of almost 60 percent of amenable deaths in low- and a “shadow market” of paid medical services (WHO middle-income countries (Kruk et al. 2018) 2009; Tonoyan and Muradyan 2012). In other Few and poorly trained human resources, in- settings, such as Latvia, the administrative procedures adequate infrastructure, outdated technology and required to be exempted from medical co-payments treatments, and insufficient material resources are and to obtain other benefits could act as a barrier for common characteristics of health care systems in older adults (European Commission 2019). Evidence most developing settings, and these factors impact for the Asia Pacific region also shows that access and older people’s access to effective, quality, and timely affordability restrictions persist in many countries healthcare. This is reflected, for example, in long there (Lagomarsino et al. 2012; Van Minh et al. 2014; waiting lines, lack of affordable transportation op- Zaidi et al. 2017). tions to health care centers, and physical barriers to access, among other obstacles (WHO 2018a). The management of chronic diseases, a key com- Adapting health care systems to an older ponent of any health care system that deals with an ag- population and a new epidemiological ing population, encounters many obstacles in low- and profile without losing focus on quality middle-income contexts. Best-practice recommenda- Improvement in health care delivery requires a de- tions that work in more developed countries usually liberate focus on the quality of health services, which fail in developing nations, due to: (i) little support for involves providing effective, safe, person-centered care self-management of chronic conditions derived from a 8   See Fondo Nacional de Recursos Uruguay webpage at http://www.fnr.gub.uy/quienes_son_beneficiarios 9   World Health Organization Fact Sheet. “Quality Health Services”. July 20th, 2020. Retrieved from: https://www.who.int/news-room/fact- sheets/detail/quality-health-services 302 CHAPTER 10 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E poor communication relationship between providers vented by eliminating the common risk factors behind and patients, which in turn is often rooted in a lack these conditions, such as tobacco and excessive alco- of time, willingness, and appropriate training among hol consumption, unhealthy diet, and lack of physical providers and is exacerbated by patients’ lack of health activity (WHO 2008). About 35 percent of dementia education, confidence, and motivation to change; (ii) cases could also be prevented by tackling nine mod- the inability of providers to follow guidelines due to ifiable risk factors: education, hypertension, obesity, lack of awareness, time constraints, or practical consid- hearing loss, late-life depression, diabetes, physical erations; (iii) poor follow-up care and information-re- inactivity, smoking, and social isolation (ADI 2019a). cording protocols; (iv) absence of community support; WHO “best buys” are simple and cost-effective (v) patients’ inability to afford the costs and cover the interventions to prevent death and disability brought distance to health care facilities; (vi) lack of resources, by chronic diseases. These “best buys” include popu- including medicines, equipment, laboratory supplies, lation-wide interventions, such as tobacco and alco- and human resources; and (vii) lack of coordination hol consumption control and regulation (through the across public, private, and alternative providers in con- application of increased taxes, regulation of publicity, texts where patients seek care from multiple providers, and laws that regulate consumption and purchase); according to their beliefs, their ability to pay, and the the promotion of healthy diets (through, for exam- seriousness of their condition (Lall et al. 2018). ple, actions to reduce salt and trans-fat consump- Within this context, health care systems need to tion); mass media communications campaigns to adapt. Taking a person-centered, holistic approach increase awareness about the importance of healthy that focuses on prevention, as well as on the integra- diets and physical activity; and timely screening for tion and continuity of care across different services, prevention of cardiovascular diseases, diabetes, and with a renewed emphasis on primary care services, is cancer (WEF and WHO 2011).10 Countries can select paramount (He, Goodkind, and Kowal 2016; WHO cost-effective interventions from the list of best buys, 2015; WHO 2018a). Developing countries have taken considering affordability, implementation capacity, decisive steps to transform their health care systems to and feasibility, according to national circumstances, deal with the new morbidity and mortality profiles of as well as the measures’ impact on health equity. their populations. Some examples are presented below. Following these recommendations, countries all over the world are making important efforts to im- a. Better management of chronic disease prove early diagnosis and management of chronic through prevention and timely diagnosis diseases and to reduce the costs of treatment. In fact, arguably, part of the observed increase in chronic Evidence shows that cost-effective policies, imple- disease prevalence rates documented in the previous mented through coordinated and integrated efforts section could be explained by better diagnosis. How- in which health care systems work in coordination ever, implementation of these type of policies has with other sectors and actors, taking a comprehensive been uneven across the globe, particularly in devel- and horizontal approach, could reduce the mortality oping countries, with recommended policies more and morbidity burden of chronic diseases (Samb et prevalent in late-dividend countries (Figure 10.5). al. 2010). Community, workplace, and school-based While some countries have taken a multisectoral programs that encourage physical activity and healthy approach (integrating health care systems with oth- diets, and individual health care interventions aimed er sectors such as education), others have adopted a at the population who are at risk of developing chron- more compartmentalized approach, focusing on spe- ic diseases (primary prevention) or at those who are cific measures to improve the situation of those with already sick in order to prevent complications (sec- chronic diseases. Still, 60 percent of countries world- ondary prevention) have proved to be effective (WHO wide have a national multisectoral strategy to man- 2017a). Up to 80 percent of heart disease, stroke, and age chronic diseases and their risk factors (WHO type 2 diabetes and 40 percent of cancers could be pre- Global Health Observatory Data Repository 2020). 10   Most updated information on the list of the “Best Buys” interventions recommended by WHO is available at World Health Organization (2017). “Tackling NCDs: ‘Best Buys’ and other recommended interventions for the prevention and control of non-communicable diseases”. Retrieved from: https://apps.who.int/iris/bitstream/handle/10665/259232/WHO-NMH-NVI-17.9-eng.pdf COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 10 303 FIGURE 10.5 Percentage of countries that have a national strategy for chronic diseases and their risk factors, by demographic stage Source: Authors’ analysis based on data from the Global Health Observatory Data Repository (WHO 2020). Note: Calculated as “Country has an operational, multisectoral national chronic diseases policy, strategy or action plan that integrates several chronic diseases and their risk factors.” For more information on what this entails, see: https://www.who.int/data/gho/indicator-metadata-registry/imr-details/4680 As an example of the above, Uruguay reorient- to encourage screening and early diagnosis among ed its entire healthcare system to increase coverage the elderly have also been taken in Armenia, Hun- and address the increasing prevalence of chronic gary, and Portugal (UN ECE 2017). diseases, strengthening primary care services and In Latin America, medications to control hyper- encouraging prevention. Broad system reforms that tension and diabetes were subjected to discounts and included measures to encourage prevention, timely subsidies to guarantee access to adequate treatment diagnosis, and access to treatment for chronic dis- in Uruguay (World Bank Independent Evaluation eases were also taken in other countries in Latin Group 2018). Similarly, in Brazil, medicines for diabe- America, such as Argentina, Brazil, Chile, Colom- tes control are subsidized (Kaselitz, Rana, and Heisler bia, Costa Rica, Cuba, Mexico, Peru, or Venezuela 2019). Among Middle Eastern countries, Morocco (Atun et al. 2015). In 2015, Croatia launched a reor- has made efforts in the treatment and control of di- ganization of hospitals that consists in reclassifying abetes by implementing a multidisciplinary strategy, hospital beds from acute beds to palliative, chron- in which a team of specialized professionals provides ic, prolonged, and day-care beds as part of a sys- individualized treatments to patients; the program tem-wide approach (European Commission 2019). encourages self-management of the disease in collab- Efforts towards improving diagnosis, treat- oration with non-governmental organizations and ment, and control targeted at specific chronic dis- subsidizes diabetes medication (Salhab et al. 2004). eases have taken place during recent years in all The case of Sri Lanka, with the establishment of regions. In Europe and Central Asia, some aging the Well Woman Clinics, is a noteworthy example countries such as Armenia, Belarus, Latvia, Lithu- in Asia. These clinics were introduced in 1996 with ania, and Slovenia can be highlighted as examples. the objective of screening perimenopausal wom- In some cases, these actions are specifically targeted en for breast and cervical cancers, as well as other at the elderly. Belarus, for example, created the so- chronic diseases, such as diabetes and hyperten- called health schools for the third age to educate old- sion. Yet coverage levels fall short of the initial aim, er people on common disease treatments and health reaching only 50 percent of women in the 35-year- promotion, while persons older than 40 in Albania old cohort, 30 percentage points below the target of are entitled to free annual checkups to identify risk 80 percent. A severe shortage of human resources factors and behaviors early (UN ECE 2017). Actions for health, especially midwifes, who are expected to 304 CHAPTER 10 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E visit homes and invite women to Well Woman Clin- et al. 2018). If revenues stemming from these taxes ics, explains this result. (UNFPA WWP 2019). are earmarked for health-related priority programs Due to historically high rates of communica- and actions targeted at individuals from low-income ble diseases, efforts towards the prevention and strata, potential health benefits could be even larger. treatment of chronic diseases in Africa have been According to the latest data from WHO (2020), meagre except in a few countries, among which two more than 80 percent of countries globally have pol- aging countries, Seychelles and Mauritius, and rela- icies aimed at discouraging unhealthy dietary habits tively younger South Africa and Tanzania stand out and encouraging physical activity.12 Although most (Aikins et al. 2010). Management of chronic dis- of them are late or post-dividend countries, a few eases is a priority in Seychelles, for example, where countries that are earlier in the demographic transi- several initiatives have been adopted such as a mul- tion have taken steps in the same direction (Figure tisectoral approach to manage chronic diseases and 10.6). Data show that between May 2016 and May their risk factors, encouraging a person-centered 2018, 24 countries or local jurisdictions implement- approach, improving health care quality standards, ed taxes on artificially sweetened beverages (World and investing in the quality of human resources for Cancer Research Fund International 2018). the long term (WHO 2018c). The budget allocated In Chile, for example, the Ministry of Health to NCD prevention activities in the country almost promoted measures that aim at controlling rising doubled from 2009 to 2015, rising from 6.4 percent levels of obesity (and its related complications) to 12.5 percent of general government spending, a among the population, such as the imposition of a clear sign of a shift in the healthcare paradigm from tax on sweetened beverages, changes in food label- a curative to a health promotion perspective (Walel- ing, advertising restrictions, and school-based in- ing et al. 2018). Targeted interventions, such as the terventions (OECD 2019a). Uruguay has followed approval of a tax on sweetened beverages and a di- a similar path, restricting the selling of unhealthy abetes management pilot program, have also been foods at schools and salt consumption in restau- supported (Waleling et al. 2018). Mauritius has had rants and bars and approving legislation on food la- a national program for the control and prevention beling, which came into effect in February 2021.13, 14 of non-communicable diseases in place since 1987 Many countries in Asia have also approved taxes and conducts focused surveys every five years.11 on sweetened beverages. Jurisdictions advancing Stand-alone population-based interventions on this front include rapidly aging countries such targeted at the prevention of the main risk factors as Malaysia, Sri Lanka, and Thailand, although Sri for chronic diseases, such as consumption taxes ap- Lanka took a step back in 2018 when it ordered a plied to tobacco or unhealthy food products, have reduction in taxes. South Africa and Seychelles have also proved to be cost-effective policies instruments also recently approved a tax on sweetened beverag- to discourage harmful habits. Although there have es.15 Other regulatory measures used by countries been claims against their use based on equity rea- to improve diet quality include the use of front-of- sons and the fact that consumption taxes are regres- package labeling; nutrient profiling; school-based sive in nature (i.e., their impact is proportionately food regulations and education, market, and retail larger on those at the lower end of the income distri- solutions; and marketing controls and regulations bution), these arguments are countered by evidence (Shekar and Popkin 2020). In Latin America, for that shows that low-income consumers benefit most instance, several countries including Chile, Mexi- from these policies in terms of health gains (Sassi co, Peru, and Uruguay have recently implemented 11   See World Health Organization, All Africa at https://allafrica.com/stories/201901070565.html 12   See preceding footnote 13   From 2007 to 2015, Uruguay also received the financial and technical assistance of the World Bank to implement a chronic disease prevention project, that focus on strengthening the country’s capacities to detect chronic diseases on a timely manner, and to reduce exposure to selected risk factors across the population (World Bank PPA Uruguay). 14   See El País Newspaper. “Etiquetado de alimentos: cambian criterios y suben valores que definen cuándo un contenido es excesivo.” Janu- ary 28th, 2021. Available at: https://www.elpais.com.uy/informacion/salud/etiquetado-alimentos-cambian-criterios-suben-valores-de- finen-contenido-excesivo.html 15   See World Health Organization, Achievements in Seychelles website at http://open.who.int/2016-17/country/SYC COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 10 305 FIGURE 10.6 Percentage of countries with policies that target main chronic diseases risk factors, by type of policies and country’s demographic stage Source: Original analysis based on data from the Global Health Observatory Data Repository (WHO 2020). Note: For selected policies. front-of-package labeling with warning labels for ferent results worldwide. Although the WHO rec- excessive amounts of sugars, total fats, saturated ommends tobacco taxes of at least 75 percent of fats, trans fats, and sodium. retail price, in 2018 only 32 countries met this cri- Salt-reduction initiatives, ranging from consum- terion, an increase of only 10 countries since 2008; er education to taxation on high-salt-content food, as a result, 90 percent of the global population lives were first taken mostly by high-income European in countries where tobacco taxation is deemed to countries and have lately become a popular strategy be insufficient by international standards (WHO to fight the rise of chronic diseases among countries 2017b). In Latin America and the Caribbean, Uru- of all income levels. Initiatives have shifted from vol- guay is regarded as the regional leader in the fight untary to mandatory or legislative approaches world- against tobacco, while important advances have wide (Trieu et al. 2015). There are suggestive exam- also been made in Brazil and Panama (Burki 2017); ples of private sector involvement, as in Uruguay progress has been somewhat slower in Argentina, and Iran, where bakeries voluntarily agreed to lower Chile, Peru, Guatemala, or Mexico (Burki 2017; Ya- the salt content in their products (Salhab et al. 2004; maguchi et al. 2017). Earmarked tobacco taxes or World Bank Independent Evaluation Group 2018). at least some strategic reinvestment of tax revenues Tobacco control policies, like banning smoking towards health care and health promotion activities in enclosed places, increasing taxes, and requiring are present in at least 43 countries, most of them health warnings on tobacco packages, have been low- and middle-income countries.16 The adoption adopted at different paces and levels and with dif- of tobacco control policies has been sluggish in Eu- 16   See Tobacco-free Kids website at www.tobaccofreekids.org 306 CHAPTER 10 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E rope and Central Asia countries, with the exception Among developed countries, the case of Isra- of Türkiye, and in Arab countries, where although el is probably one of the best-known examples of legislation is in place, actual compliance is low; how an aggressive investment in primary care can given the alarming rates of tobacco consumption improve population health in an efficient way. The among countries in the region, this lack of politi- country’s primary care level is grounded on its large cal commitment is worrisome (Smith and Nguyen workforce (approximately 7,000 primary care doc- 2013; Maziak et al. 2014). tors in 2017) and is strongly community-oriented, with an extended geographical network of primary b. Strengthening primary and care community clinics which also provide ambula- community care tory specialist care (Rosen, Waitzberg, and Merkur 2015; Clarfield et al. 2017). The role of primary care as a key strategy to achieve Also worth noting is the case of Canada, which universal coverage and a person-centered and inte- has recently implemented an innovative person-cen- grated approach to health services has been widely tered primary care model in several provinces that recognized, at least since the WHO and UNICEF focuses on health promotion and chronic disease Alma Ata Declaration of 1978.17 Strengthening pri- management, including mental health services. The mary care means giving more prominence to health initiative – called My Health Teams – brings together promotion and prevention, ambulatory and outpa- teams of providers to care for patients with chronic tient care, homebased interventions, community conditions. These teams are comprised of health care participation, and to coordination and integration workers who have different skills and expertise, usu- between primary and higher care levels (WHO 2015). ally a physician, a nurse practitioner, and other health A good primary care system should show the professionals who may be in the same clinic, commu- following characteristics: (i) be accessible and pro- nity, or even connected online. The aim of the teams vide the first point of contact for the patient with is to be the first point of care for people with chronic the health system; (ii) be comprehensive, providing conditions, even providing care after hours, while en- a range of services that extend from prevention and gaging patients in their own care and ensuring acces- health promotion to timely detection of diseases, sible, comprehensive, and coordinated person-cen- while avoiding unnecessary referrals to specialists; tered care (Chateau et al. 2017; OECD 2019b). (iii) deliver person-, family-, and community-cen- Developing countries like Sri Lanka, Thai-land, tered care, allowing for continuity of care through- or Viet Nam in Asia; Brazil, Chile, Costa Rica, and out the life cycle and taking into account the specific Uruguay in Latin American; and Seychelles in Sub- characteristics and needs of individuals; (iv) coor- Saharan Africa have undergone a series of health dinate services not only across different levels, but care reforms that built on the expansion of also across different service providers and different primary care to achieve universal coverage, while at sectors outside the healthcare system; and (v) work the same time adapting to the new epidemiological longitudinally, in the sense that the relationship profile of the population through a reinforced em- between the patient and the health care provider is phasis on the prevention and management of chron- sustained over time (Guanais et al. 2018). ic conditions (Atun et al. 2015; Waleling et al. 2018). Successful examples of strengthening primary In Sri Lanka, as part of the government’s strat- care initiatives have been implemented in developing egy to manage the increase in chronic diseases, in countries, often engaging the community. Although 2011 the Ministry of Health established Healthy many of them concentrate on child and maternal Lifestyle Centers. These centers function within health, initiatives that focus on chronic diseases and the country’s primary care units, with a twofold mental health management are also found (Kruk et objective: detecting risk factors for chronic diseases al. 2010). Enhancing the capacity of the first level of through timely screening and improving access to care for prevention, risk factor detection, and timely specialized care for those diagnosed or at great risk diagnosis of chronic diseases is among the goals of of developing a chronic condition (Mallawaarach- such initiatives (Almeida et al. 2018). chi at al. 2016). Although initially the target pop- 17   See WHO (1978) “Declaration of Alma-Ata”. Retrieved from: https://cdn.who.int/media/docs/default-source/documents/almaata-decla- ration-en.pdf?sfvrsn=7b3c2167_2 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 10 307 ulation comprised individuals aged 35 to 65 years, and hypertension, health education, and, if needed, the age upper limit was later removed (UN ES- to help patients adhere to recommended treatments CAP 2016), and by 2016 the country’s 826 centers and stay motivated (Dmytraczenko and Almeida had screened 25.5 percent of the population aged 2015; PHCPI Brazil n.d.). Given their close links to 40 to 65 (Mallawaarachchi at al. 2016). Screening the families, these agents allow for continuity of care; of younger people is also important when thinking moreover, they can refer families to other social and about an aging population, as the prevention and welfare programs, improving socio-sanitary coordi- timely management of chronic conditions earlier in nation. Since 2006, the program has increased the life improve health at older ages. attention given to the elderly, training health profes- Countries that are less advanced in the demo- sionals in matters related to older adults’ health, such graphic transition, such as Peru and Mexico, have as the assessment of functional status and frailty, and also invested in strengthening the primary care the establishment of self-help groups in the commu- focus on chronic diseases. For example, the public nity that encourage healthy behaviors (WHO 2015). health insurance under Peru’s Seguro Integral de Sa- Also, in Brazil, the example of Rocinha, the lud includes an extensive primary care package that biggest slum (favela) in Rio de Janeiro, is worth covers many chronic diseases; in Mexico, the Seguro mentioning. In 2012, a pilot project was launched to Popular expanded beneficiaries’ access to ambulato- support the health of older people living in the city’s ry health care services, including a benefits package largest slums through the design of individualized that covers a set of primary care interventions, mak- health plans and follow-up in close collaboration ing them free at the point of delivery (Han 2012). with three family health clinics in the communities. An important reliance on the community and The evaluation of the program showed good results the family are at the base of many of these programs. and was extended to the whole state of Rio de Ja- The role of community-driven initiatives is crucial, neiro (WHO 2015). Family-based models are also not only to reduce barriers to access, but also to pro- the foundation of the primary care delivery system vide timely and quality care at all levels, to help in in Costa Rica, El Salvador, and Estonia. In Costa the integration of primary, secondary, and tertiary Rica, multidisciplinary health teams (EBAIS as per care levels, and to raise awareness and education the Spanish acronym) are responsible for providing about the consequences of aging and the best ways care to 1,000 families each and pay home visits to to deal with them (HelpAge International 2014a). the elderly; in Estonia, family doctors receive spe- Because of cultural similarities with the communi- cial training in the management of chronic diseas- ty they care for, community health workers are seen es in order to reduce specialist care; moreover, any as an important link between the formal health care patient with a chronic disease has the right to get system and those in need of care (Woldie et al. 2018). an appointment within three days of requesting it Community health workers are as good as, or some- (PHCPI Estonia n.d.; PHCPI Costa Rica n.d.). times even better than, formal health care workers in The role of the community in the delivery of delivering preventive, promotive, and curative ser- primary care is also highlighted by successful exam- vices, particularly when they are integrated with the ples in Asia, where the prevention and management formal health care system and work in settings that of chronic diseases is increasingly gaining attention provide in-service training, financial incentives, ad- (HelpAge International 2014a). In Thailand, pri- equate infrastructure and supplies, regular monitor- mary health care, with a prominent role for “vil- ing, supervision, and evaluation (Woldie et al. 2018). lage health volunteers,” is considered the backbone In Brazil, the Programa Saude da Familia serves of the health care system. These workers focus on as a role model for a community approach to pri- prevention and health promotion and act as the mary care. Multidisciplinary primary health teams, first point of contact between primary care and composed of a physician, a nurse, and four to six the rest of the health care system (Tejativaddhana community health agents, are assigned up to 1,000 et al. 2018). As part of a strategic plan to reorient families each, acting as a point of entry to the health its health care system towards the management of care system for them. The community health agents chronic diseases, Malaysia has made an important have a key role in the health delivery system: each of investment in primary health care that focuses on them is assigned up to 150 families and is respon- prevention and screening for major chronic diseas- sible to visit them once a month to provide basic es at the community, workplace, and school levels, preventive care, including screening for diabetes also relying on voluntary workers (Yiengprugsawan 308 CHAPTER 10 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E et al 2017). In Viet Nam, intergenerational self-help Evidence regarding the impacts of primary care clubs for the elderly promote healthy aging activities approaches to managing chronic diseases in devel- and provide health checkups and homecare oping countries is scarce but promising (Kruk et support services, along with other services, such al. 2010; Macinko et al. 2009; Macinko et al. 2016). as self-care trainings, information on older people’s There is evidence that community-based programs rights, and help building social networks or generating are effective for the prevention and control of an income. The project has the support of HelpAge chronic disease, particularly for tobacco cessation International and works with volunteers, who are programs and diabetes and blood pressure control trained by paid retired health care professionals. (Jeet et al. 2017). In South Africa, after the imple- More than 16,000 caregivers provide ongoing care mentation of a nurse-run chronic disease program, for at least 10,000 people in the 3,000 clubs that exist disease control improved by 68 percent in the case in the country (AHWAIN 2020). of patients with hypertension, 82 percent in the case A few relatively younger countries have started to of patients with diabetes, and 84 percent in asthmat- move in the same direction, in response to the recent ic individuals (Coleman, Gill, and Wilkison 1998). increase in the prevalence of chronic conditions and In Latin America, countries that have invested in foreseeing the demographic and epidemiological building a strong primary care sector, like Chile, changes to come. In South Africa and rural Ethiopia, Costa Rica, and Uruguay, have seen an increase in for instance, chronic disease clinics run by trained the use of preventive screening services for breast nurses have proved to be effective for the and cervical cancer (Almeida et al. 2018). A recent management of chronic diseases (Mamo et al. 2007; study focusing on Brazil, Colombia, El Salvador, Ja- Coleman, Gill, and Wilkison 1998). In Bangladesh, maica, Mexico, and Panama shows that, along with the 13,000 community clinics that exist in rural areas multimorbidity rates, the quality of the primary care have recently incorporated chronic disease screening sector is a key variable in determining people’s abil- as part of their functions. In India, chronic diseases ity to pay for the medical services they need (Mac- and palliative care have become part of the remit of inko et al. 2019). Community-led initiatives have the country’s “Accredited Social Health Activists,” proved to be effective in the management and con- who serve as community health providers in rural trol of diabetic patients (Kaselitz, Rana, and Heisler areas (WHO 2018b). Also in India, the state of Kerala 2019). Different evaluations show that implementa- has made important advancements in strengthening tion of the Programa Saude da Familia program in primary care with a strong focus on community Brazil has expanded home health care among the engagement. Although, until recently, efforts were elderly, particularly among functionally dependent mostly focused on maternal and child health, the state individuals from lower socioeconomic groups who has started to plan for an aging population, with the have been hospitalized in the previous year, hence development of geriatric wards and facilities.18 contributing to reducing health inequalities (Thumé Faith-based organizations are an important part et al. 2011; Honorato dos Santos de Carvalho 2013). of community initiatives, although more data is Despite the progress, there is still much to do. In needed to estimate their coverage (Kagawa et al. many countries, primary care clinics are regarded by 2012). A 2015 report by The Lancet highlights the role many individuals as low-quality care, leading them of these organizations in Africa, particularly Christian to bypass this level and go directly to the second and Islamic facilities (Olivier et al. 2015). In Asia, a level of care. Different studies have highlighted the recent study for Viet Nam finds that Catholic and public mistrust of primary care in Asian countries, Buddhist organizations are important in providing such as China, Malaysia, Sri Lanka, and Viet Nam, both short- and long-term support for older adults in for example, but also in richer countries like Japan need of health and social services, including medical or the Republic of Korea (Chen et al. 2019; screenings, healthcare counseling, psychological Yiengprug-sawan et al. 2017; Lee et al. 2019; Gaál et support, and even surgery funding aid (Hun et al. al. 2015). Sometimes, particularly in lower-resource 2023). Anecdotal evidence also points to an important contexts, this perception is justified. A study for role of thee organizations in Latin America. Viet Nam found that 35 percent of the more than 18 See Improving Primary Health Care Performing Initiative. Country cases and promising Practices. Retrieved from: https://www.improv- ingphc.org/country-case-studies-promising-practices COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 10 309 100,000 commune health clinics that exist in the The situation is even more acute in developing country do not have doctors, and many of them suf- countries, although some progress has been made fer from a lack of medicines and equipment, ham- during recent years, particularly in rapidly aging pering service delivery (Somanathan et al. 2014). countries. In 2017, only 32 countries worldwide had The lack of trust in primary care is also reflect- adopted a national dementia plan, most of them ed in patients’ unwillingness to follow physicians’ high-income countries that are already advanced in recommendations about diets, physical activity, and the demographic transition, except for Chile, Costa screening (Yiengprugsawan et al. 2017). Improving Rica, Indonesia, and Mexico (ADI 2017).19 By 2019, patients’ health literacy, but also the communica- 30 more countries had national plans in develop- tion skills of providers, is crucial to overcome these ment (ADI 2019) (Figure 10.7). Some countries, problems. For example, a study done in six Latin such as China, Kuwait, Lebanon, Morocco, Tuni- American and Caribbean countries finds that a large sia, and Uruguay, have recently launched national percentage of primary care patients report commu- mental health programs that include older people as nication problems with their providers, such as: not part of their target populations (UN ESCWA 2017; being given the opportunity to ask questions (28 UN ESCAP 2017; MSP Uruguay 2020). percent of respondents), not having things explained The lack of a clear commitment regarding pub- clearly enough (26 percent), leaving with health prob- lic funds sometimes prevents plans from advancing lems unsolved (31 percent), and not receiving help to and being put into practice (ADI 2019a). For exam- coordinate care (61 percent) (Macinko et al. 2015). ple, the Uruguayan government approved a Mental Another issue is the lack of integration between Health Plan that includes dementia in 2017, but un- primary care and other levels of care, which com- til 2020 no resources had been allocated, prevent- promises the continuity of care, a crucial aspect of ing its implementation.20 Peru approved a national the management and prevention of chronic diseases law to develop a national plan for dementia in 2013; (Dmytraczenko and Almeida 2015; Singh, Cassels while the plan was supposed to be developed in and Travis 2018). Although some efforts have been 90 days, six years later it had not been approved.21 made in this direction, integration initiatives in low- Similar experiences have been reported in Argen- and middle-income countries generally focus on tina, India, Slovenia, and Türkiye (ADI 2019a). specific diseases or conditions, evidencing the lack Relatively younger countries have also made of a horizontal, system-wide vision that considers the important progress, albeit outside the framework of multiple health conditions that affect older adults and a National Dementia Policy. In Nepal, for example, the need to integrate health and social services (Dru- Alzheimer treatment is provided at no cost for the etz 2019; Mounier-Jack, Mayhew, and Mays 2017). elderly (Williamson 2015). Early detection of cogni- tive decline in adulthood is encouraged in Indone- c. Managing dementia and mental health sia as part of a life-course approach to mental health (WHO 2015). In the Middle East and North Africa, The management and prevention of cognitive and some countries that are less advanced in the demo- neurological degenerative diseases, like Alzheimer’s graphic transition, such as Iraq and Sudan, offer men- disease and other dementias, is a priority in the tal health programs for older people in mental insti- agenda of health systems in most developed coun- tutions or long-term residences, rather than as part tries. However, even in OECD countries, plans and of their national health policies (UN ESCWA 2017). priorities are not always put into practice: there is A large degree of misinformation and cultural still a large degree of underdiagnosis, and the qual- stigma about dementia still prevails in many coun- ity of life among people with dementia is poor, with tries. For example, 66 percent of people worldwide higher-than-recommended prescription levels of think that developing dementia is a natural part of antipsychotic drugs and a lack of adequately trained the aging process, while 63 percent and 67 percent medical personnel (OECD 2018b). of people living with dementia in Asia and Africa, 19   See Alzheimer’s Disease International website at https://www.alz.co.uk/dementia-plans ; 20   See Uruguay Presidencia. Press Release July 22nd, 2020. Retrieved from: https://www.presidencia.gub.uy/comunicacion/comunica- cionnoticias/salud-mental-presupuesto 21   See Alzheimer’s Disease International website at https://www.alz.co.uk/news/national-dementia-plan-confirmed-for-peru 310 CHAPTER 10 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E FIGURE 10.7 Countries that have national dementia plans in place, and percentage of their populations aged 80 and above Source: Alzheimer’s Disease International (2020) (https://www.alz.co.uk/dementia-plans). respectively, were targets of jokes (ADI 2019b). about the disease and defeat main misconceptions. Mechanisms to ensure accurate and timely diag- The aim is that those who attend the training ses- nosis are also needed. Even in OECD countries, sion become a support for people with dementia underdiagnosis reaches almost 50 percent (OECD within their families and communities.23 The Alz- 2018b). This figure is expected to be much higher in heimer Café initiative, first developed in the Neth- lower- and middle-income countries. erlands and then replicated in 15 countries around The development of dementia-friendly com- the world, is another example of a simple and munities 22 in different parts of the world is a pow- easy-to-implement approach to a dementia-friend- erful tool to raise awareness about the issue. Among ly initiative. The aim is to provide a safe social set- the most successful examples of such initiatives is ting where people with dementia and their caregiv- the Dementia Friends movement that comprises ers can meet to share their experiences.24 16 million volunteers in 50 countries, following Among developing countries that are advanced a model first implemented by the UK and Japan. in the demographic transition, the case of China is The model is simple, consisting of face-to-face or worth highlighting as an example of the develop- online training sessions that educate people about ment of dementia-friendly communities: the cam- dementia in order to change the way people think paign “Memory Health in the Community,” where 22   These are defined as “a place or culture in which people with dementia and their careers are empowered, supported and included in so- ciety, understand their rights and recognize their full potential.” See Alzheimer’s Disease International. Dementia Friendly Communities webpage at https://www.alz.co.uk/dementia-friendly-communities/principles 23   See preceding footnote 24   See Alzheimer’s Disease International. Alzheimer’s Cafés webpage a https://www.alz.co.uk/dementia-friendly-communities/alzheimer-cafe COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 10 311 memory specialists, social workers, and commu- by professionals organize training and recreation- nity advocates give lectures to increase awareness al activities for caregivers, is a good initiative that and understanding of the disease and encourage works in 10 cities (ADI 2017). Following the exam- memory screening, is present in 50 cities. Also, the ple of Goa, in India, community-based interventions Yellow Bracelet Project distributes free GPS-guided to support caregivers of people with dementia were bracelets to ensure the safety of people with demen- established in Russia, providing caregivers with ba- tia, preventing them from getting lost. The project sic information about the disease, advice on how was implemented in 233 cities, distributing 500,000 to manage it, and information on the availability of bracelets between 2012 and 2016 (ADI 2017). public services. A similar initiative was implemented Although the most common initiatives in devel- in Peru, a relatively younger country (WHO 2015). oping countries are concentrated on raising aware- ness about the condition, there are examples of pol- d. Adapting human resources icies in other areas, as well. For example, Bermuda and Jamaica established training programs for for- Health care professionals of all types, including mal caregivers of people with dementia (UN ECLAC doctors, nurses, community workers, and other 2017), while Trinidad and Tobago has set up special health care support staff, will need to adapt to meet provisions for people with dementia in its disaster the challenges of an aging population. Health care relief national program (UN ECLAC 2017). The pi- workers need not only to understand which are the lot project “Forget Me Not,” carried out by HelpAge main ailments that affect the elderly, their comor- International and European donors between 2013 to bidity patterns, and their main risk factors, but also 2015 in Bolivia, Colombia, and Peru, is also worth to recognize signs of frailty that could lead to func- mentioning. The program aimed to improve diagno- tional dependency, and to identify warning signals sis and care for people with dementia, as well as pro- of abuse and neglect. They also need to be able to viding support for caregivers and increasing aware- engage older people in health promotion activities ness within the population (HelpAge International and understand how to communicate effectively 2015a). International consensus exists that this is one with them (WHO 2015). of the areas in which countries are further behind Despite the increasing recognition of popula- when it comes to preparing for an aging population. tion aging as a challenge that health care systems Finally, there are few support programs for fam- need to address, the training and adaptation of ily members, who usually bear the main responsibil- human resources is still an area in which progress ity of caring for people with dementia, as discussed has been sluggish. Training is needed at all levels below, in paragraphs 4.16 and 4.17 (Manes 2016). of care, from nurses and caregivers (see Section 4, The Italian Up-Tech program is a successful initia- paragraphs 4.56 to 4.61) to medical doctors. Age- tive that aims at reducing family caregivers’ burden ism, i.e., negative attitudes towards older people through the training and hiring of social workers and towards people because of their age, is not rare who provide continuous support to the caregivers in health care settings. It is observed in the use of a of persons with dementia in the form of face-to- patronizing tone; dismissing health problems and face counseling sessions and follow-up telephone associating them with the supposed natural pro- calls, among others. The social worker acts as a care cess of aging; not involving older people in consul- manager and can coordinate care between different tations or decisions; spending less time in consul- care settings and between hospitals (Amelung et al. tations; and even not performing some diagnostic 2014). Some developing countries have taken steps procedures. Evidence across a variety of settings in this direction, but more needs to be done. has shown how ageism negatively impacts older For example, in Chile, the Ministry of Health people’s quality of life and the quality of care they gives financial support to local governments for the receive (Levy et al. 2020; Ben-Harush et al. 2016). implementation of daycare centers for people with Ageism also raises the question of what it means to dementia, which function as respite services for provide adequate care for older patients. caregivers. However, as per information from 2016, In most developing countries (and in many de- these centers covered only 0.3 percent of the popula- veloped ones), the number of geriatricians is insuf- tion with low or moderate dementia in the country ficient, and medical personnel, in general, lack the (Molina et al. 2020). In China, Dementia Caregiver necessary training to treat older adults. In many cas- Support groups, where volunteers previously trained es, medical curricula do not even include gerontolog- 312 CHAPTER 10 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E ical or geriatric training (WHO 2015). A survey from Financial pressures and the challenge of 1999, comprising 36 countries, showed that, at that keeping costs contained time, 19 percent of medical schools in high-income countries, 43 percent in transitional economies, Developing countries are facing these challenges and 38 percent in middle- and lower-income coun- under restrictive fiscal constraints. Although more tries did not have any training in geriatric medicine than 80 percent of the world’s population lives in (WHO 2015). More recent research done in Latin low- and middle-income countries, they account America and the Caribbean suggests little change for only 20 percent of global health spending (WHO (López and Reyes-Ortiz 2015). Data from Uruguay, 2018b). An analysis of trends for the period 2000- one of the oldest countries in Latin America, shows 2016 shows that this situation might be changing, as that, in 2014, there were only 81 geriatric doctors in total health expenditure has been growing faster in the country (Aguirre 2014). A similar situation can low- and middle-income countries when compared be seen in the Middle East and North Africa, where to their higher-income counterparts (WHO only in Kuwait, Lebanon, Morocco, Oman, Sudan, 2018b). and Tunisia is geriatrics considered a separate spe- Projected demographic changes are expected to cialty in medical school (UN ESCWA 2017). As a exert additional pressures on health care expenditure, result, the ratio of geriatricians is less than one per as older people account for a large share of health care 100,000 older persons in most Arab countries, except utilization, both measured as health expenditure and for Bahrain, Lebanon, and Tunisia (Sibai et al. 2017). as utilization of bed-days (Rechel et al. 2009). More- The progress made in this area has been mea- over, the shift of the burden of morbidity from ger (WHO 2015), although some efforts are worth infectious to chronic diseases pushes costs upwards, as highlighting. In Latin America, as a response to the the latter usually require more expensive treatments workforce shortage, the Pan-American Health Or- and follow-up for extensive periods of time (Guanais ganization, the Inter-American Centre for Social et al. 2018). There is some debate, however, about the Security Studies, and the Latin American Acade- extent of the impact of population aging on health my of Medicine for Older Persons have designed care costs, as some argue that factors such as a specialized course in the management of health technology and proximity to death have a much services for the elderly (UN and HelpAge Inter- larger impact than age per se (Rechel et al. 2009.). national 2012). Aging countries in the Caribbean, The costs of aging in terms of health care re- such as Jamaica and Barbados, have taken steps to sources could be largely contained by promoting a include formal geriatric and gerontological training healthy aging approach. Lifetime health care costs for in schools (UN ECLAC 2017). In Asia, there have obese individuals or smokers are significantly higher been efforts to improve the quality and quantity than for their healthier counterparts (Yang and Hall of human resources in the area of older people’s 2008; Sloan et al. 2004). However, estimates by the health, such as the inclusion of geriatric care in the OECD show that, even under a healthy aging scenario medical curriculum in Bangladesh and Pakistan, that assumes that all gains in life expectancy are the integration of geriatric medicine and services translated into healthy life years, population aging in hospitals in Malaysia, and the establishment of alone will make total health expenditure increase from geriatric training programs in Armenia, Pakistan, 5.5 percent of GDP in 2010 to 6.3 percent of GDP in Sri Lanka, and Thailand (Williamson 2015; UN ES- 2060 in OECD countries, and from 2.4 percent to 3.4 CAP 2017; Yiengprugsawan et al. 2017). In Middle percent of GDP in the so-called BRIICS countries Eastern countries, geriatrics is still considered a rel- (Brazil, Russia, India, Indonesia, China, and South atively new field, although recently it has started to Africa) (de la Maissoneuve and Martins 2013). A be recognized as a specialty of its own, particularly recent study of Latin American countries estimates in rapidly aging countries, such as Kuwait, Lebanon, that health care expenditures attributable to people Morocco, and Tunisia, but also in relatively young- aged 65 or more are expected to increase from 2.2 er countries such as Oman or Sudan (UN ESCWA percent of GDP in 2020 to 4.8 percent in 2050 (Aranco 2017). Despite the efforts made to extend and im- et al. 2022). A closer look shows that in late-dividend prove gerontological and geriatric trainings, mak- and some early-dividend countries that are aging at a ing these specialties an attractive option for young relatively faster speed, for example Türkiye and Israel, professionals remains a challenge (UN ESCWA the impacts of demographic change on health care 2017; UN ESCAP 2017). expenditures are expected to be larger (Figure 10.8). COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 10 313 FIGURE 10.8 Projected increase in health care expenditures due to demographic effects, 2010-2060 (in percent- age points) Source: Adapted from de la Maissoneuve and Martins (2013), Table 3. Note: The non-OECD average and the OECD average include countries not shown in the graph These results are in line with recent IMF projec- 2016). For countries where health care insurance is tions that estimate that health expenditure will in- funded through contributions linked to participation crease in all regions in the world from 2015 to 2100 in the formal labor market, as is the case in almost all due to population aging. Latin America will be the Latin American countries, the fiscal outlook is even region that experiences the largest increase, moving worse as the increase in old-age dependency ratios from a total health expenditure of 4.4 percent of GDP to shrinks the contributory base for funding. 13.5 percent during the period. For emerging Eu- The economic costs of chronic diseases are ropean countries, the increase will be from 5 percent also well documented. Assuming that no efforts are to 11.9 percent, and in emerging Asian countries made to control the rising trends in prevalence, it is health expenditures are projected to rise from 2.3 per- estimated that in low- and middle-income countries cent to 7.1 percent of GDP over the period (Flamini et cumulative losses due to cardiovascular disease, al. 2018). Another study for the Asia Pacific region diabetes, cancer, and chronic respiratory diseases estimates that, from 2015 to 2030, health care expen- could reach up to US$ 7 trillion during the period diture per capita could experience a 9-fold increase in 2011-2025. In developing countries, estimates show Viet Nam, and almost a 5-fold increase in China. The that cardiovascular disease, cancer, chronic respira- factors behind the projected trends are the projected tory disease, diabetes, and mental health disorders increase in the population older than 65, the project- are going to be responsible for a reduction in eco- ed increase in medical costs, and the projected in- nomic output equal to $27.6 trillion in China and crease in the demand for long-term care (Hedrich et al. India from 2012 to 2030 (Bloom et al. 2014). 314 CHAPTER 10 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E INCREASING RISK OF FUNCTIONAL DEPENDENCY AND CHALLENGES FOR LONG-TERM CARE SYSTEMS Highlights • In rapidly aging countries, rates of functional dependency are expected to rise with the expected increase in the share of older people in the population and the prevalence of chronic diseases. • Thus, an increase in expenditures related to long-term care is expected, given the increasing demand for long- term care and the need to provide support to informal caregivers. • The rising demand emerges in a context of reduced availability of traditional informal and family care and in the absence of other affordable, good-quality care options in many countries. • The provision of long-term care services for people with diminished functional abilities needs immediate attention, given changes at the societal and family levels and the current scarcity of affordable and high-quality private services. • The role of the government is crucial. Important gaps still exist in meeting the needs of functionally dependent older adults. Progress towards the development of long-term care services has been slow, and little is known about implementation, especially in low- and middle-income settings. • Key challenges for long-term care services include: (i) achieving accessibility and affordability while ensuring adequate quality, (ii) developing a person-centered approach to care, respecting older people’s dignity and rights; (iii) encouraging - whenever possible - the provision of home care services; (iv) supporting informal caregivers as needed. Investing in the training and professionalization of human resources and the improve- ment of their work conditions is essential to improve the quality of care that older people receive. • Coordination and integration between the social care and health care sectors is central to achieve a continuum of care for the older person and to obtain much-needed efficiency gains. Rising long-term care needs a. More people with functional dependency Increased age is usually associated with physical im- eases of old age, in others, they are simply the result pairment and a decline in sensory and cognitive func- of a “lifelong accumulation of molecular and cellular tions, as well as with a debilitation of the immune sys- damage” (WHO 2015, p. 52). These transformations tem (WHO 2015). While in some cases the changes may have important consequences for a person’s au- are linked to the onset of chronic conditions and dis- tonomy and ability to function independently. BOX 10.3 How is functional dependency measured? The loss of functional capacity is usually measured through difficulties and/or the need of assistance in performing a set of activities of daily living, which are classified as basic activities of daily living (ADL) or instrumental activities of daily living (IADL) (WHO 2015). The former are basic activities that are necessary for survival and self-care, like eating, dressing, using the toilet, or bathing (WHO 2004). The latter are more complex activities that involve some degree of cognitive function and interaction with the external world, such as cooking, shopping, using public transport, managing personal finances, and taking care of one’s own health (WHO 2004). Abilities are lost in the inverse order in which they are gained, so instrumental activities are usually the first to be affected (Katz et al.1983; Dunlop et al. 1997). COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 10 315 Cross-country differences in the share of the with difficulties in basic daily activities could tri- older population reporting difficulties with basic ple in countries in Latin America and the Carib- or instrumental activities of daily living (Box 10.3) bean, rising from 8 million today to 23 million or, can be substantial, with lower- and middle-income its equivalent, 15.7 percent of the population aged countries showing higher rates of functional de- 65 years or more (Aranco, Ibarrarán, and Stampini pendency at old ages than higher income countries, 2022). Similarly, the proportion of elderly with dif- probably due to differences in subjacent health sta- ficulties in basic activities of daily life is projected tus (WHO 2015). Results from the Study on glob- to increase by 169 percent in Thailand from 2014 to al AGEing and Adult Health (SAGE) show that, in 2050; when future increases in education levels are China, around 35 percent of people 75 years old or considered, the figure is estimated at 127 percent more need some type of assistance for at least one (Loichinger and Pothisiri 2018). Increases are also basic activity of daily living; this figure reaches 45 projected in other developing regions, such as Eu- percent in South Africa, more than 60 percent in rope and Central Asia (World Bank 2015). Mexico, and around 80 percent in India and Russia Results could be even more problematic if the (WHO 2015). In Thailand, almost 25 percent of peo- prevalence of chronic diseases increases, in which ple older than 80 need help with basic activities of case increasing rates of functional dependency are daily living, and the figure reaches 70 percent in the expected to be seen even when comparing the same case of instrumental activities (Knodel et al. 2015). age groups over time. Some evidence already points In contrast, fewer than 20 percent of people aged 75 in this direction. Data from Mexico shows that the or more have difficulties with basic activities in Den- proportion of people 80 years old or more with dif- mark, the Netherlands, or Switzerland (WHO 2015). ficulties in performing basic activities of daily life The series of Longitudinal Surveys of Social increased from 30 percent in 2001 to 44 percent Protection, carried out in various countries in Lat- in 2015; in the case of difficulties with instrumen- in America, show that the prevalence of difficulties tal activities, the share rose from 31 percent to 39 in performing basic activities of daily living among percent during the same period (López-Ortega and people 80 years old or more reaches 30 percent in Aranco 2019). Data for China also show an increase Chile (2015) and around 12 percent in El Salva- in the rates of disability for the “younger old” (peo- dor (2014), Paraguay (2015), and Uruguay (2013), ple from 60 to 74 years old) between 1987 and 2006, while the Costa Rican Longevity and Healthy Aging although the trend is reversed for individuals 75 Study (2009) shows a prevalence of more than 40 years old or more (Liu et al. 2009). percent. Figures for instrumental activities of daily Long-term care services are services provided living, when available, are much larger, and reach with the objective of assisting functionally depen- 23 percent in Uruguay and 60 percent in Costa Rica dent individuals with the performance of their daily (Aranco et al. 2018). Data based on the Mexican activities. Services offered can be varied, and their Health and Aging Study (2015) show somewhat suitability usually depends on the level of depen- lower degrees of dependency both on basic and in- dency of the beneficiary. Services that are often part strumental activities than the SAGE survey, both of the menu include: long-term residential care, around 40 percent for individuals aged 80 or more home care, daycare centers, tele-assistance services, (Aranco et al. 2018). and support to informal caregivers (Table 10.3). In rapidly aging countries, rates of function- It is argued that, unlike the case of health care al dependency are expected to increase following services, for which increases in projected demand the rise in the share of older people in the popu- due to aging are more uncertain, the increase of lation. It is estimated that by 2050, due only to de- long-term care expenditures is certain to happen as mographic pressures (i.e., maintaining functional the share of older individuals in the population in- dependency rates constant), the number of persons creases (Rechel et al. 2009). 316 CHAPTER 10 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E Table 10.3 Long-term care services by type Residential care Services provided in institutions and that include housing, meals, help with daily activities, and basic health care services. These services usually focus on individuals with severe degrees of dependency. Home care services Services provided in the house of the dependent person. Usually includes a personal assistant who helps with daily activities, but can also include a wider array of services, such as meal delivery or help with household chores. These services are suitable for individuals with all levels of dependency. Day centers Services offered at facilities for a number of hours a day but not including accommodation. These usually focus on preventive activities and promoting healthy aging, so their target population is auton- omous individuals or people with only mild levels of dependency. Tele-assistance services Information technology-based services that monitor and assist the dependent person remotely (emer- gency hotlines, fall detectors, and others.). They usually focus on autonomous individuals or individuals with mild levels of dependency. Services for caregivers Support services for unpaid caregivers, which may include training, counseling, respite services, and special work leave. Source: Adapted from Cafagna et al. (2019) b. Reduced availability of informal care networks Traditionally, the role of caregiver has been main- people should lie mainly with the family instead of ly a responsibility for the family, especially women, with the market or the state (Batthyány, Genta, who receive no payment for these tasks. Cultural and Perrota 2012). This may partly explain why reasons explain the overarching role of the family, older people from lower socioeconomic strata are particularly women, when it comes to caregiving more prone to live in extended family households. responsibilities. Traditional values regarding filial re- Yet changes in female labor participation and sponsibilities towards older people are strongly em- family structures are shrinking the network of avail- bedded in many cultures, sometimes hindering the able family caregivers and posing additional development of other public and private solutions to challenges for countries when it comes to cater to the needs of older people. Children’s obliga- developing sustainable long-term care strategies. tion to take care of their parents is even regulated by Female labor participation rates have increased law in several countries in Asia, including Cambodia, markedly in Latin American and the Caribbean; India, Singapore, Sri Lanka, and Viet Nam (HelpAge increases have also been seen Sub-Saharan Africa International 2015b; UN ESCAP 2017). Some East- and South-Eastern Asia and the Pacific (ILO 2016). ern European countries (for example, Serbia and Consequently, women have less time to dedicate to Latvia) also stipulate through legislation the non-remunerated activities, such as caring for older obligation of children to take care of their parents members of the family. And when women do (Stokic and Bajec 2018; Rajevska 2018). Care for provide such care, the opportunity costs are higher. older adults is also seen mainly as the responsibility Declining fertility rates and migration flows of the family in Middle Eastern countries (Hussein have reduced the typical family size. As a result, a and Ismail 2017). greater number of older people are living This is also the case in many countries in Latin alone, with fewer close relatives to rely on. In America. In Uruguay, for instance, almost half of Latin America and the Caribbean, on average, the population younger than 70 agrees that older the proportion of older people living alone people should be taken care of exclusively by went from 9 percent circa 1990 to 13 percent family mem-bers, while around 30 percent think circa 2010 (UN 2017c). In Uruguay, for that the best option is collaboration between a instance, 26 percent of people older than 65 live family member and a formal personal assistant. alone, and more than 30 percent live only with Only around 20 percent consider formal care their spouse. These figures are also high among (either through a paid assistant or through the functionally dependent population, 24 institutionalization) to be the best option percent of whom live alone, and another 24 (Batthyány, Genta, and Perrota 2012). Individuals percent only with their spouse (Monteiro and of lower socioeconomic status are more likely to Paredes 2016). Similar patterns are seen in other believe that the responsibility of taking care of older Latin American countries, with the percentage of COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 10 317 people living alone being high even for the “oldest A study conducted in China, Mexico, Nigeria, old.” Evidence from other regions also points to a re- and Peru (Mayston et al. 2014) shows that in these duction in family size (Bongaarts and Zimmer 2002; countries most of the help is provided by either United Nations 2017). In Asia, the proportion of old- the recipient’s spouse or a child and that—except er people living alone increased from 18 percent in in rural China, where care responsibilities seem 1990 to 27 percent in 2010, and the proportion living to be shared equally between genders—caretakers with children decreased from 73 percent to 64 percent are mostly women. The study also shows that only in the same period (UN 2017c). In Thailand, for in- in the urban regions of China and Peru are care stance, the percentage of older people living with chil- responsibilities shared with paid workers. In Uru- dren declined from 77 percent in 1986 to 55 percent guay, data from 2013 show that among older adults in 2014 (Knodel et al. 2015). Similar trends were seen who receive help with their daily activities, almost in Eastern Europe and in Middle Eastern countries. 76 percent rely on unpaid family members to pro- In general, the probability of living alone or vide such care; in most cases the help comes from a living only with a partner is higher the higher the woman (Aranco and Sorio 2019). In Chile, the per- person’s socioeconomic status, as people of higher centage of functionally dependent older adults who socioeconomic status tend to have fewer children, receive family help is 70 percent, exclusively (46.4 they are usually in better health (and hence, less de- percent); with estimates showing that between 70 pendent on help), and they can more easily afford to percent and 90 percent of this unpaid help is pro- hire good-quality private care in case of need. vided by women (Molina et al. 2020). In Colom- At a time when the demand for care is rising in bia, the 2015 SABE survey shows that 84 percent of developing countries, the reduction in the availabil- those providing care to the elderly are women, and ity of informal care brought about by these changes only 7 percent receive some kind of payment for the leads to a large share of elderly people with func- task, either in cash or in kind (UN 2017d). Similar- tional limitations not receiving any help whatsoev- ly, in Thailand, 90 percent of older people receiving er. In fact, it is estimated that, in Thailand, among assistance with their daily activities report getting the population aged 60 years or more, 36 percent of it from a family member, daughters being the main those needing assistance do not receive any (Knodel caregivers, followed by spouses (Knodel et al. 2015). et al. 2015), while in Uruguay this percentage reach- Globally, unpaid care work is estimated at ap- es almost half of those requiring help with either proximately 3 percent of economic output; most un- basic or instrumental daily activities (Aranco and paid caregivers are family members, usually women; Sorio 2019). Evidence for China shows that only and a large proportion of this care work is associated around 50 percent of older individuals requiring with non-communicable disease-related care (The help in using the toilet have someone to assist them Lancet Commission on Women 2015).25 In some (Glinskaya and Feng 2018). countries the contribution of unpaid work to GDP is even larger. For example, in many Latin American c. Women are still the main unpaid countries it has been estimated to reach 20 percent, caregivers for older people with women responsible for 70 percent of this con- tribution (ONU Mujeres 2018; Vaca 2021). In the absence of public solutions for long-term care The consequences that care responsibilities in many aging countries, families, and particularly have for family members’ (very often women’s) women, are still the main unpaid caregivers for older health, wellbeing, and labor market participation dependent people. This situation forces many caregiv- are well documented. More than 48 percent of ers—again, principally women—to either forego par- caregivers in rural China were forced to give up ticipation in the labor market altogether or to endure or reduce paid working time to fulfill their care- excessive working hours, when combining paid and giver responsibilities; this number reaches almost unpaid work. This is true in most countries, regard- 40 percent in Nigeria and rural Mexico (Mayston less of their level of development, but it is even more et al. 2014). In Poland, around 66 percent of care- acute in developing countries, where both public and givers do not participate in the labor market, while private alternatives to long-term care are limited. in Slovakia 28 percent of inactive women declare 25   https://www.thelancet.com/commissions/women-health-2015 318 CHAPTER 10 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E caregiving responsibilities (either towards children excluded from coverage due to restrictive means-test- or towards older individuals) as one of the main ing accessibility rules (Scheil-Adlung 2015). reasons behind not seeking a job (Sowa-Kofta 2018; In developing countries, the lack of state-orga- Gerbery and Bednárik 2018). A study that includes nized services, coupled with barriers to accessibility data from Chile, Colombia, Costa Rica, and Mexi- and affordability of services, leads to extremely low co shows that women taking care of parents are less levels of coverage, issues of poor quality, and long likely to participate in the labor market and that, waiting lists (Glinskaya et al. 2024). A recent review when they do, they work for a smaller number of of the state of the art in 26 Latin American and Ca- hours (Stampini et al. 2020; cf. Gatti et al. 2024). ribbean countries shows that the region performs The burden of caregiving responsibilities to- poorly when it comes to coverage and quality of ser- ward family members also takes its toll in terms of vices. The countries that perform better in terms of the emotional and physical health of the caregivers. coverage are Argentina and Costa Rica, but even in A multi-country survey conducted in 11 countries in these countries only 20 percent of dependent people Latin America, India, and China highlights the emo- older than 65 receive some public long-term care ser- tional strains suffered by caregivers of people with vices (Aranco et al. 2022). Small and relatively richer dementia (Prince et al. 2012). Research from Chile countries like Barbados achieve just 15 percent cov- (Flores, Rivas and Seguel 2012; Villalobos 2018), Co- erage, while Uruguay has 11 percent and Chile only 7 lombia (Giraldo and Franco 2006), and Philippines percent. Most of the other countries studied have an (Varona et al. 2007), among others, also confirms this. estimated coverage of less than 5 percent. Regarding quality, the report shows that, despite most countries Country responses in long-term care having quality standards for residential care in place, the level of compliance is very low, and challenges a. General context are even starker in the case of home care (Aranco et al. 2022). Results also show that the bulk of human As a result of the increasing demand for long-term resources working in the field do not receive proper care services and the shrinking network of family training for the tasks they are expected to perform. caregivers, there is a gap between the needs for and In countries where no formal long-term care the supply of care. This gap will widen as populations system is in place, like Serbia, the problem is particu- age if no actions are taken (Glinskaya et al. 2024). larly acute: waiting lists for admission to institutional Given the barriers to developing a private long-term and daycare centers are long and the level of coverage care insurance market, and the prohibitive costs of fi- very low. In 2016, only 8.5 percent of older individ- nancing older-adult care out-of-pocket (Barr 2010), uals in Serbia received some kind of long-term care public intervention is critical to ensure that individ- service, in the form of residential care, home and day uals have the support they need to continue living as care, or cash allowances (Stokic and Bajec 2018). In autonomously as possible as they grow old (Caruso, Hungary the figure reaches 10.2 percent (Gal 2018). Galiani and Ibarrarán 2017; Glinskaya et al. 2024). High levels of out-of-pocket payments could also be The development of national long-term care sys- a barrier to effective coverage: for instance, such pay- tems is relatively new on the political agenda global- ments represent around 50 percent of relevant costs ly. While some countries (particularly high-income in the Czech Republic (Maly 2018), while in Latvia, OECD countries) have well-developed national sys- 90 percent of the older person’s pension is used to tems in place, in low- and middle-income countries cover the cost of institutional care (Rajevska 2018). the provision of long-term care services is – at best Efforts should be made to design and imple- – characterized by a multiplicity of small-scale inter- ment long-term care systems that are affordable, ventions with limited coordination and integration. accessible, ensure minimum quality, and are per- Even in countries where public solutions do exist, son-centered, respecting older people’s dignity and coverage is low and institutional factors often affect rights and, whenever possible, improving their accessibility. According to a study done for 46 devel- functional abilities (WHO 2015). In what follows, oped and developing countries in 2015, 48 percent of we provide examples of some actions taken in that older adults were not legally covered for their long- direction by developing countries. term care needs, and an additional 46 percent were COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 10 319 b. Development of national long-term care population – i.e., the oldest and the poorest – could policies in developing countries be a good strategy when, due to budget limitations, universal coverage is not possible (WHO 2015). Yet, Many countries have started to raise the issue of these restrictions could leave an important share of long-term care on the political agenda, either with the dependent population uncovered (Villalobos the aim of addressing the problems posed by cur- 2018). In Chile and Costa Rica, for example, there rent systems – when they exist – and/or designing are systems in place, but their geographical coverage new solutions to ensure that older people receive the is limited. The Chilean system Chile Cuida, which care they need. In 2021, Costa Rica approved a 10- started in 2017, covered only 22 of the 360 counties year policy program to implement a care system at in the country by 2020. The system targets the most the national level. Colombia is in the process of de- socioeconomically vulnerable population aged 18 signing a national long-term care system, as is Saint years or more with moderate to severe levels of de- Maarten (UN ECLAC 2017), while relatively young- pendency. One of its main objectives is to coordinate er countries such as Paraguay and Ecuador have re- the fragmented and uncoordinated public supply of cently included the subject on their public agenda. long-term care services in Chile. The case of Costa The process of designing and implementing a Rica is similar: the country’s long-term care program long-term care system is not easy and one size does (Red de Atención Progresiva para el Cuido Integral de not fit all (Glinskaya et al. 2024). Countries need to las Personas Adultas Mayores) is implemented at the consider at least four factors when planning such a local level, in counties/communities that express in- system (Medellin, Ibarrarán, and Stampini 2018): terest and that meet certain prerequisites in terms (1) Who are the beneficiaries? (2) Which benefits of financial resources and management capacities. will the system cover, and how will they be deliv- The program covers functionally dependent people ered? (3) Which mechanisms should be used to en- over 65 years old from low socioeconomic settings sure the quality of services? and (4) How will the who have no other support networks available. The system be financed? benefits are provided in the form of cash subsidies In Latin America and the Caribbean, the case to purchase services/goods that range from home of Uruguay is worth highlighting. As one of the care help to home adaptations, transportation, food, countries with the oldest populations in the region, and medicines, among others. Uruguay has also been the first to design and imple- Progress has been slower in developing coun- ment an integrated national long-term care system, tries in Asia, despite their advanced stage in the de- the Sistema Nacional Integrado de Cuidados. This mographic transition. The heterogeneity of both the is a tax-funded system, which came into operation stage of economic development and the stage of the at the end of 2015. The system covers children and demographic transition across Asian countries ex- persons with functional dependency due to aging plains, in part, the different state of long-term care or disability. Due to budget constraints, the system policies in the region (Esquivel and Kaufmann 2017). focuses on the most vulnerable populations, with While high-income countries such as Japan, Korea, the aim to progressively increase its coverage over and Singapore have integrated and coordinated long- time (cf. Glinskaya et al. 2024). Those aged 80 or term care policies, the situation is different in low- more (or younger than 30) with severe dependency and middle-income countries. Although in some are entitled to 80 hours a month of home care; those countries governments are stepping in to close the 65 and older with mild levels of dependency can go gap between the growing demand for long-term care to publicly funded day centers and have the right to services and the declining support from informal access a basic teleassistance service after they turn caregivers, the implementation of national long-term 70 years old. In the case of homecare services and care policies is still at early stages. In China, some telecare, the percentage of the cost financed by the local governments (for example, the cities of Qing- government depends on household income, with dao, Wuxi, and Suzhou) have launched experimental individuals of higher socioeconomic status being long-term care insurance schemes that cover home, subjected to higher copayments. institutional, and hospital care, but a national coor- Given that functional dependency is usual- dinated policy is still not in place (UN ESCAP 2017). ly concentrated at older ages and that it correlates The situation is highly heterogenous in Eastern negatively with socioeconomic status (Aranco et al. European countries. Bulgaria approved a national 2022), targeting the most vulnerable sectors of the strategy for long-term care in 2018 (Bogdanov and 320 CHAPTER 10 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E Georgieva 2018), while Slovakia is currently work- c. Strengthening home care and ing on the design and approval of an integrated and community services coordinated national system (Gerbery and Bednár- ik 2018); steps to integrate social and health care All over the world, in recent years, there has been services, and to transform long-term care services a tendency towards deinstitutionalization of care, into an independent policy field, are being taken in while home and community care services are in- Hungary (Gal 2018). Some countries have taken ac- creasing in importance (Aranco and Ibarraran tions to improve long-term care service coverage for 2020). In addition to being aligned with older peo- populations living in rural areas. These include Rus- ple’s preferences, “aging in place,” when possible, sia, with the construction of new nursing homes, has advantages both in terms of economic costs for and Poland, with its Care 75+ program, which pro- governments and of older adults’ mental and physi- vides partial financial support to local authorities cal health (WHO 2015; Boland et al. 2017). for the provision of services (Sowa-Kofta 2018). Several countries have undertaken initiatives Despite the existence of stand-alone initiatives to improve the care of the elderly, with or without in countries such as Ghana or Tanzania, the only the framework of a national and integrated system. three countries that have developed national efforts In many OECD countries, the emphasis has been to implement national long-term care policies in on policies that encourage an aging-in-place ap- Sub-Saharan Africa are Mauritius, Seychelles, and proach, promoting home and community services South Africa (WHO 2017c). that allow older people to live independently as Efforts to design and implement long-term care long as possible (UN ECE 2017; Spasova et al. 2018; systems reflect countries’ commitment to bridge the Colombo et al. 2011). The 2015 long-term care sys- gap between increasing demand and shrinking sup- tem reform in the Netherlands, for example, lim- ply of long-term care. However, in most cases more ited the option of institutional care to those cases needs to be done. Financing is, most of the time, where home care is not feasible; similarly, reforms one of the main constraints for further progress in Germany (2015-2017) and in Japan (in the late (Glinskaya et al. 2024). Budgets devoted to long- 1990s) favored the utilization of home care services term care services are often inadequate, resulting in by making them more attractive than institutional low levels of coverage and limited services. services (Aranco and Ibarraran 2020). In Uruguay, for example, the cost of the system Following these global trends, developing in 2017 amounted to 0.04 percent of GDP, a figure countries have taken steps to strengthen the im- that was estimated to be five times lower than what portance of home-based long-term care services. would be necessary to achieve the explicit objective In Latin America and the Caribbean, for instance, of the system of reaching 60 percent coverage by home care services are the backbone of the newly 2020 (SNIC 2018; Matus-Lopez 2017). In fact, data created systems in Uruguay and Chile, and in Ar- from 2018 show that the system covers only 6 per- gentina the government subsidizes the hiring of a cent of the country’s dependent elderly (SNIC 2019). personal assistant for older adults who need help in Similarly, initial projections in Chile were for the sys- their daily activities. While in Argentina and Chile tem to cover 70 municipalities in 2018, 181 in 2019, the benefit is targeted at the economically vulner- 263 in 2020, and 345 in 2021. The lack of financial re- able population, in Uruguay it covers all severely sources and political commitment have made it im- dependent older individuals, but cost-sharing is possible to reach these aims, and in 2020 the system means-tested (Aranco and Ibarraran 2020). was in place in only 22 municipalities (Barraza 2017). Development of home care programs and ser- An alternative to the lack of resources is to fund vices has also been seen in many Caribbean coun- the system through compulsory insurance, as is the tries: in Antigua and Barbuda and Barbados, the case in some developed countries, such as France, public sector runs a free home care program that Germany, Japan, Korea, or the Netherlands (Aranco assists older people with basic daily activities, while and Ibarrarán 2020). Among developing countries, in Trinidad and Tobago means-tested home care this source of financing has not taken hold, despite services are provided; a home health care service for some recent discussions in the context of the pen- those who have been released from hospital has re- sion reform in Chile (Macías Muñoz 2019). cently been launched in Guyana (UN ECLAC 2017). COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 10 321 Other Caribbean countries that have taken steps daily activities; home care can include a wide array in this direction are Anguilla, Bermuda, Cayman of other services that, even if they do not involve the Islands, Dominica, Jamaica, Saint Lucia, and Saint assistance of a professional caregiver, can help older Vincent and the Grenadines. In Bermuda, the gov- people to remain in their homes as long as possible ernment not only provides a home care subsidy but (Aranco and Ibarraran 2020). Countries such as Es- also offers services that help older adults to remain in tonia, Israel, Italy, or Ukraine help with errands and their home for as long as possible, such as meal de- daily chores (UN ECE 2017). In Latvia, where mu- liveries and handyman services (UN ECLAC 2017). nicipalities have the obligation to provide home care Within Middle Eastern economies, home care in cases where the dependent person does not have programs are available in Jordan, Kuwait, Oman, any other family support, home services include not and the West Bank and Gaza, while in Lebanon only assistance services but also laundry, meal de- such services are provided mainly by the private livery, and teleassistance (European Commission sector (Hussein and Ismail 2017; UN-ESCWA 2019). In Chile, local governments have the freedom 2017). In Kuwait, services are provided for free to decide which services to include as part of the sys- (Hussein and Ismail 2017). tem’s benefits through the so-called Specialized Ser- The development of home care services seems vices (Servicios Especializados), depending on the to be somewhat more advanced across Eastern Euro- specific needs of the local older-adult population. pean countries, regardless of the existence of a com- Other countries, such as Myanmar and Sri Lanka, prehensive national system. In 2013, Lithuania im- provide loans for home adaptations and encourage plemented a home care program (Integral Assistance the construction of dwelling units for parents adja- Development Programme) that offers nursing and so- cent to the family’s main house (Williamson 2015). cial home services for dependent people and advice to Other programs, such as home delivery of meals for family caregivers (UN ECE 2017). A mobile team of older people, are available in Lebanon and the West specialists (social workers, psychologists, and others) Bank and Gaza (Hussein and Ismail 2017). that provide at-home services to the older population Alternative solutions to support older people has been developed as a solution to cover populations in remaining in a home-like context have been de- in rural areas in Latvia (Rajevska 2018); because of veloped for cases in which staying at home is not this and other efforts toward de-institutionalization, possible; in Belarus, Estonia, and Russia, for in- expenditure for home-based long-term cares services stance, foster families that provide accommodation increased 27 percent between 2014 and 2016 (Raje- for older persons in exchange for a monthly state vska 2018). Bulgaria has also taken action to discour- payment are common (UN ECE 2017; European age institutionalization and promote community and Commission 2019). Foster care for older people is family-based home services by expanding the range also a standard practice in Croatia and has been of services provided by its public system, while Slova- encouraged since the approval of the country’s Fos- kia approved a strategy for the de-institutionalization ter Families Act in 2012; however, the number of of social services (including long-term care services) adults in foster care increased only marginally from in 2011 (European Commission 2019). 2011 to 2015, rising from 1,409 to 1,581 over the In Sub-Saharan Africa, the case of Seychelles period (Stubbs and Zrinščak 2018; European Com- is worth noting: a home care scheme has been in mission 2019). In Tunisia, the government pays a place in the country since 1987, where the older modest sum to families that are willing to host older person can choose the caregiver, including a family persons in their homes through a Surrogate Family member, but no standards for quality are set (i.e., program (UN ESCWA 2017). the caregiver does not receive any kind of training). The role of community care has been particu- In addition, the fact that the program is funded by larly important for the promotion of home care in government revenues alone puts its fiscal sustain- Asian countries (UN ESCAP 2017). Within the ability in jeopardy (WHO 2017c). Association of Southeast Asian Nations (ASEAN) It is important to highlight that the range of countries,26 volunteer-based home care programs for benefits included as home care services is not lim- the elderly have proliferated as a way to help older ited to the help offered by a personal assistant with people remain in their homes for as long as possible. 26 ASEAN countries include Brunei, Cambodia, Indonesia, Laos, Malaysia, Myanmar, Philippines, Singapore, Thailand, and Viet Nam. 322 CHAPTER 10 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E Through a joint initiative, HelpAge International, must pay their living expenses out of pocket. By ASEAN, and the Republic of Korea have collaborated 2014, the initiative covered around 12 percent of with NGOs, government ministries, community villages in the country (Glinskaya and Feng 2018). volunteers, families, and older people to implement a Community-based care is also being encour- volunteer-based home care model across ASEAN aged by the government and international organi- countries. The model is based on the Korean zations in Sri Lanka (UN ESCAP 2017). Some com- experience and adapted to the local context of each munity initiatives encourage integration between country (HelpAge International 2014b). The health and social care, as in East-Timor, Singapore, implementing NGOs are responsible for selecting and and Thailand (HelpAge International 2013). C om- managing the volunteers. Volunteers receive basic munity care has also been gaining momentum in training before they start working, and their tasks Europe. In the Netherlands, nurses are the main range from providing companionship and support to pillars of a model that has proved successful for help with daily activities; in countries with weak many years and is now well known worldwide. The health care systems (Myanmar, Laos, and the so-called Buurtzorg approach (meaning “neighbor- Philippines) the provision of basic health services can hood care”), run by a not-for-profit o rganization, also be arranged (HelpAge International 2014b). The relies on self-managed teams of (no more than 12) criteria for selecting participants vary across countries, highly qualified nurses that are responsible for be- but usually services are targeted to dependent older tween 40 and 60 patients in each geographic area. people from low-income settings. Volunteers and Working closely with general practitioners and beneficiaries are matched, considering not only the other health care professionals, the nurses provide needs of the older person but also the skills and comprehensive and personalized care to patients in preferences of the unpaid volunteers. Aims of the their own homes, with services that can range from project include fostering an institutional status for administering medications to help with basic activi- home care ser-vices and encouraging the ties or preparing a simple meal (Kreitzer et al. 2015). development of guidelines and policies in the Less-developed countries in Eastern Europe countries where the services are implemented. This have followed the same path. For example, in Bul- was accomplished in Cambodia, Indonesia, Malaysia, garia, an Action Plan for the implementation of Myanmar, the Philippines, and Viet Nam. Policies a national long-term care system over the period and guidelines already existed in Singapore and 2018-2021 anticipated the completion of 100 new Thailand (HelpAge International 2014b). community-based social services (European Com- Community care services are also important mission 2019). Community day and home-based outside ASEAN countries. In China, for example, the care is also a distinctive feature of the provision development of so-called “time banks” in some of long-term care services in Serbia, while in Lith- provinces is an example of an innovative way of uania community services are expected to take a organizing community-based care. Under this scheme, leading role by 2030 under the country’s “Guide- time is used as a currency: relatively younger persons lines for the Deinstitutionalization of Social Care care for their older peers and receive, in exchange, a Homes for the Disabled, Children Deprived of Pa- time credit that allows them to use an equivalent rental Care, and Adult Disabled Persons” (Lazut- number of hours of care when they need it (UN ES- ka, Poviliunas, and Zalimiene 2018). CAP 2017). Another popular model of home-based As in the case of health care, faith-based organiza- care in the country is the Virtual Elder Care Home, tions also play a role in the delivery of long-term care, where local governments maintain an information particularly when it comes to residential and nursing hotline and service center, which acts as a referral for services. Information regarding the importance of older people in need of services. Providers are these institutions in terms of coverage and services reimbursed by the government for services provided provided is, however, limited, and more research is to vulnerable dependent people (Feng et al. 2012). The needed in order to quantify their contribution, Happiness Homes are a community-based solution for The establishment of day centers is another way older people living in rural areas. They consist of to enable older adults to stay in their homes for as village homes in which older adults live together and long as possible, although usually this type of service help each other. Housing and utilities are provided free targets individuals with lower levels of dependency by the government, as are living expenses for those (Cafagna et al. 2019). In many countries in Eastern who have no source of income; those who can do so Europe, such centers are usually provided and fund- COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 10 323 ed by local governments, as is the case in Bulgaria, include control and inspection-based mechanisms, Poland, and Serbia (Slavina et al. 2018), and with the the disclosure of information by public agencies to aid of not-for-profit organizations (UN ECE 2017). encourage competition, and the professionalization Another example of day centers developed at of the long-term care workforce (Mor, Leone, and the local level is seen in Brazil. Under the program Maresso 2014). Irrespective of the approach ad- “Brasil, Amigo do Idosso” (Brazil, Friend of the El- opted, the role of the government in setting quality derly), which encourages local governments to car- standards and enforcing their compliance is crucial. ry out elder-friendly initiatives, the state of São Pau- In developing countries this might be a critical lo has expanded the number of day and community challenge. In fact, the evidence shows that in many care centers for older adults as a way to promote low- and middle-income countries, private care ser- their levels of autonomy and social participation.27 vices are often not regulated, compromising quality In Uruguay, nine day-centers were built under the standards, and allowing situations of abuse to take new long-term care system for adults who are mod- place (WHO 2015). Even when private or public erately dependent (SNIC 2020). In Chile, the Min- solutions do exist, the quality of services is often istry of Health runs day centers dedicated to people poor and unregulated. This is true for all regions. that have been diagnosed with dementia. The 2017 Madrid International Plan of Action on Despite this progress, home and community Ageing progress report for the Middle East region, services are still under-developed in many countries, for example, highlighted that, in most countries in hindering the process of deinstitutionalization. More the region, the lack of a legal framework to regu- than 50 percent of public expenditure on long-term late and control these institutions undermined their care in Estonia, for example, is devoted to institu- quality (UN ESCWA 2017). Lack of monitoring and tional care (Paat-Ahi and Masso 2018); evidence for regulation can lead to situations of abuse, especially Latvia shows that, despite efforts toward deinstitu- in institutional settings. Recent global reviews doc- tionalization, demand and funding for institutional ument that 64 percent of the staff working in long- care has remained stable (Rajevska 2018). Romania term care institutions have admitted to engaging in has witnessed an important increase in both public abusive practices, while the figure is 15.7 percent in and private residential homes since 2006 (Pop 2018). community settings (Yon et al. 2017; Yon et al. 2019). Similarly, in Poland, where the number of daycare Within this context, efforts to improve the quali- centers has increased recently as part of the govern- ty of services have been undertaken, whether through ment initiative Senior+, lack of funding has prevent- the inspection and control of facilities and providers, ed further progress and challenges the sustainability the training of caregivers, or both. However, these of the program (Sowa-Kofta 2018). agendas are particularly challenging. Regulations While encouraging aging in place and commu- regarding quality standards, such as infrastructure nity-led alternatives may be an interesting policy and personnel requirements in institutional facili- option, it still requires strong government steward- ties, are more prevalent across countries, as – by law ship and regulatory oversight to ensure that quality – most require institutions to meet certain criteria standards are met (Glinskaya et al. 2024). Investing to operate. In some countries, these regulations are in the proper training of human resources, as high- relatively new. For example, quality standards for res- lighted in the following paragraphs, is also of para- idential and daycare facilities in Lithuania were first mount importance to guarantee a person-centered approved in 2006 (Romas, Poviliunas, and Zalimiene approach to care in these settings and ensure that 2018) and in Romania in 2012 (Pop 2018). In Trini- older people are protected from abuse and neglect. dad and Tobago, actions are being taken to strength- en the control and inspection of care facilities, and d. Improving the quality of services and Uruguay has also set new standards for institutional investing in human resources residences and reinforced inspection mechanisms (UN ECLAC 2017; Aranco and Sorio 2017). Developed countries use several mechanisms to Nevertheless, the gap between what is stated guarantee the quality of services provided. These in the legislation and actual practice is sometimes 27   See Governo do São Paulo. Press Release: “Conheça as ações do Programa São Paulo Amigo do Idoso” 22nd December, 2015. Retrieved from: https://www.saopaulo.sp.gov.br/ultimas-noticias/conheca-as-acoes-do-programa-sao-paulo-amigo-do-idoso/ 324 CHAPTER 10 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E large. In Uruguay, for example, data from 2019 show developed ones, the sector attracts many migrant that fewer than 2.7 percent of long-term care insti- workers, who must often accept lower wages and tutions have the social and sanitary authorization worse working conditions given the lack of alter- required by law to operate.28 Lack of supervision is native opportunities (Addati et al. 2018). Paid care a widespread problem across developing countries workers may also suffer from emotional stress that in all regions. A review of the country reports in the could affect their mental and physical wellbeing. ESPN Thematic Report on Challenges in Long-term For those providing home care, the isolated nature Care from 2018 highlighted this as a common prob- of the job and the existence of conflicts with family lem in most Eastern European countries. members due to privacy issues or delimitation of Closely related to the quality of services is the responsibilities can make the situation even harder. need to ensure an adequate workforce, both in Poor working conditions make it challenging to terms of quantity and quality. Nearly a decade ago, recruit and retain workers in the sector, compro- it was already estimated that the shortfall of formal mising the viability of services (Addati et al. 2018). caregivers had already reached about 13.6 million Investing in the training of human resources globally (Scheil-Adlung 2015). Recent estimates in the care sector not only improves the quality for Latin America and the Caribbean show that the of care, but also gives recognition and value to need for caregivers for older adults in that region the caregiving task, opening opportunities for alone could reach 14 million by 2050 (Villalobos et career paths, increasing salaries, and hence al. 2022). Improving pay and working conditions attracting more people to the sector, creating a could make the sector more attractive for poten- virtuous cycle. Given the over-representation of tial workers (WHO 2015). In developing coun- women in this sector, with adequate policies, tries, where levels of labor informality are high, this population aging can increase women’s oppor- represents an important challenge (Cafagna et al. tunities in the formal labor market in low- and 2019). Also, it is important to ensure that caregiv- middle-income countries, and equip women with ers have the necessary skills by requiring that they knowledge, skills, professional careers, and busi- engage in proper education and training courses. In ness opportunities. most countries, this would imply updating the ed- Some countries have started to invest in the ucational curriculum and even the development of training and certification of formal caregivers, although specialized educational institutions (WHO 2015). there is still much to be done. Even in OECD countries, Given the intrinsic characteristics of the task educational requirements for long-term care workers at hand, caregivers need not only to have technical are low, and only Canada, Denmark, Germany, and skills (how to move or bathe a bedridden person, Korea have a career structure for formal caregivers for example), but they also need to be trained in (OECD 2020). In developing countries, some examples a wide array of softer skills such as patience and are worth noting. Uruguay has made important efforts interpersonal communication, particularly in the in this regard, as training and accreditation of formal case of patients with cognitive decline. They also caregivers have become compulsory since 2015 (SNIC need to be able to identify when the dependent per- 2015). Formal training is also required in Chile (UN and son might need other services, such as health care HelpAge International 2012; Molina et al. 2020). In or additional social care services. Asia, aging countries such as Thailand and China are Even in cases in which care is paid, it is often among the countries that require a formal certification provided by workers with no formal training or course for those working as caregivers (HelpAge expertise (WHO 2015). Women are more likely International 2015b), while other countries such as Fiji, to work in the care sector, and female occupations Sri Lanka, and Viet Nam encourage voluntary training tend to pay less well than typical male occupations. for community caregivers (Williamson 2015). Paid care jobs are characterized by low wages, infor- However, as in the case of institutional care, mality, work overload, poor career prospects, and, sometimes these requirements are not enforced. in general, a lack of recognition of the work done. For example, in China, data from 2015 show that, In many countries, especially in relatively more at that time, two-thirds of the workers in the sec- 28   Based on interview with Adriana Rovira from INMAYORES published in La Diaria newspaper article, June 2019. Retrieved from: https:// salud.ladiaria.com.uy/articulo/2019/6/solo-25-de-los-940-residenciales-para-ancianos-tienen-habilitacion-del-msp/ COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 10 325 tor lacked the required formal training (Glinskaya Uruguay, where one of the main benefits provided and Feng 2018). A similar situation has been re- by the system is a subsidy to hire a formal caregiv- ported in Serbia. There, while licensed home and er at home, the system does not allow family mem- daycare services are required to employ properly bers to apply for this subsidy (i.e., the person hired trained staff, the lack of a regulated labor supply should be external to the family). In the policy dis- forces many families to rely on the informal labor cussions preceding the implementation of the sys- market (Stokic and Bajec 2018). In other cases, tem, the argument was explicitly made that paying such as Uruguay, the lack of an adequate supply family members would reinforce traditional gender of educational opportunities prevents caregivers roles. This is not the case in Chile, where family from engaging in the necessary trainings, forcing caregivers who are socioeconomically vulnerable the government to relax the requirements initial- can apply for a monthly stipend. While the aim of ly set regarding the accreditation of care workers; both systems is to recognize and value the work of data from 2017 show that, while 18,000 people had caregivers and ultimately to release women from the applied for the caregiver training course given by unpaid obligation of caring, the path to get there is the government, only 1,000 had been able to attend different. While in Uruguay the main objective is to and complete it (Aranco and Sorio 2019). professionalize and formalize the caregiver’s task, in Chile the main objective is to provide some mone- e. The provision of cash benefits tary relief and recognition to those doing the job, re- gardless of whether they are family members or not. An important aspect to consider when defining the Most developed countries have opted for the use combination of services to be offered is how they are of conditional transfers, which are determined based going to be delivered. One of the main decisions is on a personal-care plan that considers the specific whether services are going to be delivered in the form needs of the dependent individual and allows for of direct benefits or in the form of cash transfers. the purchase of the services established in that plan, Cash transfers have the advantage of provid- making it easier to monitor the use of the funds. This ing users with increased flexibility and freedom (as is the case in France, the Netherlands, Scotland, and they allow the dependent person or their family to Spain, for example (Aranco and Ibarrarán 2020). buy the services they choose), and they are usually Cash allowances paid either to the dependent less expensive to deliver (Cafagna et al. 2019). On person or to the main family caregiver are used in the other hand, the use of cash can be difficult to many countries in Eastern Europe, such as Bulgar- monitor. Hence, cash transfers can generate distor- ia, Croatia, the Czech Republic, Estonia, Poland, tions and inefficiencies within families, particularly Serbia, Slovakia, and Ukraine (Slavina et al. 2018; when their use is discretionary (Costa-Font et al. Gerbery and Bednárik 2018; Stubbs and Zrinščak 2018; Costa-Font and Vilaplana 2017). In Serbia, for 2018; Maly 2018; Praat-Ahi and Masso 2018; Hi- example, 72 percent of those who receive the ben- rose and Czepulis-Rutkowska 2016; European efit do not use it to buy care services (Hirose and Commission 2019). Russia approved a compen- Czepulis-Rutkowska 2016). Also, when transfers are sation benefit for caregivers of older people who not tied to the specific purchase of care services, it is cannot participate in the labor market, providing not possible to monitor the quality of such services. some financial security to family caregivers (UN- Moreover, when cash transfers are used to pay ECE 2017). While in a few countries the cash can the main caregiver, it is argued that they tend to rein- be used in the way that best suits the beneficiary force gender roles and hence are a poor instrument or their family (Austria, Lithuania, Serbia, Slove- when one of the aims of the long-term care system nia, and Slovakia, for example), in most cases the is to liberate women from the social mandate that payment is linked to the purchase of care services. assumes they are responsible for caregiving (Cafag- The use of cash benefits to support dependent na et al. 2019; Geyer and Korfhage 2015; Brimble- adults and their families in lower-and middle-in- combe et al. 2018). Conversely, others argue that come countries is not as common and, when such paying women for the care they provide at least gives programs do exist, they are generally focused on the recognition to a job they would be doing anyway socioeconomically vulnerable population. In Chile, (Holmes et al. 2010; Esser, Bilo, and Tebaldi 2019). for example, informal caregivers of severely depen- The cases of Uruguay and Chile in Latin America dent adults that belong to the poorest segments of provide excellent examples of these tradeoffs. In the population are entitled to a monthly stipend, if 326 CHAPTER 10 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E they do not participate in the formal labor market cash transfers might help to mitigate the negative (Molina et al. 2020). In Argentina, economically vul- effects that care responsibilities can have on caregiv- nerable families with a dependent elderly member ers’ physical and emotional health and ensure the receive a monetary subsidy to hire a caregiver; the quality of care (WHO 2015). amount of the subsidy depends on the economic and Support to informal caregivers is more wide- medical evaluation of each situation (Oliveri 2020).29 spread in developed countries, where an iconic ex- Monetary benefits to dependent people or their ample is that of Scotland, where the needs of unpaid caregivers exist in a very few countries in the Mid- caregivers are formally assessed by the government dle East and North Africa, such as Iraq, where a cash and a support plan is elaborated depending on the transfer program for older people with disabilities care burden and the age of the caregiver.30 Benefits has recently been implemented, or Tunisia, where in terms of labor flexibility (Canada, France, Ger- a program of government subsidies is in place. As many, Spain), counseling services (Austria, France, another way to provide additional monetary sup- Germany, the Netherlands, Spain), or respite ser- port to potential caregivers, in most Arab countries vices (Denmark, Finland, Germany, Japan, Korea, the value of pensions increases if the beneficiary has the Netherlands) are among the most common surviving parents (Hussein and Ismail 2017). ways in which developed countries support infor- Cash transfers for dependent older individ- mal caregivers (Aranco and Ibarraran 2020). uals or their caregivers are not common in Asia, In developing countries, other than cash ben- although in some cities in China monthly vouch- efits (discussed in the previous sub-section), infor- ers are given to the dependent person so that he/ mal caregivers’ support relies almost exclusively she can buy care services. Tax incentives for family on counseling and training services. Among Asian caregivers are provided in India, Malaysia, the Phil- countries, counseling services are provided in In- ippines, Singapore, and Thailand, while allowanc- dia, Iran, and Sri Lanka, while training courses are es for low-income families with a dependent elder given in China, Fiji, India, Iran, the Democratic are also given in Sri Lanka (HelpAge International People’s Republic of Korea, Myanmar, Sri Lanka, 2015b; Williamson 2015). In Sub-Saharan Africa, and Viet Nam (Williamson 2015). Examples from the government of Mauritius offers a monthly cash other regions include Chile, Latvia, and Slovenia benefit to family caregivers of people with severe (European Commission 2019). loss of functional capacity (WHO 2017c). Labor flexibility legislation generally does not include those who care for an elderly person, al- f. Other forms of support to though in Costa Rica those who care for a terminally informal caregivers ill patient have the right to unpaid leave from work and to receive a monetary subsidy from the state.31 A Acknowledging the fact that most caregiving re- law project that aims at recognizing the work of fam- sponsibilities are borne by the family, even in ily caregivers by entitling them to a cash allowance, countries where long-term care systems provide subsidizing their contributions to the health care and accessible and high-quality services, many nations pension systems (under specific conditions), and have made provisions to support unpaid caregivers, providing them with counseling and training ser- ranging from the delivery of cash benefits (as dis- vices is currently being discussed in Colombia. In cussed above) to flexible working conditions, train- Chile, the pension system reform project, currently ing, or counseling services. A long-term care sys- under discussion, includes a social security contribu- tem that supports informal caregivers by providing tions’ subsidy for caregivers of people with moderate training, respite care, labor flexibility, counseling, or or severe dependency, for a maximum of two years.32 29   See Instituto de Seguridad Social Salud y Préstamos, Argentina, webpage at https://www.insssep.gob.ar/Coberturas/Index/daafa79f- a7a9-40b4-8d51-a7ae00954052 30   See “The Carers Scotland Act 2016” at https://www.legislation.gov.uk/asp/2016/9/contents 31   Law 7756 of Costa Rica. Retreived from: https://www.pgrweb.go.cr/scij/. 32   See Ministerio de Trabajo y Protección Social Chile. Press Release, November 2nd, 2022. Retrieved from: https://www.pensionesparachile. cl/noticia-02-11-2022.html COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 10 327 Actions to promote social and health care integration poses particularly complex challenges when it comes to health care delivery and calls for a patient-cen- A person-centered and integrated approach to care tered integrated and coordinated system that brings is paramount to achieving the shift in the care para- together not only the different levels of health care digm that is necessary to tackle the challenges of an but also different care systems, such as health and aging society. According to the WHO, “integrated social systems (van der Heide et al. 2015). care refers to services that span the care continuum, The absence of solutions for long-term care are integrated within and among the different levels also affects health care systems, as sometimes older and sites of care within the healthcare and long-term people who are admitted to hospitals for acute or care systems (including within the home) and are of- short-term conditions end up staying longer than fered according to people’s needs throughout the life necessary because they lack other care alternatives course” (WHO 2015, p. 103). (WHO 2015). In fact, evidence shows that the im- The coordination between health and social plementation of long-term care systems could lead care services is key to ensure the continuity of care to a reduction in health care costs in the form of along the life cycle and its multiple health states, fewer hospital admissions and re-admissions, fewer and to prevent as much as possible older adults’ loss emergency entries, and shorter hospital stays (Cos- of functionality. Integrated care not only entails the ta-i-Font et al. 2018; Cho and Kwon 2020; Holland integration of the various medical steps involved et al. 2014; De Souza and Peixoto 2017). in the treatment of a given disease, but also the Some successful examples of integration initia- consideration of the person as a central part of the tives are worth highlighting, both at the community community and the acknowledgment that, in most (local) level and at the national level. Among devel- cases, the need for care exists even after the person oped countries, the System of Integrated Care for leaves the health care system (Amelung et al. 2014). Older Persons (SIPA) in Quebec, Canada, is a com- A comprehensive approach to care that situates the munity-based care program for older people where- individual in their social context and considers not by self-managed multidisciplinary teams of physi- only acute care but also preventive, rehabilitative, cians, nurses, social workers, therapists, home care long-term, and palliative care is needed. workers, and, sometimes, nutritionists and pharma- The aim of social care, of which long-term cists, are responsible for 160 patients each, with the care is part, is to “improve quality of life and subjec- aim of minimizing functional decline and allowing tive well-being of individuals, groups, and commu- aging at home for as long as possible (Amelung et al. nities who are in need or at risk of being in need” 2014). In the United States, a similar initiative exists, (Amelung et al. 2014, p.10). As such, the distinc- called the Program of All-inclusive Care of the Elder- tion between health and social care is sometimes ly (PACE) (Amelung et al. 2014). The Dutch Buurtz- vague, as health and social care needs could affect org model described earlier is also a well-known ex- and reinforce each other (Amelung et al. 2014). ample of a community-led integration approach that However, the integration of these two dimen- has proved to be successful. sions of care is particularly challenging, because The example of Japan, where the government they usually depend on different institutions and has set an agenda for a “Community-based Inte- are subject to different regulations and legislation grated Care System” for older people, also merits at- (Slavina et al. 2018). Moreover, in many countries, tention. The aim is to fully integrate long-term care, social care is underfunded, organized at the lo- health care, and community-provided care services cal level, and is characterized by a large degree of by 2025. The country has been working towards this participation of the community and informal net- aim since 2012 (Morikawa 2014). The model is based works, making coordination and integration with on a decentralization approach in which responsibil- the health care system even more challenging, even ity is gradually shifted from the central government for developed countries (Genet et al. 2012). to local authorities, and where prefectures have re- A lack of coordination, both within health care sponsibility for designing care policies that allow for services and between health and social care services, the most efficient allocation and coordination of the poses problems for the treatment of older people, different models of care (Tomita 2017). particularly given the increasing prevalence of mul- In Sweden, the implementation of the “Esther timorbidity and risk-factor clustering (WHO 2018a; Model” in Jönköping county (named after the real He, Goodkind, and Kowal 2016). Multimorbidity experience of an older patient) shows how a model 328 CHAPTER 10 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E that aligns different care providers in the definition Social Welfare” from Thailand, implemented as pilot of what is best for the patient could have positive programs and involving the collaboration of local impacts on the quality of care. The implementation authorities, volunteers, and older adults themselves of the model has been linked to a 30 percent reduc- (HelpAge International 2013). Another example is tion in emergency department admissions between a pilot project in four municipalities of the Litoral 1998 and 2013 and a 9 percent reduction in the 30- Norte region in the State of Sao Paulo, Brazil, that day readmission rate for patients aged 65 years or took place from November 2018 to November 2019 more.33 Collaboration and coordination between in collaboration with the International Center for providers is encouraged through: (1) quarterly Es- Integrated Care. The project, called “Transforming ther Cafes: cross-sectional meetings where provid- Together,” took an integrative, collaborative, and in- ers share stories from patients; (2) a yearly steering terdisciplinary approach in which each municipality group, which consists of a committee of community identified the objectives that were more important care managers from municipalities, hospitals, and to them in the context of integrated care and worked primary care facilities to discuss challenges across towards them (Integrated Care Foundation 2019). organizations; (3) an annual “strategy day,” where In some countries, such as Chile and Brazil, the nurses, doctors, coaches, and managers come to- health care system has the capacity to make referrals to gether for team-building exercises and to create a social services (including long-term care services), but vision for the network; and (4) ongoing inter-orga- integration does not go much further than that. More- nizational training (Tomita 2017). Due to its proven over, sometimes, even when integration is an explicit success, the model is being replicated in Singapore goal in legislation or government plans, these efforts and the United Kingdom. have not succeeded, as is the case in Bulgaria and the Besides some community-led initiatives, in de- Czech Republic (European Commission 2019). veloping countries, efforts to better integrate health In spite of these efforts, much more is needed and social care services are still incipient. Examples to ensure that older people receive continuous and include the “Bangkok 7 model” and the “Commu- integrated care that encompasses all dimensions of nity-Based Integrated Services of Health Care and care, both at the medical and at the social level. FINAL REMARKS AND KEY POLICY CONSIDERATIONS Population aging is good news; it is a consequence of a myriad of social, economic, and technological advances that allow people to live longer. It is not, however, exempt from challenges. Health care and long-term care sys- tems are particularly affected, as the demographic transition brings about an epidemiological transformation, leading to an increasing prevalence of chronic diseases and rising functional dependency rates. Countries should consider the epidemiological long-term care systems, guaranteeing timely access characteristics underlying their own demographic and quality of care, while promoting prevention and transformation to shape policy responses. If coun- healthier lifestyles, to ease the financial burden that tries observe a compression of morbidity, it would inevitably results from increasing demand. be important to focus on increasing the opportuni- Yet, even in an optimistic scenario, in which ties for older people to continue to contribute active- older people are living healthier lives than in the ly to their communities, adapting labor and pension past, the absolute increase in the number of older markets accordingly, and providing them with ac- people means an absolute increase in the need for cess to life-long learning and training opportunities, care resources (Parker and Thorslund 2007). Pop- promoting active and healthy aging (UN 2017). If, ulation aging comes with a rising prevalence of on the contrary, countries are witnessing an expan- chronic diseases, as evidenced by the increasing sion of morbidity (whereby people are living more trends in conditions that are known to be strongly years with a particular disease or ailment), the pri- age-dependent, such as cardiovascular diseases, di- ority should be to adapt and prepare health care and abetes, dementia, or chronic obstructive pulmonary 33   See “The Esther Model: How one patient redefined an entire system vision in Sweden”, Advisory Board blog post, June 21, 2018. Available at: https://www.advisory.com/blog/2018/06/esther-model COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 10 329 disease (Prince et al. 2015). As the share of older care needs than the younger population but are often adults in the population increases, the importance less able to afford care. In the absence of cost con- of geriatric neurological conditions, such as demen- tainment policies, an increased demand for health tia and Alzheimer’s disease, also grows. Depression care services exerts upward pressure on costs, chal- and musculoskeletal disorders in older adults are lenging the fiscal sustainability of systems. Coun- also major causes of disability. tries need to invest in the efficiency of health care Several characteristics of the epidemiological systems, focusing on prevention and primary care, transition in developing countries create addition- without compromising the quality of care. al pressures. These include the following facts: (i) The provision of long-term care services for Many of today’s older adults in developing coun- people with diminished functional abilities needs tries were exposed to infectious diseases in their immediate attention. Given the changes that are early childhood, with unknown effects on their taking place at the societal and family levels, and health as older adults; (ii) The double burden of the lack of affordable, high-quality private services disease common in developing countries increases in many settings, the role of the government in pro- the risk of multimorbidity; (iii) Under-diagnosis viding the long-term care services that functionally of key medical conditions and lack of affordable, dependent people need becomes crucial (Caruso, timely, and effective treatments are common prob- Galiani, and Ibarrarán 2017). lems in many developing countries; and (iv) The Besides being accessible and affordable, long- adoption of harmful habits such as smoking, alco- term care services should be person-centered, en- hol use, lack of physical activity, and unhealthy eat- couraging - whenever possible - the provision of ing has fueled the increase in chronic diseases and home care services, as aging in place has proved to the speed of the epidemiological change. be the preferred option for older people and their Thus, population aging comes with at least families. Aging in place has also been shown to de- three main challenges for health care systems in de- liver the best results in terms of older people’s men- veloping countries: (i) first, guaranteeing coverage, tal and physical health and to be cost-effective for access, and affordability of services in a context of governments (WHO 2015; Boland et al. 2017). Co- increasing demand; (ii) second, shifting the care ordination and integration between the social care model from a curative-based approach towards a and health care sectors is of paramount importance person-centered holistic and integrated approach both to achieving a continuum of care for the older that encourages the continuum of care throughout person and securing needed efficiency gains. the life course, with a greater emphasis on primary The training and adaptation of human re- health care and health promotion and prevention sources is an area in which progress has been slow. activities; and (iii) third, keeping costs under con- In most developing countries, the number of geri- trol without losing attention to quality. atricians is insufficient, and medical personnel, in How people age is, to a large extent, deter- general, lack the necessary training to treat older mined by their health earlier in life and the choices adults. The professionalization of human resources they made when young. Countries should therefore in the long-term care sector is even less developed. promote healthy lifestyles, like physical activity and Investing in the training of human resources, their healthy eating, through the entire life course (Re- professionalization, and the improvement of their chel et al. 2009). Adaptation of health care systems working conditions is essential not only to improve is critical to anticipate and successfully manage the the quality-of-care older people receive but also to consequences of an aging population, especially giv- narrow the important gender gaps that still exist in en the challenges of coverage, accessibility, adequa- terms of labor participation, financial autonomy, cy and quality of services, and financing outlined in and the burden of care responsibilities. previous sections of this chapter. Moreover, for the As this review has shown, several developing appropriate management of chronic diseases, a per- countries have designed and implemented some son-centered model focused on adequate attention of the key policies described above. In the case of to comorbidities is crucial (Grumbach 2003). health care, promising policies have aimed at in- Health care systems need to adapt and prepare. creasing coverage, adapting health care systems to Universal health care coverage should adopt distinc- new epidemiological profiles, focusing on the pre- tive characteristics in countries with aging popula- vention of chronic diseases, and promoting healthy tions, as older adults not only have different health aging. Attention has also been given to the devel- 330 CHAPTER 10 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E opment or reinforcement of mental health policies, health and social care to others across the life course. an area that was until recently somewhat neglected. Notable are the unpaid care responsibilities women Long-term care policies and programs aimed often have to bear and the consequences that these at assisting those with difficulties in performing have both for women’s mental and physical health their daily activities have also been part of the pol- and their income-generation capacity. icy agenda in some countries, although important There are important limitations regarding data gaps still exist in covering the needs of functionally related to aging and health (Box 10.4). These gaps dependent older adults. Despite efforts, progress may complicate the design of effective policies. Data has been slow and uneven (Zaidi 2018; UN 2017b). limitations were found in: (i) the determinants of The scope and depth of policies vary, and, in some healthy aging; (ii) the evolution of health and func- cases, little is known about their actual implemen- tionality as people age; (iii) the variation of health tation (Williamson 2015). inequalities among older adults over time and across Future policy responses to population aging and within countries; and (iv) the needs and prefer- should also consider the important distributional is- ences of older adults regarding health care and long- sues that come with population aging. These include term care services. These are areas where granular gender gaps; intergenerational disparities; spatial dif- and specialized data are needed to enable the changes ferences (e.g., rural vs. urban); and socioeconomic required in health care and long-term care systems. inequalities (World Bank 2021). This will require Examples include data on disability, functional de- paying more attention to the unequal burdens faced pendency, cognitive status, and mental health, as well by women both in accessing services and in providing as health care service utilization (World Bank 2021). BOX 10.4 The importance of timely and sound data To facilitate evidence-based policy decisions, data need to be comparable across countries and time, yet consensus does not exist on key definitions or on which data should be prioritized. As early as 2001, a seminal report from the National Research Council [NRC] highlighted the need for improved data on aging and stressed the importance of comparable cross-national and longitudinal data (Moore 2001). This has been echoed more recently in the milestone WHO report on Ageing and Health (2015) and subsequent work. Attention is needed to the harmonization and standardization of data collected across countries (Boersch-Supan 2016). Despite notable efforts made in countries, longitudinal, good-quality, and comparable information regarding the health status of the elderly is either lacking or difficult to access, particularly in developing countries. Data on health status (either self-re- ported or through medical diagnosis) is more commonly available than data on disability, functional dependency, risk factors, cognitive status, mental health, and health care service utilization. Health data come from different sources (ranging from censuses or nationally representative household surveys to administrative records, vital statistics, and specific/targeted surveys, among others), but none of these sources is without limitations. Few countries have specific health surveys targeted at older adults. The Health, Ageing and Retirement Surveys are an example of such surveys. They are available in China, Costa Rica, England, India, Ireland, Israel, Japan, Korea, the United States, and 27 European countries. Multiple waves have been conducted, which allows for longitudinal analysis. The WHO Surveys on Ageing and Health, carried out in several developing countries during 2006/2007 and later in 2014, offer another example of such efforts. In Latin America, the Longitudinal Social Protection Surveys (ELPS as per the acronym in Spanish, standing for “Encuesta Longitudinal de Protección Social”) are another group of surveys with specific focus on the elderly, available in Chile, Colombia, El Salvador, Paraguay, and Uruguay. Even when surveys were designed to enable comparability, differences in methodology can prevent cross-national compar- isons or generalizations. Even when the same “type” of questions are included in health-specific surveys, the exact questions are often different, making comparisons difficult. For instance, the section on functional dependency in the Uruguay ELPS survey makes it the only country administering an ELPS survey to include both basic and instrumental activities of daily living. Source: World Bank (2021). COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 10 331 BOX 10.4 The importance of timely and sound data (continued) The other ELPS countries only include basic activities, yet these activities are not standardized in ELPS surveys across countries. Census and National Household Surveys sometimes (but not always) include a health section, but in those cases, cross-country comparison is even more difficult. Administrative and medical records, which usually span the entire life of the individual, are also rich sources of information; however, harmonization across countries may require addressing issues of format and soft- ware compatibility (Moore 2001). Privacy concerns also usually surround the use of administrative records. Being able to link administrative and medical records to survey data would open immense research possibilities useful for evidence-based policy (Moore 2001). The National Research Council recommends that countries follow a system of “hierarchy of data collection modules,” where minimum sets of data are defined for each module, from basic data to increasingly elaborated datasets, that guarantee compa- rability across countries. These datasets should include, at least, “the frequency and rates for: (1) deaths and their main causes, (2) main medical conditions, (3) self-reported health status, (4) population levels of physical, social, and mental function, (5) preventive and health promotional behaviors, and (6) disabilities. Regarding health care utilization data, the minimum informa- tion should include: (1) utilization rates for important types of health services, along with the use of long-term care services; (2) personal and family expenses for formal health services; (3) rates of use of medications and devices; (4) major cultural influences on the concept of health and the use of health services (such as gender, ethnicity, geographic residence, and socioeconomic status); and (5) the use of informal and alternative and complementary health care services” (Moore 2001). Source: World Bank (2021). Finally, several issues not considered in this care and long-term care services. In many coun- chapter are part of the pressures placed by popula- tries, COVID-19 has highlighted the weaknesses of tion aging on health and long-term care systems and systems, as the lack of adequate resources and pro- countries’ responses to them. These include: the im- tocols has led to unprecedented strains on health portance of technology, the role of the private sector care systems. Long-term care services unprepared in the delivery of services, the need to strengthen for such a shock have also been disrupted, leaving palliative care services, the importance of investing many older adults without the assistance they need. in more resilient health care and social care systems, Although the impacts of the health and economic and the need to educate not only health and long- crisis generated by the pandemic are not yet fully term care workers but the whole of society to fight understood, one thing is sure: the consequences will ageism and discrimination towards the elderly. Also, be unprecedented, and all countries will need to re- the impacts of the pandemic generated by the spread vise their social protection matrix (including health of the severe acute respiratory syndrome coronavirus care and long-term care systems) in order not only 2 (SARS-CoV-2) in 2020 were not part of this review. to deal with the consequences of the current coro- Undoubtedly, the COVID-19 pandemic pres- navirus, but to be prepared for other potential crises ents a unique set of age-related challenges to health in the future (Twigg et al. 2021). REFERENCES 1. Addati, L., U. Cattaneo, V. Esquivel, and I. Valarino (2018). Care 4. ADI (Alzheimer’s Disease International) (2019a). From plan to work and care jobs for the future of decent work. International impact II: the urgent need for action. London, England. ADI. Labour Office, Geneva. 5. ADI (Alzheimer’s Disease International) (2019b). World Alzhei- 2. ADI (Alzheimer’s Disease International) (2015). World Alzheimer mer Report 2019: Attitudes to dementia. London, England. ADI. Report 2015: The global impact of dementia: An analysis of preva- 6. Afshar, S., PJ. Roderick, P. Kowal, BD. Dimitrov, and AG. Hill. lence, incidence, cost and trends. London, England. ADI. (2015). “Multimorbidity and the inequalities of global ageing: 3. ADI (Alzheimer’s Disease International) (2017). Dementia a cross-sectional study of 28 countries using the World Health Friendly Communities Global developments, 2nd Edition. Lon- Surveys.” BMC Public Health 15(1): 776. don, England. ADI. 7. Aguilera, I., A. Infante, H. Ormeño, and C. Urriola. Improving 332 CHAPTER 10 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E health system efficiency: Chile: Implementation of the Universal Dmytraczenko, P. Frenz, P. Garcia, O. Gómez-Dantés, F.M. Knaul, Access with Explicit Guarantees (AUGE) reform. WHO/HIS/HGF/ C. Muntaner and J.B. De Paula (2015). “Health-system reform CaseStudy/15.3. World Health Organization. and universal health coverage in Latin America.” The Lancet 8. Aguirre, M. (2014). La institucionalización de la vejez en Uruguay. 385(9974): 1230-1247. Paper presented at the XIII Research Seminar of the Faculty of 22. Barr, N. (2010). “Long‐term care: a suitable case for social insur- Social Ciences, Public Uruguayan University, UdelaR, Montevi- ance” Social Policy and Administration 44(4): 359-374. deo, September de 2014. 23. Barraza, M. (2017). “Dependencia y apoyo a los cuidados, un 9. AHWAIN (Asia Health and Wellbeing Initiative) (2020). “The asunto de derechos humanos” en Forttes, P. (ed) Subsistema Intergenerational Self-Help Club (ISHC) Development Model”. Nacional de Apoyos y Cuidados: Un desafío país: 60-67. Gobierno Available at: https://www.ahwin.org/helpage-vietnam-ishc/. de Chile. Dirección Sociocultural, Presidencia de la República. 10. Aikins, A, N. Unwin, C. Agyemang, P. Allotey, C. Campbell, and 24. Batthyany, K., N. Genta and V. Perrotta (2012). “La población D. Arhinful (2010). “Tackling Africa’s chronic disease burden: uruguaya y el cuidado: persistencias de un mandato de género. from the local to the global.” Globalization and Health 6(1): 1-7. Encuesta nacional sobre representaciones sociales del cuidado, 11. Alfonso-Sierra, E.; Arcila Carabalí, A.; Bonilla Torres, J.; La- principales resultados”. Serie Mujer y Desarrollo 117, CEPAL. torre Castro, M.L.; Porras Ramírez, A. and Urquijo Velásquez, 25. Ben-Harush A, Shiovitz-Ezra S, Doron I, Alon S, Leibovitz A, Go- L. 2018. Situación de multimorbilidad en Colombia 2012 lander H, Haron Y, Ayalon L. Ageism among physicians, nurses, -2016. Ministerio de Salud y Protección Social. Estudios Sec- and social workers: findings from a qualitative study. Eur J Age- toriales. Bogotá, Colombia. https://documents1.worldbank. ing. 2016 Jun 28;14(1):39-48. doi: 10.1007/s10433-016-0389-9. org/curated/en/801401550612917615/pdf/134506-SPAN- 26. Bishwajit, G., S. Tang, S. Yaya, and Z. Feng (2017). “Burden of I S H - W P- P 1 6 4 6 3 2 - O U O - 9 - M u l t i m o r b i l i d a d - e n - Co - asthma, dyspnea, and chronic cough in South Asia”. Interna- lombia-sin-formato.pdf. tional journal of chronic obstructive pulmonary disease 12: 1093- 12. Allen, L., J. Williams, N. Townsend, B. Mikkelsen, N. Roberts, C. 1099. Foster, and K. Wickramasinghe (2017). “Socioeconomic status 27. Blas, E., and A. S. Kurup, eds. (2010). Equity, social determinants and non-communicable disease behavioural risk factors in and public health programmes. World Health Organization. low-income and lower-middle-income countries: a systematic 28. Bloom, D. E., E. Cafiero-Fonseca, M.E. McGovern, K. Prettner, A. review.” The Lancet Global Health 5(3): e277-e289. Stanciole, J. Weiss, and L. Rosenberg (2014). “The macroeco- 13. Almeida, G., O. Artaza, N. Donoso, and R. Fábrega (2018). “La nomic impact of non-communicable diseases in China and atención primaria de salud en la Región de las Américas a 40 India: Estimates, projections, and comparisons”. The Journal of años de la Declaración de Alma-Ata.” Revista Panamericana de the Economics of Ageing 4: 100–111. Salud Pública 42: e104. 29. Bloom, G., Y. Katsuma, K. D. Rao, S. Makimoto, J. DC Yin, and G. 14. Amelung, V. E., A. Reichert, D. Urbanski, L. Matejevic, E. O’rior- M. Leung (2019). “Next steps towards universal health cover- dan, and E. Blatt. (2014). “Integrating Health and Social Care A age call for global leadership.” Bmj 365. global perspective of experience, best practices and the way 30. Boersch-Supan (2016). “Enhancing the Comparability of SHARE forward.” INAV–Institute for Applied Health Services Research. with HRS and ELSA”. Max Planck Institute/Social Law/Social 15. Aranco, N., M. Stampini, P. Ibarrarán, Medellín, N. (2018). “Pan- Policy, Muenchen, Germany. orama de envejecimiento y dependencia en América Latina y 31. Bogdanov, G. and L. Georgieva (2018). ESPN Thematic report el Caribe”. Resúmen de Políticas IDB-PB-273. Banco Interamerica- on challenges on long-term care: Bulgaria. The European Social no de Desarrollo. Policy Network, European Commission. 16. Aranco, N. and R. Sorio (2019). “Envejecimiento y atención a la 32. Boland, L., F. Légaré, M. Becerra Perez, M. Menear, M. Garvelink, dependencia en Uruguay”. Nota Técnica IDB-TN-1615. Banco D. McIsaac, G. Painchaud Guérard, J. Emond, N. Brière, and S. Interamericano de Desarrollo. Dawn (2017). “Impact of home care versus alternative loca- 17. Aranco, N. and Ibarrarán, P. (2020). “Servicios de apoyo person- tions of care on elder health outcomes: an overview of system- al para personas con dependencia funcional: Antecedentes, atic reviews”. BMC Geriatrics 17(1): 20. características y resultados”. Nota Técnica IDB-TN-1884. Banco 33. Bonilla-Chacin, M.E., G. Afandiyeva, and A. Suaya (2018). “Chal- Interamericano de Desarrollo. lenges on the path to universal health coverage: the experi- 18. Aranco, N., Ibarrarán, P., Stampini, M. (2022). “Prevalencia de ence of Azerbaijan”. Universal Health Care Coverage Series 28, la dependencia funcional entre las personas mayores en World Bank Group, Washington, DC. 26 países de América Latina y el Caribe”. Nota Técnica IDB- 34. Boutayeb, A., S. Boutayeb, and W. Boutayeb (2013). “Multi-mor- TN-2470. Banco Interamericano de Desarrollo. bidity of non-communicable diseases and equity in WHO East- 19. Aranco, N., Bosch, M., Stampini, M., Azuara, O., Goyeneche, L., ern Mediterranean countries.” International journal for Equity in Ibarrarán, P., Oliveira, D., Reyes, M., Savedoff, W., Torres, E. (2022). Health 12.1 (2013): 1-13. Envejecer en América Latina y el Caribe: protección social y calidad 35. Brimblecombe, N., J.L. Fernandez, M. Knapp, A. Rehill and R. de vida de las personas mayores. Banco Interamericano de De- Wittenberg (2018). “Review of the international evidence on sarrollo. Arsenijevic J, Pavlova M, Rechel B, Groot W (2016) Cat- support for unpaid carers”. Journal of Long-Term Care Septem- astrophic Health Care Expenditure among Older People with ber: 25 - 40. Chronic Diseases in 15 European Countries. PLoS ONE 11(7): 36. Burki, T.K. (2017). “Latin America makes progress on tobacco e0157765. https://doi.org/10.1371/journal.pone.0157765. control.” The Lancet Respiratory Medicine 5(6): 470. 20. Arredondo, A., and A. L. Recaman (2018). “Determinants of 37. Byers, AL., and K. Yaffe (2011). “Depression and risk of develop- uncontrolled hypertension in the context of universal health ing dementia.” Nature Reviews Neurology 7(6): 323-331. coverage in middle-income countries.” American Journal of Hy- 38. Cafagna, G., N. Aranco, P. Ibarrarán, M.L. Oliveri, N. Medellín, pertension 31(11): 1175-1177. and M. Stampini (2019). Envejecer con cuidado: atención a la 21. Atun, R., L. O. Monteiro De Andrade, G. Almeida, D. Cotlear, T. dependencia en América Latina y el Caribe. Inter-American De- COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 10 333 velopment Bank. Centre for Inclusive Growth. Working Paper 178. 39. Caruso Bloeck, M, S. Galiani, and P. Ibarrarán (2017). “Long-term 58. European Commission 2019. Joint report on health care and care in Latin America and the Caribbean? Theory and policy long-term care systems & fiscal sustainability. Country Documents considerations”. IDB Working Paper Series Nº IDB-WP- 834. 2019 Update. Institutional Paper 105. 40. Vaca, I. (2021). “Valorización económica del trabajo no remu- 59. Fang, H., K. Eggleston, K. Hanson, and M. Wu. (2019)”Enhanc- nerado de los hogares”. División de Asuntos de Género, Comis- ing financial protection under China’s social health insurance ión Económica para América Latina y el Caribe. to achieve universal health coverage.” bmj 365: l2378. 41. Chang, A. Y., Skirbekk, V. F., Tyrovolas, S., Kassebaum, N. J., and 60. Feng, Z., C. Liu, X. Guan, and V. Mor (2012). “China’s rapidly ag- Dieleman, J. L. (2019). “Measuring population ageing: an analy- ing population creates policy challenges in shaping a viable sis of the global burden of disease study 2017.” The Lancet Pub- long-term care system.” Health Affairs 31(12): 2764-2773. lic Health, 4(3), e159-e167. 61. Flamini, V., M. Galdamez, F. Lambert, M. Li, B. Lissovolik, R. 42. Chateau, D., A. Katz, C. Metge, C.C.M. Taylor, and S. McCulloch Mowatt, J. Puig, A. D. Klemm, and M Soto. Growing Pains: Is (2017). Describing Patient Populations for the My Health Team Latin American Prepared for Population Aging? International Initiative. Manitoba Centre for Health Policy. Monetary Fund, 2018. 43. Chatterji, S., Byles, J., Cutler, D., Seeman, T., and Verdes, E. (2015). 62. Flores, E., E. Rivas, and F. Seguel. “Nivel de sobrecarga en el “Health, functioning, and disability in older adults—present desempeño del rol del cuidador familiar de adulto mayor con status and future implications”. The Lancet 385: 563–75. dependencia severa.” Ciencia y Enfermería 18(1): 29-41. 44. Chen, Y., H. Ding, M/. Yu, J. Zhong, R. Hu, X. Chen, C. Wang, 63. Fries, J. F. (1980). “Aging, natural death, and the compression of K. Xie. and K. Eggleston (2019). “The Effects of Primary Care morbidity”. New England Journal of Medicine 303(3): 130–135. Chronic-Disease Management in Rural China” National Bureau 64. Gaál P., M. Csere, A. Conklin, et al. (2015). “Hungary. In: Nolte E, of Economic Research w26100. Knai C, editors. Assessing Chronic Disease Management in Euro� - 45. Cho, Y. M., & Kwon, S. (2022). Effects of public long-term care pean Health Systems: Country reports”. Copenhagen (Denmark): insurance on the medical service use by older people in South European Observatory on Health Systems and Policies 29(8.) Korea. Health Economics, Policy and Law, 1-18. 65. Gakidou, E., A. Afshin, A. A. Abajobir, K. H. Abate, C. Abbafati, 46. Clarfield, A. M., O. Manor, G. Bin Nun, S. Shvarts, Z. S. Azzam, A. K. M. Abbas, F. Abd-Allah et al. (2017). “Global, regional, and Afek, F. Basis, and A. Israeli (2017). “Health and health care in national comparative risk assessment of 84 behavioural, envi- Israel: an introduction.” The Lancet 389(0088): 2503-2513. ronmental and occupational, and metabolic risks or clusters of 47. Coleman, R., G. Gill, and D. Wilkinson (1998). “Noncommunica- risks, 1990–2016: a systematic analysis for the Global Burden of ble disease management in resource-poor settings: a primary Disease Study 2016”. The Lancet 390(10100): 1345-1422. care model from rural South Africa.” Bulletin of the World Health 66. Gal, R.I. (2018). ESPN Thematic report on challenges on long-term Organization 76(6): 633-640. care: Hungary. The European Social Policy Network, European 48. Colombo, F., A. Llena-Nozal, J. Mercier, and F. Tjadens (2011). Commission. “Help Wanted? Providing and Paying for Long-Term Care” OECD 67. Garg, S. et al. (2020). “Hospitalization rates and characteristics Health Policy Studies. OECD Publishing. of patients hospitalized with laboratory-confirmed coronavirus 49. Costa-Font, J., S. Jimenez-Martin, and C. Vilaplana-Prieto disease 2019—COVID-NET, 14 States, March 1–30, 2020.” Cen- (2018). “Thinking of incentivizing care? The Eefect of demand ter for Disease Control and Prevention. Morbidity and Mortality subsidies on informal caregiving and intergenerational trans- Weekly Report 69. fers”. IZA Discussion Papers 11774. 68. Gatti Roberta, Daniel Halim, Allen Hardiman and Shuqiao Sun. 50. Costa‐Font, J. and C. Vilaplana‐Prieto (2017). “Does the expan- 2024. “Gendered Responsibilities, Elderly Care, and Labor Sup- sion of public long‐term care funding affect saving behaviour?. ply: Evidence from Four Countries,” in Unlocking the Power of Fiscal Studies, 38(3): 417-443. Healthy Longevity: Compendium of Research for the Healthy 51. Costa-Font, J., Jimenez-Martin, S., & Vilaplana, C. (2018). Does Longevity Initiative. Washington D.C.: World Bank. long-term care subsidization reduce hospital admissions and 69. GBD DALYs and HALE Collaborators (2017). “Global, regional, utilization? Journal of health economics, 58, 43-66. and national disability-adjusted life-years (DALYs) for 359 dis- 52. De la Maisonneuve, C. and J. Oliveira Martins (2013). “Public eases and injuries and healthy life expectancy (HALE) for 195 spending on health and long-term care: a new set of projec- countries and territories, 1990–2017: a systematic analysis tions.” OECD Economic Policy Papers. for the Global Burden of Disease Study 2017”. The Lancet 392: 53. Dmytraczenko, T. and G. Almeida (eds.) (2015). Toward universal 1859–922. health coverage and equity in Latin America and the Caribbean: 70. Genet, N., W. Boerma, M. Kroneman, A. Hutchinson and R.B. evidence from selected countries. The World Bank. Saltman eds. (2012). Home care across Europe: current structure 54. Druetz, T. (2019). “Integrated primary health care in low-and and future challenges. European Observatory on Health Sys- middle-income countries: a double challenge.” BMC medical tems and Policies, World Health Organization. ethics 19(1): 89-96. 71. Gerbery, D. and R. Bednárik (2018). ESPN Thematic report on 55. Dunlop, D. D., S. L. Hughes and L.M. Manheim (1997). “Disability challenges on long-term care: Slovakia. The European Social Pol- in activities of daily living: patterns of change and a hierarchy icy Network, European Commission. of disability”. American Journal of Public Health, 87(3): 378-383. 72. Geyer, J. and T. Korfhage (2015). “Long - term care insurance 56. Esquivel, V.R., and A. Kaufmann. Innovations in care: New con- and carers’ labor supply: a structural model”. Health Economics cepts, new actors, new policies. Friedrich Ebert Stiftung. 24(9):1178-1191. 57. Esser, A., C. Bilo, and R. Tebaldi (2019). “How can cash transfer 73. Giraldo Molina, C.I and G.M. Franco Agudelo. “Calidad de vida programmes work for women and children? A review of gen- de los cuidadores familiares: Life Quality among Family Carers.” der-and child-sensitive design features” International Policy Aquichan 6(1): 38-53. 334 CHAPTER 10 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E 74. Glassman, A., T. A. Gaziano, C.P. Bouillon Buendia, and F. C. Guanais take. Published in Health Systems & Reform 2021, vol. 7, no. de Aguiar (2010). “Confronting the chronic disease burden in Lat- 2, e1911067 https://doi.org/10.1080/23288604.2021.1911067. in America and the Caribbean.” Health Affairs 29 (2): 2142-2148. 95. Hoang, H. T., Nguyen, T., & Medina, C. K. (2023). Faith-based or- 75. Glinskaya, E., and Z. Feng (eds) (2018). Options for aged care in ganizations’ support for older adults in Vietnam: A comparison China: Building an efficient and sustainable aged care system. of Catholic and Buddhist efforts. Social Work and Social Sciences The World Bank, Washington D.C. Review, 24(1). 76. Glinskaya, Elena, Xiaohui Hou, Zhanlian Feng, Marco Angrisani, 96. Hoogendijk, E.; Rijnhart, J. Kowal, P.; Pérez-Zepeda, M.; Cesari, Guadalupe Suarez, Jigyasa Sharma, Drystan Phillips, et al. 2024. M.; Abizanda, P.; Flores Ruano, T.; Schop-Etman, A.; Huisman, “Demand for and Supply of Long-Term Care for Older persons M.; Dent, E. (2018) Socioeconomic inequalities in frailty among in Low- and Middle-Income Countries,” in Unlocking the Power older adults in six low- and middle-income countries: Re- of Healthy Longevity: Compendium of Research for the Healthy sults from the WHO Study on global AGEing and adult health Longevity Initiative. Washington D.C.: World Bank. (SAGE), Maturitas, Volume 115, 2018: 56-63, ISSN 0378-5122, 77. Goeppel, C., P. Frenz, P. Tinnemann, and L. Grabenhenrich https://doi.org/10.1016/j.maturitas.2018.06.011. (2014). “Universal health coverage for elderly people with 97. Holmes, R. and N. Jones (2010). “Cash transfers and gendered non-communicable diseases in low-income and middle-in- risks and vulnerabilities: lessons from Latin America.” Overseas come countries: a cross-sectional analysis.” The Lancet 384: S6. Development Institute. 78. Goldman, N. (2001). Social inequalities in health. Annals of the 98. Holland, S. K., Evered, S. R., & Center, B. A. (2014). “Long-Term New York Academy of Sciences 954(1):118–139. Care Benefits May Reduce End-of-Life Medical Care Costs.” Pop- 79. Gruenberg, E. (1977). “The failure of success”. Milbank Memorial ulation Health Management, 17(6), 332–339. fund Quarterly/ Health and Society 55: 3–24. 99. Honorato dos Santos de Carvalho, V.C., Rossato, S.L., Fuchs, F.D. 80. Grumbach, K. (2003). “Chronic illness, comorbidities, and the et al. (2013). “Assessment of primary health care received by need for medical generalism”. Annals of Family Medicine 1(1): 4-7. the elderly and health related quality of life: a cross-sectional 81. Guanais, F., F. Regalia, R. Pérez-Cuevas, M. Anaya. eds. (2018). study”. BMC Public Health 13 (605). From the patient’s perspective: experiences with primary health 100. Howse, K. (2006). “Increasing life expectancy and the compres- care in Latin America and the Caribbean. Inter-American Devel- sion of morbidity: a critical review of the debate”. Oxford Instute opment Bank. of Ageing. Working Paper 206. 82. Han, W. (2012) “Health care system reforms in developing 101. Huenchan, S. ed. (2010). Ageing, human rights and public pol- countries.” Journal of public health research 1(3): 199. icies. CEPAL. Serie Libros de la CEPAL – Desarrollo Social 100. 83. He Y, Lam TH, Jiang B et al. (2009). “Combined effects of tobac- 102. Hughes, D, and S. Leethongdee. (2007). “Universal coverage co smoke exposure and metabolic syndrome on cardiovas- in the land of smiles: lessons from Thailand’s 30 Baht health cular risk in older residents of China”. Journal of the American reforms.” Health Affairs 26 (4): 999-1008. College of Cardiology 53: 363–71. 103. Hussein, S., and M. Ismail (2017). “Ageing and elderly care in 84. He, W., D. Goodkind, and P. Kowal (2016). “An aging world: the Arab region: policy challenges and opportunities.” Ageing 2015, international population reports.” Washington: US Gov- International 42(3): 274-289. ernment Printing Office. Washington DC. 104. Ibrahim, M. M, and A. Damasceno (2012) “Hypertension in de- 85. http://www.census.gov/library/publications/2016/demo/ veloping countries.” The Lancet 380 (9841): 611-619. P95-16-1.html. 105. International Labour Office (ILO) (2016). Women at work: trends 86. Hedrich, W., J. Tan, B. Chalmers, and J. Yeo. “Advancing into the 2016. Geneva: ILO. Golden Years: Cost of Healthcare for Asia Pacific’s Elderly.” Marsh 106. International Diabetes Federation. (2013). IDF Diabetes Atlas. & McLennan Companies Asia Pacific Risk Center: Singapore. 6th Edition. 87. HelpAge International (2013). “Care in old age in Southeast 107. International Diabetes Federation. (2019). IDF Diabetes Atlas. Asia and China: situational analysis”. 9th Edition. 88. HelpAge International (2014a). “Why health systems must 108. Jeet, G., J. S. Thakur, S. Prinja, and M. Singh (2017). “Communi- change: Addressing the needs of ageing populations in low- ty health workers for non-communicable diseases prevention and middle-income countries”. and control in developing countries: evidence and implica- 89. HelpAge International (2014b). “Home care for older people: tions.” PloS One 12(7): e0180640. The experience of ASEAN countries”. 109. Kagawa, R. C., Anglemyer, A., & Montagu, D. (2012). The scale 90. HelpAge International (2015a). Forget Me Not. Improving de- of faith-based organization participation in health service de- mentia care in Andean countries. Project Summary. HelpAge livery in developing countries: systemic review and meta-anal- International. Regional Development Centre for Latin America ysis. PloS one, 7(11), e48457. and the Caribbean. 110. Kaselitz, E., G. K. Rana, and M. Heisler (2017). “Public policies 91. HelpAge International (2015b). Community-based social care in and interventions for diabetes in Latin America: a scoping re- East and Southeast Asia HelpAge International. East Asia and view.” Current Diabetes Reports 17 (8): 65. Pacific Regional Office. 111. Katz, S. (1983). “Assessing self-maintenance: activities of daily 92. Hirose, K. and Z. Czepulis-Rutkowska (2016). “Challenges in living, mobility, and instrumental activities of daily living”. Jour- Longterm Care of the Elderly in Central and Eastern Europe.” ILO nal of the American Geriatrics Society 31(12): 721-727. DWT and Country Office for Central and Eastern Europe, Budapest. 112. Khanam, M. A., P. K. Streatfield, Z. N. Kabir, Qiu, C. Cornelius, and 93. Hoang, V.M. and H.Y. Duong (2018). “Health and health care for Å. Wahlin (2011). “Prevalence and patterns of multimorbidity older people in Vietnam”. Healthy Aging Research 7:15 among elderly people in rural Bangladesh: a cross-sectional 94. Hoang-Vu Eozenou, P.; Neelsen, S. and Smitz, M.F. 2021. Finan- study”. Journal of Health, Population, and Nutrition 29(4): 406-414. cial Protection in Health among the Elderly – A Global Stock- 113. Kieny, M.P.; Evans,T.; Scarpetta, S.; Kelley, E.T.; Klazinga, N.; Forde, COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 10 335 I.; Veillard, J.; Leatherman, S.; Syed, S.; Kim, S.; Nejad, S.; Don- Journal of Aging and Health 21(2): 266-285. aldson, L. 2018. “Delivering quality health services: a global 128. Loichinger, E. and W. Pothisiri (2018). “Health prospects of old- imperative for universal health coverage.” Washington, D.C.: er persons in Thailand: the role of education.” Asian Population World Bank Group. http://documents.worldbank.org/curat- Studies 14(3): 310-329. ed/en/482771530290792652/Delivering-quality-health-ser- 129. López-Ortega, M. and Aranco, N. (2019). “Envejecimiento y vices-a-global-imperative-for-universal-health-coverage. atención a la dependencia en México”. Nota Técnica IDB-TN 114. Kilpi, F., L. Webber, A. Musaigner, A. Aitsi-Selmi, T. Marsh, K. Rt- 1614. Banco Interamericano de Desarrollo. veladze, K. McPherson, and M. Brown. (2014). “Alarming pre- 130. López, J. H. and C. A. Reyes-Ortiz (2015). “Geriatric education in dictions for obesity and non-communicable diseases in the undergraduate and graduate levels in Latin America.” Gerontol- Middle East.” Public Health Nutrition 17 (5): 1078-1086. ogy & Geriatrics Education 36(1): 3-13. 115. Knaul, F.M., R. Wong, H. Arreola-Ornelas, O. Méndez, R. Bitran, 131. Lustigova, M., D. Dzurova, H. Pikhart, R. Kubinova, and M. Bobak A. Carlos Campino, C.E. Flórez Nieto, U. Giedion, D. Maceira, M. (2018). “Cardiovascular health among the Czech population at Rathe and M. Valdivia. (2011). “Household catastrophic health the beginning of the 21st century: a 12-year follow-up study.” expenditures: a comparative analysis of twelve Latin American Journal of Epidemiology and Community Health 72(5): 442-448. and Caribbean Countries”. Salud Pública de México 53: s85-s95. 132. Luy, M. and Minagawa, Y., 2014. “Gender Gaps-Life expectancy 116. Knodel, J., B.P Teerawichitchainan, V. Prachuabmoh, and W. and proportion of life in poor health.” Health Reports 25(12): 12. Pothisiri (2015). “The situation of Thailand’s older population: 133. Macías Muñoz, O. (2019). “Reforma al sistema chileno de pen- An update based on the 2014 Survey of Older Persons in Thai- siones”. XVII Convención Anual de la Asociación de Administrado- land.” Research Collection School of Social Sciences 1948. Institu- ras de Fondos Mutuos de Chile (AAFM). tional Knowledge at Singapore Management University. 134. Macinko, J., B. Starfield, and T. Erinosho (2009) “The impact 117. Kreitzer, M.J., K. A. Monsen, S. Nandram, and J. De Blok (2015). of primary healthcare on population health in low-and mid- “Buurtzorg Nederland: a global model of social innovation, dle-income countries.” The Journal of ambulatory care manage- change, and whole systems healing.” Global Advances in ment 32(2): 150-171. Health and Medicine 4(1): 40-44. 135. Macinko, J., F.C. Guanais, P. Mullachery, and G. Jimenez (2016). 118. Kronfol, N.M. (2012). “Access and barriers to health care delivery “Gaps in primary care and health system performance in six in Arab countries: a review”. Eastern Mediterranean Health Jour- Latin American and Caribbean countries.” Health Affairs 35(8): nal 18 (12): 1239-1246. 1513-1521. 119. Kruk, M.E., D. Porignon, P. C. Rockers, and W. Van Lerberghe 136. Macinko, J., F.C.D Andrade, B. P. Nunes, and F. C. Guanais (2019). (2010). “The contribution of primary care to health and health “Primary care and multimorbidity in six Latin American and systems in low-and middle-income countries: a critical review Caribbean countries.” Revista Panamericana de Salud Pública of major primary care initiatives.” Social Science & Medicine 43: e8. 70(6): 904-911. 137. Mallawaarachchi, DS V., S.C. Wickremasinghe, L.C. Somatunga, 120. Kruk, M. E., Gage, A. D., Joseph, N. T., Danaei, G., García-Saisó, S., V.T.S.K Siriwardena, and N.S. Gunawardena (2016). “Healthy & Salomon, J. A. (2018). Mortality due to low-quality health sys- Lifestyle Centres: a service for screening noncommunicable tems in the universal health coverage era: a systematic analysis diseases through primary health-care institutions in Sri Lanka.” of amenable deaths in 137 countries. The Lancet, 392(10160), WHO South-East Asia journal of public health 5(2): 89-95. 2203-2212. 138. Maly, I. (2018). ESPN Thematic report on challenges on long-term 121. Kujawska, J. (2017) “The Efficiency of Post‐Communist Coun- care: Czech Republic. The European Social Policy Network, Euro- tries’ Health Systems.” Advances in Health Management 23: 93- pean Commission. 111. 139. Mamo, Y., E. Seid, S. Adams, A. Gardiner, and E. Parry (2007). “A 122. Lagomarsino, G., A. Garabrant, A. Adyas, R. Muga, and N. Otoo. primary healthcare approach to the management of chronic “Moving towards universal health coverage: health insurance disease in Ethiopia: an example for other countries.” Clinical reforms in nine developing countries in Africa and Asia.” The Medicine 7(3): 228-231. Lancet 380 (9845): 933-943. 140. Manes, F. (2016). “The huge burden of dementia in Latin Amer- 123. Lall, D., N. Engel, N. Devadasan, K. Horstman, and B. Criel. “Mod- ica.” The Lancet Neurology 15(1): 29. els of care for chronic conditions in low/middle-income coun- 141. Manton, KG. (1982). “Changing concepts of morbidity and tries: a ‘best fit’framework synthesis.” BMJ Global Health 3 (6): mortality in the elderly population”. Milbank Memorial Fund e001077. Quarterly/Health and Society 60: 183–244. 124. Lee, H.Y, Juhwan Oh, J. R. Moon, and S. V. Subramanian (2019). 142. Mathers, CD., Stevens, GA., Boerma, T., White, R. A., & Tobias, M. “Use of high-level health facilities and catastrophic expendi- I. (2015). “Causes of international increases in older age life ex- ture in Vietnam: can health insurance moderate this relation- pectancy.” The Lancet 385(9967): 540-548. ship?” BMC health services research 19(1): 318. 143. Matus-López, M. (2017). “Análisis prospectivo de los servicios 125. Levy BR, Slade MD, Chang ES, Kannoth S, Wang SY. Ageism Am- de cuidado a la dependencia en Uruguay”. Unpublished. In- plifies Cost and Prevalence of Health Conditions. Gerontologist. ter-American Development Bank. 2020 Jan 24;60(1):174-181. doi: 10.1093/geront/gny131. 144. Mayston, R., M. Guerra, Y. Huang, A.L. Sosa, R. Uwakwe, I. Acosta, 126. Llibre Rodriguez, JJ., C. P. Ferri, D. Acosta, M. Guerra, Y. Huang, K. P. Ezeah, S. Gallardo, V. Montes de Oca, H. Wang, M. Guerchet, S. Jacob, E. S. Krishnamoorthy, A. Salas, Sosa, A.L., I. Acosta and Z. Liu, M. Sanchez, P. Lloyd-Sherlock and M. J. Prince (2014). M.E. Dewey (2008). “Prevalence of dementia in Latin America, “Exploring the economic and social effects of care dependence India, and China: a population-based cross-sectional survey.” in later life: protocol for the 10/66 research group INDEP study.” The Lancet 372 (9637): 464-474. Springerplus 3 (1): 379. 127. Liu, J., G. Chen, X. Song, I. Chi and X. Zheng (2009). “Trends in 145. Maziak, W., R. Nakkash, R. Bahelah, A. Husseini, N. Fanous, and T. disability-free life expectancy among Chinese older adults.” Eissenberg. “Tobacco in the Arab world: old and new epidem- 336 CHAPTER 10 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E ics amidst policy paralysis.” Health Policy and Planning 29(6): “Renewing priority for dementia: Where do we stand?” OECD 784-794. Policy Brief. 146. McEniry, M. (2013). “Cross-national comparisons of health dis- 163. Organization for Economic Co-operation and Development parities among aging populations in Latin America, the Carib- and World Health Organization (OECD and WHO) (2019a). bean, Asia and Africa”. Population Studies Center, University of “Chile: A healthier tomorrow. Assessments and recommenda- Michigan. tions”. OECD Reviews of Public Health. 147. McEniry, M., and J. McDermott (2015). “Early-life conditions, 164. Organization for Economic Co-operation and Development and rapid demographic changes, and older adult health in the World Health Organization (OECD and WHO) (2019b). “Realising developing world.” Biodemography and social biology 61 (2): the full potential of primary health care”. OECD Policy Brief. 147-166. 165. Organization for Economic Co-operation and Development 148. Medellin, N., Ibarrarán, P. and Stampini, M. (2018). “Cuatro ele- and World Health Organization (OECD and WHO) (2020). mentos para diseñar un sistema de cuidados”. Banco Interam- Who Cares? Attracting and retaining care workers for the elderly. ericano de Desarrollo, Nota Técnica IDB-TN-1438. OECD Health Policy Studies, OECD Publishing, Paris, https://doi. 149. Miranda, J. J, S. Kinra, J. P. Casas, G. D. Smith, and S. Ebrahim org/10.1787/92c0ef68-en. (2008). “Non‐communicable diseases in low‐and middle‐in- 166. Organization for Economic Co-operation and Development come countries: context, determinants and health policy.” and World Bank (OECD and WB)(2020), Health at a Glance: Tropical Medicine & International Health 13 (10): 1225-1234. Latin America and the Caribbean 2020, OECD Publishing, Paris, 150. Molina, H., L. Sarmiento, N. Aranco and P. Jara (2020). “Envejec- https://doi.org/10.1787/6089164f-en. imiento y atención a la dependencia en Chile”. Nota Técnica 167. ONU Mujeres (2018). “El trabajo de cudados: una cuestión de IDB-TN-2004. Inter-American Development Bank. derechos humanos y políticas públicas”. ONU Mujeres, Entidad 151. Monteiro, L., and M. Paredes. “Arreglos de convivencia en la ve- de las Naciones Unidas para la Igualdad de Género y el Empod- jez en Uruguay: perfiles específicos para una política de cuida- eramiento de las Mujeres. dos.” Papeles de Población 22(87): 133-160. 168. Organización Panamericana de la Salud (OPS)(2017). “Día mun- 152. Montenegro Torres, F. (2013). “Costa Rica Case Study: Primary dial de la hipertensión: Conoce tus números”. Health Care Achievements and Challenges within the frame- 169. Oliveri, L. (2020). “Envejecimiento y atención a la dependencia work of the Social Health Insurance”. UNICO Studies Series 14, en Argentina”. Nota Técnica IDB-TN-2044. Inter-American Devel- The World Bank, Washington D.C. opment Bank. 153. Mor, V., T. Leone and A. Maresso eds. (2014). Regulating long- 170. Olshansky, S. (1991). “Trading off longer life for worsening term care quality: an international comparison. Cambridge Uni- health”. Journal of Aging and Health 3: 194–216. versity Press. 171. Omran, AR. 1971. “The epidemiologic transition. A theory of 154. Morikawa, M. (2014). “Towards community-based integrated the epidemiology of population change”. The Milbank Memo- care: trends and issues in Japan’s long-term care policy.” Inter- rial Fund Quarterly 49(4): 509-538. national Journal of Integrated Care 14: e005. 172. Palloni, A., G. Pinto-Aguirre, and M. Peláez. (2002). “Demo- 155. Mounier-Jack, S., S. H. Mayhew, and N. Mays. “Integrated care: graphic and health conditions of ageing in Latin America and learning between high-income, and low-and middle-income the Caribbean”. International Journal of Epidemiology 31(4): country health systems.” Health policy and planning 32(4): iv6-iv12. 762-771. 156. Moore, E. (2001). “Preparing for an aging world: the case for 173. Parker, MG. and M. Thorslund. (2007). “Health Trends in the El- cross‐national research”. Panel on a Research Agenda and New derly Population: Getting Better and Getting Worse”. The Geron- Data for an Aging World, National Research Council. Washing- tologist 47 (2):150–158. ton, DC, National Academy Press. 174. PHCPI Brazil (Primary Healthcare Performance Initiative Brazil). 157. Ministerio de Salud Pública (MSP) Uruguay. Plan Nacional de n/d. “Brazil: A community-based approach to comprehensive Salud Mental. 2020-2027. primary care”. 158. Nitrini, R., C. Bottino, C. Albala, N. Santos Custodio Capunay, C. 175. PHCPI Costa Rica (Primary Healthcare Performance Initiative Ketzoian, J. J. Llibre Rodriguez, G. E. Maestre, A.T. A. Ramos-Cer- Costa Rica). n/d. “Costa Rica: Universal health coverage and queira, and P. Caramelli (2009). “Prevalence of dementia in Lat- community-based health teams create effective care”. in America: a collaborative study of population-based cohorts.” 176. PHCPI Estonia (Primary Healthcare Performance Initiative Esto- International Psychogeriatrics 21(4) : 622-630. nia). n/d. “Estonia: Establishing family medicine as a specialty 159. Olivier, J., Tsimpo, C., Gemignani, R., Shojo, M., Coulombe, H., to strengthen primary health care”. Dimmock, F., ... & Wodon, Q. (2015). Understanding the roles of 177. Pitheckoff, N. (2017). “Aging in the Republic of Bulgaria.” The faith-based health-care providers in Africa: review of the evi- Gerontologist 57 (5): 809-815. dence with a focus on magnitude, reach, cost, and satisfaction. 178. Pop, L. (2018). ESPN Thematic report on challenges on long-term The Lancet, 386(10005), 1765-1775. care: Romania. The European Social Policy Network, European 160. Organization for Economic Co-operation and Development Commission. (OECD) (2017), Preventing Ageing Unequally, OECD Publishing, 179. Praat-Ahi, G. and M. Masso (2018). ESPN Thematic report on Paris, https://doi.org/10.1787/9789264279087-en. challenges on long-term care: Estonia. The European Social Poli- 161. Organization for Economic Co-operation and Development cy Network, European Commission. and World Health Organization (OECD and WHO) (2018a). 180. Prentice, A. M. (2006). “The emerging epidemic of obesity in Health at a Glance: Asia/Pacific 2018: Measuring Progress to- developing countries.” International Journal of Epidemiology wards Universal Health Coverage. OECD Publishing, Paris. 35(1):93-99. 162. Organization for Economic Co-operation and Development 181. Prince, M., H. Brodaty, R. Uwakwe, D. Acosta, C.P. Ferri, M. Guer- and World Health Organization (OECD and WHO) (2018b). ra, Y. Huang, K.S. Jacob, J.J. Llibre Rodriguez, A. Salas, and A.L. COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 10 337 Sosa (2012). “Strain and its correlates among carers of people co-residence of parents and children among older Kuwaiti with dementia in low‐income and middle‐income countries. men and women: What are the significant correlates?” Journal A 10/66 Dementia Research Group population‐based survey.” of Cross-Cultural Gerontology 26(2): 157-174. International journal of geriatric psychiatry 27 (7): 670-682. 199. Shekar, M. and Popkin, B. eds. 2020. Obesity: Health and Eco- 182. Prince, M. J., Wu, F., Guo, Y., Robledo, L. M. G., O’Donnell, M., nomic Consequences of an Impending Global Challenge. Hu- Sullivan, R., and Yusuf, S. (2015). “The burden of disease in old- man Development Perspectives series. Washington, DC: World er people and implications for health policy and practice”. The Bank. doi:10.1596/978-1-4648-1491-4. Lancet, 385(9967): 549-562. 200. Sibai, A. M., Semaan, A., Tabbara, J., and Rizk, A. (2017). “Ageing 183. Rajevska, F. (2018). ESPN Thematic report on challenges on long- and health in the Arab region: Challenges, opportunities and term care: Latvia. The European Social Policy Network, Europe- the way forward”. Population Horizons 14(2): 73-84. an Commission. 201. Singh, P.K., A. Cassels, and P. Travis (2018). “Primary health care 184. Ramli, A. S., and S. W. Taher. (2008). “Managing chronic diseas- at forty: reflections from south-east Asia.” World Health Organi- es in the Malaysian primary health care–a need for change.” zation. Regional Office for South-East Asia. Malaysian family physician: the official journal of the Academy of 202. Sloan, F. A., J. Ostermann, D. H. Taylor Jr, C. Conover, and G. Pi- Family Physicians of Malaysia 3 (1): 7. cone (2004). The price of smoking. MIT press, 2004. 185. Rampal, L., Rampal, S., Khor, G.L., Zain, A.M., Ooyub, S.B., Rah- 203. Smith, J.P, and Majmundar, M. eds. (2012). Aging in Asia: Find- mat, R.B., Ghani, S.N. and Krishnan, J., (2007). “A national study ings from new and emerging data initiatives. Committee on on the prevalence of obesity among 16,127 Malaysians”. Asia Population, Division of Behavioral and Social Sciences and Ed- Pacific journal of clinical nutrition, 16(3). ucation, National Research Council. Washington, D.C. 186. Rechel, B., Doyle, Y., Grundy, E. and McKee, M. (2009). How can 204. Smith, O. and S.N. Nguyen. (2013). Getting better: improving health health systems respond to population ageing. Copenhagen: system outcomes in Europe and Central Asia. The World Bank. WHO Regional Office for Europe. World Health Organization. 205. SNIC (Sistema Nacional Integrado de Cuidados) (2015). Plan 187. Rechel, B., M. Ahmedov, B. Akkazieva, A. Katsaga, G. Khod- Nacional de Cuidados 2016-2020. Ministerio de Desarrollo So- jamurodov, and M. McKee (2012). “Lessons from two decades cial, Uruguay. of health reform in Central Asia.” Health Policy and Planning 206. SNIC (Sistema Nacional Integrado de Cuidados) (2018). Informe 27(4): 281-287. Anual 2017. Ministerio de Desarrollo Social, Uruguay. 188. Romas L., A. Poviliunas and L. Zalimiene (2018). ESPN Thematic 207. SNIC (Sistema Nacional Integrado de Cuidados) (2019). Informe report on challenges on long-term care: Lithuania. The European Anual 2018. Ministerio de Desarrollo Social, Uruguay. Social Policy Network, European Commission. 208. SNIC (Sistema Nacional Integrado de Cuidados) (2020). Cuida- 189. Rodda, J., Z. Walker, and J. Carter (2011). “Depression in older dos rinde cuentas: informe mensual, marzo 2020. Ministerio de adults.” BMj 343. Desarrollo Social, Uruguay. 190. Rosen, B., H. Samuel, and S. Merkur (2009) “Israel Health system 209. Somanathan, A., A. Tandon, H. Lan Dao, K. L. Hurt, and H L. review”. Health Systems in Transition 17(6): 253-253. Fuenzalida-Puelma. (2014). Moving toward universal coverage 191. Sadana, R., A. Soucat, and J. Beard (2018) “Universal health cov- of social health insurance in Vietnam: assessment and options. erage must include older people.” Bulletin of the World Health The World Bank, Washington, D.C. Organization 96 (1): 2. 210. Somkotra, T., and L. P. Lagrada (2009). “Which Households 192. Salhab, N., J. Yartey, S. Rawaf, P. Musgrove, M. Claeson, I. Kickbus- Are At Risk Of Catastrophic Health Spending: Experience In ch, C. Jenkins, D. Robalino, A. Downing, B. Sadrizadeh, and A.M. Thailand After Universal Coverage: Exploring the reasons why Pierre-Louis (2004). Public health in the Middle East and North some households still incur high levels of spending—even un- Africa: meeting the challenges of the twenty-first century. WBI der universal coverage—can help policymakers devise solu- Learning Resources Series Washington, D.C.: World Bank Group. tions.” Health Affairs 28 (1): w467-w478. 193. Samb, B., N. Desai, S. Nishtar, S. Mendis, H. Bekedam, A. Wright, 211. Souza, D. K. de, & Peixoto, S. V. (2017). “Estudo descritivo J. Hsu, A. Martiniuk, F. Celletti, K. Patel and F. Adshead (2010). da evolução dos gastos com internações hospitalares por “Prevention and management of chronic disease: a litmus test condições sensíveis à atenção primária no Brasil, 2000-2013.” for health-systems strengthening in low-income and mid- Epidemiologia e Serviços de Saúde, 26(2), 285-294. dle-income countries.” The Lancet 376 (9754): 1785-1797. 212. Spasova, S., R. Baeten, S. Coster, D. Ghailani, R. Peña-Casas and 194. Sandoya, E. (2016). “Diabetes y enfermedad cardiovascular en B. Vanhercke (2018). Challenges in LTC in Europe: a study of na- Uruguay”. Revista Uruguaya de Cardiología 31: 505-514. tional policies. European Commission. 195. Sassi, F., A. Belloni, A. J. Mirelman, M. Suhrcke, A. Thomas, N. 213. Sowa-Kofta, A. (2018). ESPN Thematic report on challenges on Salti, S. Vellakkal, C. Visaruthvong, B. M. Popkin, and R. Nugent long-term care: Poland. The European Social Policy Network, (2018): “Equity impacts of price policies to promote healthy be- European Commission. haviours.” The Lancet 391(10134): 2059-2070. 214. Stampini, M., Olivieri, L., Ibarrarán, P., Londoño, D., Rhee, H. J., 196. Scazufca, M., P. R. Menezes, H. P. Vallada, A. L. Crepaldi, M. Pa- James, G. M. (2022). “Working les to take care of parents? Labor stor-Valero, L. MS Coutinho, V. D. Di Rienzo, and O. P. Almeida market efects of family long-term care in Latin America”. Working (2008). “High prevalence of dementia among older adults from Paper Series IDB-WP-1105. Inter-American Development Bank. poor socioeconomic backgrounds in Sao Paulo, Brazil.” Interna- 215. Stokic, L.P. and J. Bajec (2018). ESPN Thematic report on chal- tional Psychogeriatrics 20(2): 394-405. lenges on long-term care: Serbia. The European Social Policy 197. Scheil-Adlung, X. (2015). Long-term care protection for older per- Network, European Commission. sons: a review of coverage deficits in 46 countries. UNECE Working 216. Storr, Carla L., Hui Cheng, Jordi Alonso, Matthias Angermeyer, Group on Ageing, Geneva, Switzerland, ILO. Ronny Bruffaerts, Giovanni De Girolamo, Ron De Graaf et al. 198. Shah, N. M., H. E. Badr, K. Yount, and M. A. Shah. “Decline in (2010). “Smoking estimates from around the world: data from 338 CHAPTER 10 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E the first 17 participating countries in the World Mental Health 234. UN ECA (United Nations Economic Commission for Africa) Survey Consortium.” Tobacco control 19(1): 65-74. (2017). The Third Review and Appraisal Cycle of the implementa- 217. Stubbs, P. and S. Zrinščak (2018). ESPN Thematic report on chal- tion of the Madrid International Plan of Action on Ageing in Africa lenges on long-term care: Croatia. The European Social Policy for the period 2012 – 2017. United Nations. Network, European Commission. 235. UN ECE (United Nations Economic Commission for Europe) 218. Suwannaphant, K., W. Laohasiriwong, N. Puttanapong, J. (2017). Synthesis Report on the implementation of the Madrid Saengsuwan, and T. Phajan. “Association between socioeco- International Plan of Action on Ageing in the ECE region between nomic status and diabetes mellitus: The National Socioeco- 2012 and 2017. United Nations, Working Group on Ageing. nomics Survey, 2010 and 2012.” Journal of Clinical and Diagnos- 236. United Nations Economic Commission for Latin American and tic Research 11 (7):LC18-LC22. the Caribbean (UN ECLAC) (2017a). Caribbean Synthesis Report 219. Tejativaddhana, P., D. Briggs, O. Singhadej, and R. Hinoguin on the implementation of the Madrid International Plan of Action (2018). “Developing primary health care in Thailand.” Public Ad- on Ageing and the San Jose Charter on Rights of Older People in ministration and Policy. Latin America and the Caribbean. United Nations. 220. Tham, T.Y., T.L. Tran, S. Prueksaritanond, J. S. Isidro, S. Setia, and V. 237. United Nations Economic Commission for Latin American and Welluppillai (2018). “Integrated health care systems in Asia: an the Caribbean (UN ECLAC) (2017b). Fourth Regional Intergov- urgent necessity.” Clinical interventions in aging 13: 2527-2538. ernmental Conference on Ageing and the Rights of Older Persons 221. Thumé, E., L.A. Facchini, G. Wyshak, and P. Campbell (2011). “The in Latin America and the Caribbean. United Nations. utilization of home care by the elderly in Brazil’s primary health 238. United Nations Economic Commission and Social Commission care system.” American journal of public health 101(5): 868-874. for Asia and the Pacific (UN ESCAP ) (2017). Government actions 222. Tomita, K. (2017). “Japanese healthcare at a crossroads (1): towards the implementation of the Madrid International Plan of toward integrated community care”. The Tokyo Foundation for Action on Ageing, 2002: Achievements and remaining challenges. Policy Research. United Nations. 223. Tonoyan, T., and L. Muradyan (2012). “Health inequalities in Ar- 239. United Nations Economic and Social Commission for Western menia-analysis of survey results.” International Journal for Equity Asia (UN ESCWA) (2017). Third review and appraisal of the Ma- in Health 11(1): 32. drid International Plan of Action on Ageing. United Nations. 224. Trieu, K., B. Neal, C. Hawkes, E. Dunford, N. Campbell, R. Ro- 240. United Nations Population Fund Well Woman Programme (UN- driguez-Fernandez, B. Legetic, L. McLaren, A. Barberio, and J. FPA WWP) (2019). National Strategic Plan 2019-2023. Family Webster (2015). “Salt reduction initiatives around the world–a Health Bureau, Ministry of Health, Sri-Lanka. systematic review of progress towards the global target.” PloS 241. United Nations General Assembly (UNGA). 2019. Political ONE 10(7): e0130247. Declaration of the High-level Meeting on Universal Health 225. Twigg, J., Bardasi, E., Garcia, G. (2021). Covid-19 has exposed Coverage “Universal health coverage: moving together to the fragilities of aging countries. World Bank blog. Available build a healthier world”, September 2019 https://www. at: https://ieg.worldbankgroup.org/blog/covid-19-has-ex- un.org/pga/73/wp-content/uploads/sites/53/2019/07/FI- posed-fragilities-aging-countries NAL-draft-UHC-Political-Declaration.pdf. 226. United Nations (UN). 2002. Political Declaration and Madrid In- 242. Van der Heide, I, S. Snoeijs, M. Gabriella Melchiorre, S. Quattri- ternational Plan of Action on Ageing. Second World Assembly ni, W. Boerma, F. Schellevis, and M. Rijken. (2015). “Innovating on Ageing, Madrid, Spain. care for people with multiple chronic conditions in Europe: An 227. United Nations (UN). 2015a. World population ageing report overview.” ICARE4EU. Health Programme of the European Union. 2015. New York: Department of Economic and Social Affairs, 243. Van Minh, H., N. S. Pocock, N. Chaiyakunapruk, C. Chhorvann, H.A. United Nations 2015. Duc, P. Hanvoravongchai, J. Lim, D.E. Lucero-Prisno III, N. Ng, N. 228. United Nations (UN). 2017a. World population ageing 2017: Phaholyothin and A. Phonvisay (2014). “Progress toward universal Highlights. New York: Department of Economic and Social Af- health coverage in ASEAN.” Global Health Action 7 (1): 25856. fairs, United Nations. 244. Varona, R., T. Saito, M. Takahashi, and I. Kai (2007). “Caregiving in 229. United Nations (UN). 2017b. Population ageing and policy op- the Philippines: A quantitative survey on adult-child caregivers’ tions in the Arab region. United Nations Population Fund, Arab perceptions of burden, stressors, and social support.” Archives of States Regional Office. Gerontology and Geriatrics 45(1): 27-41. 230. United Nations(UN). 2017c. Living arrangements of older per- 245. Vathesatogkit, P, G. D. Batty, and M. Woodward (2014). “Socio- sons: a report on an expanded international dataset. New York: economic disadvantage and disease-specific mortality in Asia: Department of Economic and Social Affairs. Population Divi- systematic review with meta-analysis of population-based sion, United Nations. cohort studies.” J Epidemiol Community Health 68 (4): 375-383. 231. United Nations (UN). 2017d. Una Mirada sobre el envejec- 246. Verity, R., L.C. Okell, I. Dorigatti, P. Winskill, C. Whittaker, N. Imai imiento ¿Dónde están varios países latinoamericanos a 15 et. al. (2020). “Estimates of the Severity of Coronavirus Disease años del Plan de Acción Internacional de Madrid? Population 2019: A Model-Based Analysis.” The Lancet Infectious Diseases Division, United Nations. 20(6): 669-677. 232. United Nations (UN). 2019. World Population Ageing 2019: High- 247. Villalobos, P. 2018. Long-term care systems as social security: lights. New York: Department of Economic and Social Affairs, the case of Chile. Health Policy and Planning, Volume 33, Issue United Nations. 9, November 2018, Pages 1018–1025, https://doi.org/10.1093/ 233. United Nations (UN) and HelpAge International (2012). Overview heapol/czy083 of available policies and legislation, data and research, and institu- 248. Villalobos, P. “Is aging a problem? Dependency, long-term care, tional arrangements relating to older persons - Progress since Ma- and public policies in Chile.” Revista Panamericana de Salud drid. United Nations Population Fund and HelpAge International. Pública 42: e168. COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 10 339 249. Villalobos, P., Oliveira, D., Stampini, M. (2022). “Estimación de global tobacco epidemic 2017: monitoring tobacco use and pre- las necesidades de recursos humanos para la atención de vention policies. World Health Organization. personas mayores con dependencia de cuidados en América 269. World Health Organization (WHO) (‎ 2017c)‎. Towards long- Latina y el Caribe”. Nota Técnica IDB-TN-02556. Banco Interam- term care systems in sub-Saharan Africa. WHO Series on Long- ericano de Desarrollo. Term Care. 250. Wang, B., R. Li, Z. Lu, and Y. Huang (2020). “Does comorbidity 270. World Health Organization (WHO) (2018a). The state of food se- increase the risk of patients with COVID-19: evidence from me- curity and nutrition in the world 2018: building climate resilience ta-analysis.” Aging (Albany NY) 12(7): 6049-6057. for food security and nutrition. Food & Agriculture Organization. 251. Wilkinson, R. G., and M. Marmot, eds. (2003). Social determi- 271. World Health Organization (WHO) (2018b). Public Spending on nants of health: the solid facts. World Health Organization. Health: A Closer Look at Global Trends. World Health Organization. 252. Williamson, C. (2015). “Policy mapping on ageing in Asia and 272. World Health Organization (WHO) (2018c). Country coopera- the Pacific: Analytical report”. HelpAge International, East Asia/ tion strategy at a glance: Seychelles. Pacific Regional Office, Chiang Mai, Thailand. 273. World Health Organization (WHO) (2019). Global report on 253. Woldie, M., G.T. Feyissa, B. Admasu, K. Hassen, K. Mitchell, S. trends in prevalence of tobacco use 2000-2025, third edition. Ge- Mayhew, M. McKee, and D. Balabanova (2018).”Community neva: World Health Organization. health volunteers could help improve access to and use of es- 274. World Health Organization (WHO) (2020). “Obesity and over- sential health services by communities in LMICs: an umbrella weight – Key facts”. Retrieved from: https://www.who.int/news- review.” Health Policy and Planning 33(10): 1128-1143. room/fact-sheets/detail/obesity-and-overweight. 254. Woolf, AD., J. Erwin and L. March (2012). “The need to address 275. Wu, F., Guo, Y., Chatterji, S., Zheng, Y., Naidoo, N., Jiang, Y., ... and the burden of musculoskeletal conditions.” Best practice & re- Manrique-Espinoza, B. (2015). “Common risk factors for chron- search Clinical rheumatology 26(2): 183-224. ic non-communicable diseases among older adults in China, 255. Workie, N.W., E. Shroff, A. S. Yazbeck, S.N, Nguyen, and H. Kara- Ghana, Mexico, India, Russia and South Africa: the study on magi. “Who Needs Big Health Sector Reforms Anyway? Sey- global AGEing and adult health (SAGE) wave 1”. BMC public chelles’ Road to UHC Provides Lessons for Sub-Saharan Africa health, 15(1), 1-13. and Island Nations.” Health Systems & Reform 4(4): 362-371. 276. Xu, F., X. M. Yin, M. Zhang, E. Leslie, R. Ware and N. Owen (2006). 256. World Bank. 2015. “World Bank Group Gender Strategy (FY16– “Family average income and diagnosed type 2 diabetes in ur- 23): Gender Equality, Poverty Reduction and Inclusive Growth.” ban and rural residents in regional mainland China”. Diabetic World Bank, Washington, DC. Medicine 23(11): 1239-1246. 257. World Bank. 2021. World Bank Support to Aging Countries: 277. Yamaguchi, N., N. Pilnik, J. De La Garza, L. Ashton, A. Garcia, An Independent Evaluation. Independent Evaluation Group. E. Bianco, and G. Kevorkof. “Tobacco Control Policies in Latin Washington, DC: World Bank. America”. Journal of Thoracic Oncology 12(1): S56-S57. 258. World Bank and International Monetary Fund (2016). Develop- 278. Yang Z., Hall A. (2008). “The financial burden of overweight and ment goals in an era of demographic change. Global Monitoring obesity among elderly Americans: the dynamics of weight, Report 2015/2016. longevity, and health care cost”. Health Services Research 259. World Bank Independent Evaluation Group (2018). Project 43(3):849–868. performance assessment report. Oriental Republic of Uruguay. 279. Yiengprugsawan, V., J. Healy, H. Kendig, M. Neelamegam, P. Non-communicable diseases prevention project (IBRD-74860). Karunapema, and V. Kasemsup. “Reorienting health services 260. World Cancer Research Fund International (2018). Building to people with chronic health conditions: diabetes and stroke momentum: lessons on implementing a robust sugar sweetened Services in Malaysia, Sri Lanka and Thailand.” (2017). Health Sys- beverage tax. tems & Reform 3(3): 171-181. 261. World Economic Forum and World Health Organization (WEF 280. Yon, Y., C. R. Mikton, Z. D. Gassoumis, and K. H. Wilber. “Elder and WHO) (2011). From Burden to “Best Buys”: Reducing the abuse prevalence in community settings: a systematic review Economic Impact of Chronic Diseases in Low- and Middle-In- and meta-analysis.” The Lancet Global Health 5(2): e147-e156. come Countries. 281. Yon, Y., M. Ramiro-Gonzalez, C. R. Mikton, M. Huber, and D. 262. World Health Organization (WHO) (2008). Preventing Chronic Sethi. “The prevalence of elder abuse in institutional settings: a Diseases: A Vital Investment: WHO Global Report 2008. Geneva: systematic review and meta-analysis.” European Journal of Pub- World Health Organization. lic Health 29(1): 58-67. 263. World Health Organization (WHO) (2009). “Armenians struggle 282. Zaidi, A. (2016). “Sustainable Development Goals have put for health care and medicines”. Bulletin of the World Health Or- ageing back onto the agenda”. Global AgeWatch Index Blogs, 08 ganization 87(7): 485-564. June 2016. 264. World Health Organization (WHO) (2011a). Global Health and 283. Zaidi, S., R. Bennet and R.C. Summer (2017). The Madrid In- Aging. WHO; US National Institute of Aging. ternational Plan of Action on Ageing: Where is Eastern Europe 265. World Health Organization (WHO) (2011b). Global status report and Central Asia region fifteen years later? United Nations Pop- on noncommunicable diseases 2010. WHO; US National Insti- ulation Fund. tute of Aging. 284. Zaidi, S., P. Saligram, S. Ahmed, E. Sonderp, and K. Sheikh. 266. World Health Organization (WHO) (2015). World report on age- (2017). “Expanding access to healthcare in South Asia.” bmj 357 ing and health. World Health Organization. (j1645): 1-4. 267. World Health Organization (WHO) (‎ 2017a)‎. Tackling NCDs: ‘best 285. Zaidi, A. (2018). “Implementing the Madrid Plan of Action on buys’ and other recommended interventions for the prevention and Ageing: What have we learned? And, where do we go from control of noncommunicable diseases. World Health Organization. here?”. Human Development Reports. United Nations Develop- 268. World Health Organization (WHO) (‎ 2017b)‎ . WHO Report on the ment Programme. 340 CHAPTER 10 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CONSULTED SURVEYS AND DATABASES i. Encuesta de Hogares de Propósitos Múltiples de El Salvador viii. GBD (Global Burden of Disease and Collaborators Results (2018). Retrieved from: https://onec.bcr.gob.sv/encues- Tool) (2019) Retrieved from: http://ghdx.healthdata.org/ ta-de-hogares-de-propositos-multiples-ehpm/ gbd-results-tool ii. Encuesta Longitudinal de Protección Social (ELPS), Chile ix. United Nations. Department of Economic and Social Affairs. (2015). Available upon request. World Population Prospects (UNWPP) (2019). Retrieved from: https://population.un.org/wpp2019/ iii. Encuesta Longitudinal de Protección Social (ELPS), El Salvador (2013). Retrieved from: https://onec.bcr.gob.sv/metadatos/ x. World Health Organization (WHO) (2004). “A glossary of terms index.php/catalog/12 for community health care and services for older persons”. WHO Centre for Health Development, Ageing and Health iv. Encuesta Longitudinal de Protección Social (ELPS), Paraguay Technical Report, 5 (2013). Retrieved from: https://www.stp.gov.py/v1/encues- ta-longitudinal-de-proteccion-social-elps/ xi. World Health Organization Global Health Estimates. Global Burden of Disease (2018). Retrieved from: v. Encuesta Longitudinal de Protección Social (ELPS), Uruguay https://www.who.int/healthinfo/global_burden_disease/en/ (2013). Retrieved from: https://www.elps.org.uy/ xii. World Health Organization Global Health Observatory Data vi. Encuesta de Salud y Bienestar del Adulto Mayor en Perú Repository (2020). Retrieved from: https://www.who.int/data/ (2012). Retrieved from: http://webinei.inei.gob.pe/anda_inei/ gho/indicator-metadata-registry/imr-details/4680 index.php/catalog/584 xiii. World Bank Open Data website. Retrieved from: vii. United Nations World Population Prospects (2019). Retrieved https://data.worldbank.org/ from: https://population.un.org/wpp/ 11 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E C H A P T E R 11 341 Demand for and Supply of Long- Term Care for Older Persons in Low- and Middle-Income Countries Elena Glinskaya a, Xiaohui Hou b, Zhanlian Feng c, Marco Angrisani d,e, Guadalupe Suarez c, Jigyasa Sharma b, Drystan Phillips d, Jenny Wilkens d, Jinkook Lee d,e, Yeeun Lee Yoo d, Samuel Lau d, Hae Yeun Park d, and Yizhou Chen d a Social Protection and Labor, West Africa Region, World Bank b Health, Nutrition, and Population Global Practice, World Bank c Research Triangle Institute d Center for Economic and Social Research, University of Southern California e Department of Economics, University of Southern California 342 C H A P T E R 11 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E INTRODUCTION In virtually all countries across the world, the population is aging, and consequently, both the number and pro- portion of older people are growing. Notably, the aging process is accelerating at a more rapid pace in low- and middle-income countries (LMICs) than in high-income countries (HICs). In 2020, older people aged 65 and over accounted for roughly 8 percent of the population in middle-income countries. This proportion is project- ed to double in 2050 to over 16 percent. Even in low-income countries, which typically have much younger populations, the proportion of older people aged 65 and over will jump from about 3 percent in 2020 to over 5 percent in 2050 (United Nations 2019). In comparison, the proportion of individuals aged 65 and over in HICs will increase more gradually, from 18 percent in 2020 to 27 percent by 2050. In 2020, there were an estimated 727 million people aged 65 and over worldwide, of whom over two-thirds (68 percent) lived in LMICs (United Nations 2019). Projections indicate that by 2050 the global population aged 65 and over will more than double, to top 1.5 billion and more than three-quarters (77 percent) of older individuals will live in LMICs. While the population is growing older and life expec- rently unable to meet all of their basic daily needs tancy is increasing steadily across LMICs, the burden due to functional limitations (World Health Orga- of non-communicable diseases (NCD) and age-re- nization 2020b). The majority of these older people lated disabilities are also on the rise (Kazibwe, Tran, are from LMICs, where the challenge for meeting and Annerstedt 2021). Recent research on disability their care needs is particularly acute, due to the ab- trends among older adults in 23 LMICs forecasts that sence of robust healthcare and LTC systems. the prevalence rate of people living with severe activ- The LTC research literature to date has been ity limitations will change very little in the next 30 dominated by studies in HICs with mature LTC years; however, the absolute number of persons with systems. At the same time, in many LMICs, formal activity limitations will increase as the sheer size of (paid) LTC services are emerging in the public sector older population swells (Weber and Scherbov 2020). and especially in the private market, but they are in Thus, despite anticipated gains in life expectancy, the early stages of development and not well document- number of years in later life in ill health and with ed. The lack of LTC research and empirical knowl- disability is expected to increase. Meanwhile, there edge in LMICs is an impediment to evidence-based is growing concern that family-based elder care policy planning, formulation, and evaluation in these alone—traditionally the mainstay of old age support countries (Lloyd-Sherlock 2014, Aboderin 2019). and long-term care (LTC)—is no longer sufficient In this chapter, we provide an overview of the to meet the escalating demand for LTC services for demand for LTC and describe the current landscape rapidly aging populations across LMICs (Feng 2019). of supply of LTC services for old persons across six Yet, in most countries, LTC has not received middle-income countries (MICs), namely China, high enough priority on the national policy agenda India, Malaysia, Mexico, Poland, and Romania. The (Scheil-Adlung 2015). To help all countries address demand-side assessment is based on the analysis of the aging challenges, the United Nations launched individual-level datasets that are part of the Inter- the Decade of Healthy Ageing 2021−2030, to pro- national Network of Health and Retirement Studies. mote healthy aging and improve the lives of older The supply-side evaluation is a synthesis of findings people (World Health Organization 2020a). It iden- from a review of relevant literature, both published tifies “access to long-term care for people who need and grey, identified through a series of web searches it” as a priority area for action, which is crucial and databases as well as national and international for older people with declining physical or mental policy documents published in English in the past capacity to optimize functional ability and live in- 22 years (2000-2022).1 The review covers selected dependently. Current estimates indicate that more LMICSs and MICs in Asia (where we have most of than 142 million older people, or 14 percent of the data), Central and South America, Southern Eu- world population aged 60 years and over, are cur- rope and Middle East, and Sub-Saharan Africa. 1   This review also draws from the authors’ earlier work, including (i) “Understanding the “state of play” of long-term care provision in low- and middle-income countries” , International Social Security Review Vol. 75, No. 3-4(2022), and (ii) “Primary Health Care Centered Integrated Care for an Aging Population”, World Bank 2022. COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E C H A P T E R 11 343 What is LTC for older persons? According to the definition used by the European people is discussed mainly in the context of integrat- Commission and the Social Protection Committee, ing the full range of services in a continuum of care for LTC refers to “a range of healthcare and social care ser- older adults with LTC needs. Furthermore, although vices and assistance, for people who, as a result of mental LTC encompasses both formal and informal care, and and/or physical frailty and/or disability and/or old age, informal care is by far the more prevalent modality over an extended period of time depend on help with use in virtually all countries across the world, our sup- daily living activities, and/or need some permanent ply-side discussion focuses on formal LTC provision, nursing care” (European Commission 2021). As such, which is typically provided by paid carers, in contrast LTC covers both medical care and non-medical ser- to informal LTC that is typically provided by unpaid vices and support for older adults. In this chapter, we family members, friends or neighbors. The focus on focus primarily on the latter, namely, the non-medical the formal care is intentional: the objective of this long-term services and support to meet the every-day chapter is to raise awareness about the limitations that care needs of older adults with physical and/or cog- LMICs face in assuring that families have choices with nitive impairments. Medical or health care for older respect to LTC for their older members. A DEMAND SIDE VIEW - CURRENT LTC LANDSCAPE IN MICS In this section, we analyze the demand for long-term care (LTC) among older adults across six middle-income countries – China, India, Malaysia, Mexico, Poland and Romania. The analysis is based on individual-level data comparable across these countries which are collected and maintained by the International Network of Health and Retirement Studies, see Box 11.1. Long-term care demand is measured by the prevalence of self-re- ported limitations in activities of daily living (ADLs) and instrumental activities of daily living (IADLs),which are the essential activities that individuals should perform without assistance to lead independent lives. ADLs and IALDs are commonly used to assess an individual’s functional status and quality of life (Clark et al. 2004; Edemekong et al. 2022). Receipt of long-term care is measured by self-reported formal and informal care utili- zation. The analysis considers differences across several socio-demographic characteristics that influence both the need for, and the receipt of, long-term care. Specifically, we consider two age brackets (e.g., 65 to 79 and 80 and above), gender, urban and rural status, education categories (those with no education, those with primary education, and those with secondary education or more) and household income (low, middle, and high income tertiles). We also separate individuals into three groups according to their living situation: those living alone, those living with one other person, and those living with two or more other people. BOX 11.1 Data used for the demand side analysis The demand for long-term care (LTC) among older adults across six middle-income countries – China, India, Malaysia, Mexico, Poland, and Romania – is based on individual-level data comparable across these six countries which are collected and main- tained by the International Network of Health and Retirement Studies. For China, we use Wave 4 (2018) of The China Health and Retirement Survey (CHARLS), which includes respondents aged 45 years and older and their spouses, regardless of age. For India, we use Wave 1 (2017-2019) of the Longitudinal Aging Study in India (LASI), which interviews respondents aged 45 years and older and their spouses, regardless of age. For Malaysia, we use Wave 1 (2018) of the Malaysia Ageing and Retirement Sur- vey (MARS), which includes the three oldest respondents in the household aged at least 40 years and older. For Mexico, we use Wave 5 (2018) of the Mexican Health and Aging Study (MHAS), which includes respondents aged 50 years and older and their spouses regardless of age. For Poland and Romania, we use Wave 8 (2019-2020) of the Survey of Health, Ageing and Retirement in Europe (SHARE), which interviews respondents aged 50 years and older and their spouses, regardless of age. We limit our analyses to respondents aged 65 years and older, and use survey sample weights, when available, to estimate the prevalence of long-term care needs and receipt at the population level. As such, our estimates should be interpreted as popula- tion-level estimates for a particular year and country, and do not reflect prevalence accurately at another time or in another country. 344 C H A P T E R 11 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E Limitations in activities of daily living (ADLs) This section documents the prevalence of difficul- harmonized variable is coded as having difficulty. ties with activities of daily living (ADLs) separately MARS asks about receiving help with ADLs in for six countries. The five ADLs considered are: (i) this form: “In the last week, did you need any assis- dressing, (ii) bathing, (iii) eating, (iv) getting in and tance in performing the following activities (exclud- out of bed, and (v) using the toilet. Difficulties with ing any difficulty that you expect to last less than 3 each of these five ADLs have been elicited by each months)?” If the respondent reports being able to do country-specific health and retirement study and it all by themselves, then the harmonized variable is included in the harmonized datasets. coded as not needing help. If, however, the respon- It is worth noting that a few cross-study dif- dent reports that they sometimes need some help, ferences exist in the survey questions used to cap- always need some help, or always need help, then the ture difficulties with activities of daily living. LASI, harmonized variable is coded as needing help. MHAS, SHARE, and CHARLS ask about difficulties The difference in question wording between with ADLs in a similar way: “Because of a physical MARS and the other studies appears to be substan- or health problem do you have difficulty with any tial. MARS provides a measure of whether an in- of these activities? Exclude any activities you expect dividual needs help with ADLs, whereas the other to last less than three months.” In LASI, SHARE, studies provide a measure of whether an individual and MHAS, respondents answer “yes” or “no” to has difficulties with ADLs. Needing help with and this question, separately for each ADL. MHAS re- having difficulty with an activity are conceptually spondents can also answer with a “can’t do,” which different. This prevents comparability between Ma- is coded as having difficulty in the harmonized laysia and other countries. Because of that, we will variable, or “doesn’t do,” which is coded as having comment about Malaysia’s results separately, with- difficulty only if respondents report receiving help out attempting a comparison with the other coun- for that activity. For each activity in CHARLS, if the tries. We should also point out that, in our analyses, respondent reports they have no difficulty, then the we exclude individuals from MHAS who answered harmonized variable is coded as not having difficul- that they don’t do the activity and don’t receive help. ty. On the other hand, if the respondent reports that Since LASI, SHARE, and CHARLS did not give re- they have difficulty but can still do it, they have dif- spondents this option, it is not possible to identify ficulty and need help, or they cannot do it, then the the same group of people in these other studies. TABLE 11.1  Individual ADLs (Age 65+) Having difficulty Needing help China India Mexico Poland Romania Malaysia Dressing 11.74 8.58 14.90 14.82 14.06 3.26 Bathing 15.55 9.08 8.68 11.81 12.00 3.77 Eating 5.44 9.56 4.60 5.61 6.66 2.39 Getting in/out of bed 11.64 13.31 10.42 9.04 13.05 5.51 Using toilet 17.88 19.57 9.72 6.08 7.83 3.77 Prevalence estimates for China, India, Mexico, Poland, and Romania are weighted Prevalence estimates for Malaysia are not weighted Table 11.1 summarizes the estimated prevalence ty. In contrast, difficulty dressing is less prevalent in of having difficulty or needing help with each ADL China and India, with rates closer to 10 percent. Old- across countries. The following patterns emerge. We er Indians and Chinese exhibit the highest prevalence find similarly high prevalence rates of around 14 per- rate of difficulty with using the toilet, approaching cent in difficulty dressing for older individuals in Po- 20 percent, while their counterparts in Mexico, Ro- land, Romania, and Mexico. In these three countries, mania, Poland report significantly lower prevalence difficulty dressing is the most prevalent ADL difficul- rates ranging from 6 percent to 10 percent. COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E C H A P T E R 11 345 Difficulty eating is one of the least prevalent To gauge the extent to which the prevalence of difficulties across countries. Older Indians have having difficulty with at least one ADL varies across the most difficulty eating compared to residents of socio-economic groups, we estimate multivariate other countries, with a prevalence rate around 9.5 logistic models relating the indicator for having percent. The prevalence rate of difficulty eating in at least one ADL difficulty (for all countries except Romania, Poland, China, and Mexico is between 5 Malaysia) and the indicator for needing help with at percent and 7 percent. Getting in and out of bed is least one ADL (for Malaysia) to various individuals’ one of the most prevalent ADL difficulties, with a characteristics. This exercise allows us to infer how rate of about 13 percent in India and Romania, and strongly each individual characteristic correlates between 9 percent and 12 percent in Poland, Mex- with ADL difficulty/needing help holding the oth- ico, and China. In terms of difficulty bathing, older ers constant. We perform the estimation of the lo- Chinese exhibit the highest prevalence rate of about gistic regression separately for each country and 15 percent, followed by those in Poland and Roma- report the corresponding marginal effects in Table nia, with a rate around 12 percent, and residents of 11.2. Below, we first comment about the results in India and Mexico, with a rate around 9 percent. columns 1-5 of Table 11.2, next we comment about Malaysia’s prevalence rates of needing help the results in column 6 referring to Malaysia. with each activity are relatively low, ranging from The results reveal that, holding other factors 2 percent, for eating, to 5.5 percent, for getting in constant, those aged 80 and over are significantly and out of bed. more likely than those aged 65-79 to experience We construct and indicator taking value one difficulty with any ADL. The difference in the like- if the respondent reports having difficulty with at lihood of experiencing difficulty with at least one least one ADL and zero otherwise. Figure 11.1 plots ADL between these two age groups ranges from the prevalence of having difficulty with at least one 14 percentage points in Mexico to 20 percentage ADL across countries, excluding Malaysia. As can points in Romania. The association between gen- be seen about 28 percent of older adults in China der and difficulty with an ADL exhibit different and India have difficulty with at least one ADL; this patterns across countries. Relative to men, women proportion is 8 percentage points lower (at around are about 7 percentage points more likely to experi- 20 percent) in Mexico, Poland and Romania. For ence at least one ADL difficulty in China, India, and the MARS dataset, we construct an indicator tak- Mexico. In contrast, women are 6 percentage points ing value one if the respondent reports needing help less likely than men to experience difficulty with an with at least one ADL and zero otherwise. Figure ADL in Poland. Among Romanian older adults, we 11.2 shows that around 7 percent of older adults in do not detect any significant difference by gender. Malaysia need help with at least one ADL. Education is a strong correlate of having an ADL FIGURE 11.1 Having difficulty with any one ADL (Age 65+)     FIGURE 11.2 Needing help with any one ADL (Age 65+) 346 C H A P T E R 11 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E difficulty across all countries, although the features education gradient is less apparent in India, where of this association are country specific. In China the likelihood of experiencing an ADL difficulty is and Mexico there is an apparent education gradient between 5 and 6 percentage points higher among as the likelihood of experiencing difficulty with at those with less than secondary education relative least one ADL increases monotonically as the lev- to those with secondary education. In both Poland el of education decreases. In Mexico those with no and Romania, the fraction of older adults with no education and those with primary education are 11 education is small. Likely because of that, the mar- and 6 percentage points more likely to experience ginal effect of having no education is not significant. an ADL difficulty relative to those with secondary Yet we still detect a noticeable increase in the prob- education, respectively. Similarly, in China the like- ability of having an ADL difficulty of about 10 per- lihood of having at least one ADL difficulty is 8 and centage points when comparing individuals with 6 percentage points higher, respectively, for individ- secondary and primary education. uals with no education and only primary education Rural residents are more likely to have any ADL compared to those with secondary education. The difficulty in comparison to urban residents only in TABLE 11.2 Marginal Effects from Logistic Regressions of Having Difficulty/Needing Help with at Least One ADL China India Mexico Poland Romania Malaysia1 Age (ref: Age 65-79) Age 80+ 0.160*** 0.152*** 0.144*** 0.158*** 0.199*** 0.230*** (0.017) (0.009) (0.019) (0.027) (0.042) (0.040) Gender (ref: Men) Women 0.064*** 0.063*** 0.076*** -0.060** -0.002 0.087*** (0.014) (0.008) (0.018) (0.027) (0.040) (0.027) Education (ref: Secondary or higher) No education 0.076*** 0.046*** 0.108*** 0.092 -0.028 0.190*** (0.022) (0.012) (0.032) (0.066) (0.083) (0.037) Primary education 0.058*** 0.058*** 0.061** 0.099*** 0.109** 0.086*** (0.021) (0.012) (0.028) (0.030) (0.045) (0.032) Urban/rural (ref: Urban) Rural 0.052*** -0.000 0.011 -0.046 0.021 0.072*** (0.014) (0.008) (0.018) (0.028) (0.045) (0.027) Household income per person (ref: High income) Low income 0.076*** 0.034*** 0.002 0.067* 0.012 0.021 (0.024) (0.009) (0.024) (0.036) (0.047) (0.033) Middle income 0.053** 0.022** 0.038* 0.033 0.043 0.058 (0.024) (0.009) (0.022) (0.034) (0.046) (0.036) Missing income 0.044* 0.042 (0.023) (0.056) Living arrangement (ref: Lives with two or more) Lives alone 0.010 0.000 0.012 0.039 -0.010 0.010 (0.021) (0.015) (0.022) (0.038) (0.055) (0.047) Lives with one other -0.007 -0.002 -0.004 -0.004 -0.037 -0.040 (0.015) (0.005) (0.021) (0.029) (0.045) (0.030) N 6,967 21,255 6,930 1,205 685 1,375 Standard errors are reported in parentheses. *, **, *** indicates significance at the 90%, 95%, and 99% level, respectively. 1 The dependent variable for Malaysia is the cumulative number of ADL/IADLs a respondent needs help with, not the number they have difficulty with, as it is for all other countries COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E C H A P T E R 11 347 China, where living in a rural area is associated with a three IADLs elicited by all studies and available in 5 percentage point increase in the probability of hav- the harmonized dataset, namely: (i) difficulty pre- ing an ADL difficulty. Per-capita household income paring meals, (ii) difficulty shopping, and (iii) diffi- correlates with the likelihood of having difficulty culty taking medications. with an ADL in some countries but not in others. As for the questions asking about difficulty with China exhibits a steep income gradient: relative to ADLs, there exist some cross-study differences in individuals in the highest income tertile, those in the how survey instruments elicit difficulties with IADLs. lowest tertile (low income) and in the middle tertile LASI, MHAS, SHARE, and CHARLS ask IADL ques- (middle income) are 8 and 5 percentage points more tions in a similar way: “Because of a physical or health likely to have an ADL difficulty, respectively. In In- problem do you have difficulty with any of these activ- dia, relative to individuals in the highest income ter- ities? Exclude any activities you expect to last less than tile, those in the lowest tertile (low income) and in three months.” In LASI, SHARE, and MHAS, respon- the middle tertile (middle income) are 3 and 2 per- dents answer “yes” or “no” to this question, separately centage points more likely to have an ADL difficulty, for each IADL. MHAS respondents can also answer respectively. The association between household in- with “can’t do,” which is coded as having difficulty in come and ADL difficulty is much weaker in Mexico, the harmonized variable, or “doesn’t do”, which is only Poland, and Romania. Finally, we do not find that, coded as having difficulty if the respondents report holding other factors constant, living arrangement receiving help for that activity. For each activity in are significantly associated with the likelihood of CHARLS, if the respondent reports they have no dif- having difficulty with an ADL across countries. ficulty, then the harmonized variable is coded as not The last column of Table 11.2 presents the es- having difficulty. On the other hand, if the respondent timation results for Malaysia. Given the aforemen- reports that they have difficulty but can still do it, they tioned differences between the MARS questionnaire have difficulty and need help, or they cannot do it, then and that used by all other studies, we describe these the harmonized variable is coded as having difficulty. results separately. In Malaysia, we estimate that being MARS asks about receiving help with IADLs in 80 years or older is associated with a 23 percentage this form: “In the last week, did you need any assis- point increase in the probability of needing help with tance in performing the following activities (exclud- an ADL compared to being 65-79 years. Being female ing any difficulty that you expect to last less than 3 is associated with a 9 percentage point increase in months)?” If the respondent reports being able to do the probability of needing help with at least one ADL it all by themselves, then the harmonized variable is compared to being male. We detect a steep monotonic coded as not needing help. If, however, the respon- increase in the likelihood of needing help with an dent reports that they sometimes need some help, ADL as an individual’s education level decreases. Spe- always need some help, or always need help, then the cifically, relative to those with secondary education, harmonized variable is coded as needing help. individuals with no education and individuals with The difference in question wording between primary education are 19 and 9 percentage points MARS and the other studies prevents comparability more likely to need help with an ADL, respectively. In between Malaysia and other countries. Because of Malaysia, living in a rural area rather than in an urban that, we will comment about Malaysia’s results sepa- area is associated with a 7 percentage point increase in rately, without attempting a comparison with the other the probability of needing help with at least one ADL. countries. Similarly, to what we did for the analysis of Holding other factors constant there is no significant ADL difficulties, we exclude individuals from MHAS correlation of needing help with an ADL with either who answer that they don’t do and didn’t receive help per-capita household income or living arrangements. with an IADL, a group which cannot be identified in LASI, SHARE, and CHARLS. Lastly, CHARLS uses Limitations in instrumental activities of slightly more specific wording to assess difficulty daily living (IADLs) shopping for groceries, by adding “deciding what to buy and paying for it,” and taking medication, by in- While instrumental activities of daily living (IADLs) cluding “taking the right portion of medication right are not as critical as ADLs in enabling functional on time.” We do not expect this different wording to living, they still play an integral role in determin- affect the comparability of prevalence estimates be- ing the quality of an individual’s life (Guo & Sapra, tween China and India, Mexico, Poland, or Romania. 2022). In this report, we document difficulty with Table 11.3 summarizes the estimated prev- 348 C H A P T E R 11 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E TABLE 11.3 Individual IADLs (Age 65+) Having difficulty Needing help China India Mexico Poland Romania Malaysia Preparing meals 20.49 21.69 8.18 9.64 11.64 25.94 Shopping 17.47 27.25 12.67 16.76 18.64 35.94 Taking medications 9.13 17.81 4.44 5.45 7.66 12.39 Prevalence estimates for China, India, Mexico, Poland, and Romania are weighted Prevalence estimates for Malaysia are not weighted alence of having difficulty with each individual 10 percent need help with taking medications. It is IADL, documenting great heterogeneity across worth noticing that in Malaysia prevalence rates of countries. While slightly more than 20 percent of needing help with IADL difficulties are much larger individuals aged 65 and older need help preparing than those of needing help with ADL difficulties. meals in India and China, this proportion is only We construct and indicator taking value one if about 10 percent in Mexico, Poland and Romania. the respondent reports having difficulty with at least Difficulty with shopping for groceries is the most one IADL and zero otherwise. Figure 11.3 plots the prevalent IADL difficulty across countries, with a prevalence of having difficulty with at least one IADL prevalence rate ranging from 13 percent in Mexico across countries, excluding Malaysia. Older Indians to 27 percent in India. On the other hand, difficulty are at the highest risk of experiencing difficulty with taking medications is the least prevalent IADL dif- an IADL, with an estimated prevalence rate of 36 per- ficult in each country. Older adults in India exhibit cent. In comparison, the prevalence of having diffi- the greatest difficulty, with a prevalence rate of near- culty with at least one IADL is 9 percentage points ly 18 percent, followed by older adults in China and lower in China (at 27 percent) and 16 percentage Romania, with a prevalence rate of 9 percent and 8 points lower in Romania (at 20 percent). Mexico is percent respectively, and older adults in Poland and the country with the lowest prevalence rate of having Mexico, who have prevalence rates at around 5 per- an IADL difficulty (about 15 percent). For the MARS cent.2 As far as Malaysia is concerned, the results in dataset, we construct and indicator taking value one the last column of Table 11.3 reveal that more than if the respondent reports needing help with at least one third of individuals aged 65 and older need help one IADL and zero otherwise. Figure 11.4 shows that with shopping for groceries, about a quarter need around 45 percent of older adults in Malaysia need assistance preparing meals, and slightly more than help with at least one IADL. FIGURE 11.3 Having difficulty with any one IADL (Age 65+)    FIGURE 11.4 Needing help with any one IADL (Age 65+) 2   Access to medication is significantly more common in high income countries, which may explain the overall lower prevalence rates than those seen in other IADL activities. While western medicine use has been increasing over time in these countries, traditional medicine is still frequently used and found to have the highest use in India (Oyebode et al., 2016). COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E C H A P T E R 11 349 Next, we estimate multivariate logistic models education, individuals with only primary education relating the indicator for having at least one IADL are between 4 percentage points (Mexico) and 13 difficulty (for all countries except Malaysia) and the percentage points (Poland) more likely to report an indicator for needing help with at least one IADL IADL difficulty. With the exception of Poland, an (for Malaysia) to various individuals’ characteris- education gradient is apparent in all countries. tics. We perform the estimation of these logistic re- Living in a rural area is associated with a sta- gressions separately for each country and report the tistically significant increase in the probability of corresponding marginal effects in Table 11.4. having difficulty with an IADL only in China and The results in the first five columns of Table 11.4 India. These are also the two countries where the show that, holding other factors constant, those aged likelihood of reporting an IADL difficulty is signifi- 80 and over are more likely to experience difficulty cantly higher at lower levels of household per-capita with any IADL in all countries compared to those income decreases. There is no association between aged 65 to 79. The estimated increase in the likeli- household per-capita income and the probability of hood of having difficulty with an IADL associated experiencing an IADL difficulty in Mexico, Poland, with being 80 and older appears to be highly vari- or Romania. Similarly, differences in living arrange- able across countries, ranging from 17 percentage ment are only significantly associated with the prob- points in Mexico and Poland to about 23 percent- ability of having difficulty with an IADL in China age points in China, India, and Romania. Women and India. interestingly, the estimated marginal ef- are between 7 and 9 percentage points more likely fects are rather different in these two countries. In to experience an IADL difficulty compared to men China, living alone is associated with a 9 percentage in China, India, and Mexico, whereas the estimated point decrease in the probability of having difficulty marginal effect of gender is not statistically different with an IADL relative to those living with at least from zero in Poland and Romania. Education is one two other household members; in India, this effect of the strongest correlates of difficulty with an IADL is less sizeable at 4.5 percentage points. In China, across countries. Specifically, relative to those with living with just one other person is associated with secondary education, individuals with no education a 4 percentage point decrease in the probability of are between 9 percentage points (Mexico) and 16 having difficulty with an IADL relative to those liv- percentage points (Poland) more likely to report an ing with at least two other household members; in IADL difficulty. Relative to those with secondary India, this effect is not distinguishable from zero. TABLE 11.4 Marginal Effects from Logistic Regressions of Having Difficulty/Needing Help with at Least One IADL China India Mexico Poland Romania Malaysia1 Age (ref: Age 65-79) Age 80+ 0.233*** 0.224*** 0.174*** 0.168*** 0.237*** 0.225*** (0.015) (0.010) (0.014) (0.022) (0.033) (0.040) Gender (ref: Men) Women 0.067*** 0.089*** 0.090*** -0.012 0.036 0.088*** (0.014) (0.008) (0.016) (0.024) (0.039) (0.027) Education (ref: Secondary or higher) No education 0.105*** 0.121*** 0.089*** 0.161*** 0.086 0.194*** (0.022) (0.012) (0.030) (0.054) (0.058) (0.037) Primary education 0.056*** 0.053*** 0.043 0.133*** 0.111*** 0.094*** (0.021) (0.013) (0.028) (0.025) (0.041) (0.032) Urban/rural (ref: Urban) Rural 0.033** 0.044*** 0.009 -0.002 -0.043 0.070*** (0.014) (0.009) (0.016) (0.027) (0.041) (0.027) 350 C H A P T E R 11 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E China India Mexico Poland Romania Malaysia1 Household income per person (ref: High income) Low income 0.082*** 0.028*** 0.036 0.034 0.020 0.023 (0.023) (0.010) (0.023) (0.034) (0.044) (0.033) Middle income 0.045* 0.022** 0.025 0.027 0.110*** 0.060* (0.024) (0.010) (0.020) (0.031) (0.041) (0.036) Missing income 0.054** 0.039 (0.023) (0.056) Living arrangement (ref: Lives with two or more) Lives alone -0.097*** -0.045*** -0.022 -0.048 0.031 0.007 (0.021) (0.016) (0.018) (0.034) (0.046) (0.047) Lives with one other -0.040*** -0.006 -0.029 -0.026 -0.042 -0.035 (0.014) (0.010) (0.018) (0.025) (0.042) (0.030) N 6,967 21,255 6,925 1,205 685 1,375 Standard errors are reported in parentheses. *, **, *** indicates significance at the 90%, 95%, and 99% level, respectively. 1 The dependent variable for Malaysia is the cumulative number of ADL/IADLs a respondent needs help with, not the number they have difficulty with, as it is for all other countries Even if the dependent variable of our logistic activities that an individual reports difficulty (all regression is fundamentally different for Malaysia, countries, excluding Malaysia) or needing help with we find similar associations as those described for (Malaysia only). We then estimate a linear regres- other countries. As can be seen in the last column sion to gauge the relative importance of different of Table 11.4, being age 80+ is associated with near- socio-economic factors on this outcome. For this ly a 23 percentage point increase in the probability exercise, we limit the sample to those who report at of needing help with an IADL relative to being 65- least one difficulty with an ADL or IADL in China, 79 years. Also, women exhibit a 9 percentage point India, Mexico, Poland, and Romania, and to those lower likelihood of needing help with an IADL who report needing help with at least one ADL or compared to men. Relative to those with secondary IADL in Malaysia. This sample selection criterion education, older Malaysians with no education and is justified by the large number of individuals with only primary education have a 19 and 9 percentage no ADL and IADL difficulty.3 The resulting sam- point higher likelihood of needing help with an ple sizes are well above 1,000 for China, India, and IADL, respectively. Finally, living in a rural area is Mexico, leaving ample cases within each socio-eco- associated with a 7 percentage point increase in the nomic category. For Poland, Romania and Malaysia, probability of needing help with at least one IADL sample sizes are more limited. Because of that, the compared to living in an urban area. For Malaysia, group of individuals with no education is partic- we do not detect statistical associations between ularly small in Romania and Poland, and so is the needing help with an IADL and either household group of individuals living alone in Malaysia. These per-capita income or living arrangement. limited cell sizes should be taken into account when interpreting the regression results. Level of care need To proxy for the level/intensity of an individual’s care needs, we count the number of ADL and IADL 3   Keeping these individuals in the sample would require a different estimation approach, where the probability of having difficulty (needing help) with at least one ADL or IADL is modeled separately from the number of ADLs or IADLs the individuals have difficulty (needs help) with, conditional on this number being strictly positive. Since, we have already performed the former analysis, we only focus on the latter in this section . COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E C H A P T E R 11 351 FIGURE 11.5 Average number of cumulative ADL & IADL FIGURE 11.6 Average number of cumulative ADL & difficulties (Age 65+ having difficulty with an ADL or IADL) IADL needs (Age 65+ needing help with an ADL or IADL) Figure 11.5 shows that, among individuals with laysians report needing help with 2.1 activities on at least one ADL or IADL difficulty, Romanians average, conditional on needing help with at least have the highest number across countries, with 3.5 one ADL or IADL (Figure 11.6). It is apparent that difficulties on average. Poland, China, and India most individuals who have difficulty or need help each report an average close to 3 difficulties, while with either an ADL or an IADL actually have diffi- Mexico has an average of 2.5 difficulties. Older Ma- culty or need help with multiple activities. TABLE 11.5 Linear Regressions of Count of Cumulative ADL & IADL Difficulties/Help Needs China India Mexico Poland Romania Malaysia1 Age (ref: Age 65-79) Age 80+ 0.893*** 1.062*** 0.844*** 0.091 0.496 1.052*** (0.135) (0.071) (0.180) (0.319) (0.488) (0.196) Gender (ref: Men) Women -0.178 0.203*** 0.193 -0.055 -1.349*** -0.184 (0.118) (0.060) (0.165) (0.285) (0.467) (0.144) Education (ref: Secondary or higher) No education 0.267 0.165* 0.260 0.750 0.539 0.360* (0.185) (0.088) (0.333) (0.618) (1.008) (0.192) Primary education 0.145 0.004 -0.003 1.605*** 1.743*** 0.245 (0.173) (0.092) (0.289) (0.385) (0.525) (0.181) Urban/rural (ref: Urban) Rural 0.066 -0.159** -0.086 -0.167 0.091 -0.378*** (0.121) (0.067) (0.162) (0.345) (0.513) (0.128) 352 C H A P T E R 11 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E China India Mexico Poland Romania Malaysia1 Household income per person (ref: High income) Low income 0.032 0.479*** 0.010 -0.047 0.054 -0.060 (0.195) (0.069) (0.231) (0.416) (0.617) (0.186) Middle income -0.061 0.180*** 0.334 0.255 -0.450 0.052 (0.204) (0.066) (0.237) (0.413) (0.556) (0.198) Missing income -0.151 -0.425* (0.199) (0.217) Living arrangement (ref: Lives with two or more) Lives alone -0.044 -0.187* -0.192 -0.456 -0.126 -0.831*** (0.179) (0.111) (0.194) (0.437) (0.653) (0.117) Lives with -0.056 -0.031 -0.197 -0.092 0.439 -0.174 one other (0.116) (0.025) (0.200) (0.368) (0.525) (0.162) Constant 2.700*** 2.399*** 2.098*** 2.610*** 3.326*** 2.096*** (0.195) (0.082) (0.308) (0.420) (0.713) (0.195) N 2,610 8,568 1,852 285 149 612 Standard errors are reported in parentheses. *, **, *** indicates significance at the 90%, 95%, and 99% level, respectively. 1 The dependent variable for Malaysia is the cumulative number of ADL/IADLs a respondent needs help with, not the number they have difficulty with, as it is for all other countries The regression results reported in Table 11.5 show culties than being in the high income tercile, respec- great cross-country heterogeneity in which factors tively. Living arrangement does not correlate with the may predict the degree of an individual’s care needs. number of ADL/IADL difficulties in any country. Age stands out as one of the most important predic- The last column of Table 11.5 reports the esti- tors. On average, individuals aged 80 and over have mates for Malaysia. In this context, the number of more difficulties – from 0.5 in Romania to 1 in India – ADLs/IADLs individuals need help with increas- than individuals aged 65 to 79, although the estimated es significantly with age (1 more activity for those coefficients are not statistically significant for Poland aged 80 and over relative to those aged 65-79) and and Romania. Gender is significantly associated with decreases with living in a rural area (0.4 fewer ac- the degree of care needs only in India, where women tivities for those in a rural area relative to those in have 0.2 more difficulties than men, and in Romania, an urban area) and with living alone (0.8 fewer ac- where women have 1.3 fewer difficulties than men. tivities for those living alone relative to those living Having lower education is associated with a higher with at least two other household members). number of difficulties in Poland and Romania, where we estimate that individuals with only a primary ed- Receipt of care ucation have on average more than 1 additional diffi- culty than individuals with secondary education. In In this section, we assess the degree to which an China, India, and Mexico, there no significant asso- individual’s care needs are met and by whom. We ciation between education and number of ADL/IADL note that there is significant variability across sur- difficulties. Living in a rural area correlates negatively veys in the questions eliciting whether a respondent and significantly with the number of difficulties only receives care and who provides that care. We dis- in India, where older adults in rural areas have, on av- tinguish between informal care, usually provided erage, 0.2 fewer difficulties than their counterparts in by family members and friends, and formal care, urban areas. Similarly, per-capita household income which generally includes hired helpers, govern- correlates significantly with the number of difficulties ment-provided assistance, or nursing home staff. only in India, where being in the low and middle in- In CHARLS, Chinese respondents who have come tercile is associated with 0.5 and 0.2 more diffi- difficulty with an ADL or IADL are asked whether COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E C H A P T E R 11 353 anyone ever helps with that specific activity, and, if tionship with the three people who most help, sepa- someone helps, to indicate the relationship with the rately for ADLs and IADLs. When constructing the people providing help most often. When construct- harmonized variables, spouse, child, child-in-law, ing the harmonized variables, spouse, children, grandchild, parent, other relative, and other people children-in-law, grandchildren, sibling, parents, are considered to provide informal care, while paid other relative, and other people are considered to persons are considered to provide formal. provide informal care, while paid helpers, volun- In SHARE, Polish and Romanian respondents teers, employees of a facility, and the community are separately asked whether a family member, friend, are considered to provide formal care. or neighbor from outside the household has provided In LASI, Indian respondents who report having help with ADL or IADL in the past year, and wheth- difficulty with an ADL or IADL are asked whether er someone living in their household has helped anyone helps with those difficulties and to indicate with ADL in the past year. If they report receiving the relationship with the single person who helps help, respondents are asked to indicate their relation- them most often. When constructing the harmonized ship with the helpers choosing from a list including variables, spouse or partner, child, grandchild, parent, spouse/partner, child, child-in-law, grandchild, par- sibling, other relative, ex-spouse or partner, and other ent, sibling, other relatives, and several non-relative people are considered to provide informal care, while categories like friend or neighbor. When constructing non-professional paid helpers and professionals (paid the harmonized variables, all these categories are clas- or non-paid) are considered to provide formal care. sified as informal carers. Respondents are also asked Since LASI only ascertains the relationship with the if they have received professional or paid services in person providing help most often, it is not possible to their home in the past year for ADL and IADL. In the determine to what extent the care received by an indi- harmonized variables, this is coded as formal help. vidual is obtained from informal and formal sources. In MARS, Malaysian respondents who report In MHAS, Mexican respondents who report needing at least some help are sometimes asked to having difficulty with an individual ADL or IADL report who usually helps with each ADL and IADL. are asked whether someone helps with that specific When constructing the harmonized variables, activity, and, if someone helps, to indicate the rela- spouse, child, grandchild, and other people are con- FIGURE 11.7 Care mix (Age 65+ having difficulty with an ADL or IADL) *The subsample for Malaysia is (Age 65+ needing help with an ADL or IADL) 354 C H A P T E R 11 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E sidered to provide informal care, while domestic the proportion of older adults with unmet care needs maids and professional caregivers are considered ranges from 30 percent to 70 percent across these to provide formal care. However, it is important to countries. The striking difference between India note that in MARS respondents who report needing and the other countries suggests that there is a dif- help with an activity do not have an option to choose ferent cultural understanding of family roles so that “not receiving help” for that activity. Once again, this any help received by an individual with ADL/IADL questionnaire difference makes the MARS’ variable difficulties is not necessarily recognized as such. measuring care receipt not directly comparable with As mentioned above, for Malaysia our measure of the measure of care receipt in other studies. care receipt is only available for respondents need- Figure 11.7 shows that, among all individuals ing help with ADLs/IADLs. For this reason, Figure who have difficulty with an ADL or IADL, approxi- 11.7 shows that the fraction of individuals receiving mately 70 percent of older Chinese and Romanians help in Malaysia is one. Because of that, Malaysia is receive any care, while close to 60 percent of older excluded from the next set of analyses, which inves- Mexican and Polish adults receive any care, and only tigates the association between the likelihood of re- 30 percent of older Indians receive any care. As such, ceiving care and socio-demographic characteristics. TABLE 11.6 Marginal Effects from Logistic models of receiving any care (Age 65+ having difficulty with an ADL or IADL) China India Mexico Poland Romania Age (ref: Age 65-79) Age 80+ 0.128*** 0.105*** 0.188*** 0.096 0.203** (0.024) (0.013) (0.033) (0.067) (0.080) Gender (ref: Men) Women -0.020 0.034*** 0.046 0.016 -0.148* (0.020) (0.013) (0.034) (0.064) (0.082) Education (ref: Secondary or higher) No education 0.061** -0.035* 0.050 0.388*** 0.076 (0.031) (0.019) (0.056) (0.144) (0.137) Primary education -0.038 -0.001 0.052 -0.039 0.200** (0.029) (0.019) (0.049) (0.081) (0.086) Urban/rural (ref: Urban) Rural -0.019 0.001 -0.022 0.094 -0.142 (0.020) (0.013) (0.034) (0.068) (0.087) Household income per person (ref: High income) Low income -0.062* -0.062*** 0.025 -0.080 -0.120 (0.034) (0.015) (0.043) (0.084) (0.101) Middle income -0.065** -0.021 -0.028 -0.047 0.112 (0.033) (0.014) (0.041) (0.084) (0.094) Missing income -0.073** (0.033) Living arrangement (ref: Lives with two or more) Lives alone -0.210*** -0.182*** -0.118*** 0.040 0.112 (0.027) (0.029) (0.040) (0.087) (0.105) Lives with one other -0.032 -0.017 -0.120*** 0.033 0.075 (0.021) (0.040) (0.041) (0.078) (0.087) Count of ADL/IADLs 0.078*** 0.053*** 0.142*** 0.077*** 0.027* (0.006) (0.002) (0.015) (0.017) (0.015) N 2,610 8,567 1,851 278 149 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E C H A P T E R 11 355 In Table 11.6, we estimate logistic regressions age points, respectively. Being female is associated relating care receipt for individuals with at least with a 3 percentage point increase in the probabil- one ADL or IADL difficulty and socio-demograph- ity of receiving any care in India, but not in other ic characteristics, controlling for the number of countries. The effect of education varies significantly ADL/IADL difficulties. Not surprisingly, this latter across countries. Relative to those with secondary variable correlates positively and significantly with education, individuals with a lower level of educa- the probability that an individual receives any care. tion are between 20 and 40 percentage points more However, the strength of this correlation differs likely to receive any care in Poland and Romania. across countries. One more ADL/IADL difficulty is In China, those with no education are 6 percentage associated with a 14 percentage point increase in the points more likely to receive care relative to individ- likelihood of receiving care in Mexico, with about uals with secondary education. In India and Mexico, an 8 percentage point increase in China and Poland, there is no significant association between education with a 5 percentage point increase in India, and with and care receipt. A lower level of per-capita house- only a 3 percentage point increase in Romania (sig- hold income is associated with a lower probability of nificant at the 10% confidence level). Conditional receiving care in China and India, while the income on the number of ADL/IADL difficulties, relative to marginal effects for Mexico, Poland, and Romania being 65-79 years, being aged 80 and over increases are not statistically different from zero. Relative to the probability of receiving any care from a mini- individuals living with at least other two household mum of 10 percentage points in India and Poland members, those living alone are 12, 18, and 21 per- (not significant) to a maximum of 20 percentage centage points less likely to receive care in Mexico, points in Romania, with China and Mexico having India, and China, respectively. more intermediate increases of 13 and 19 percent- TABLE 11.7 Marginal Effects From Logistic models of receiving formal care (Age 65+ having difficulty with an ADL or IADL) China India Mexico Poland Romania Age (ref: Age 65-79) Age 80+ 0.018* 0.002** 0.012 0.093 -0.017 (0.010) (0.001) (0.009) (0.058) (0.045) Gender (ref: Men) Women -0.005 0.003** 0.006 -0.048 0.014 (0.011) (0.001) (0.010) (0.052) (0.037) Education (ref: Secondary or higher) No education -0.005 -0.004** -0.030* 0.105 0.000 (0.016) (0.002) (0.016) (0.092) (.) Primary education -0.012 -0.002** -0.008 -0.023 0.003 (0.015) (0.001) (0.012) (0.062) (0.046) Urban/rural (ref: Urban) Rural -0.023** -0.002 0.012 -0.014 -0.138*** (0.010) (0.001) (0.011) (0.049) (0.045) Household income per person (ref: High income) Low income -0.005 -0.002 -0.033** 0.059 -0.193*** (0.014) (0.001) (0.013) (0.067) (0.069) Middle income -0.025 -0.001 -0.029** -0.002 -0.041 (0.022) (0.001) (0.013) (0.065) (0.034) Missing income -0.002 (0.015) 356 C H A P T E R 11 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E China India Mexico Poland Romania Living arrangement (ref: Lives with two or more) Lives alone 0.051*** 0.003*** 0.023 0.099 -0.001 (0.017) (0.001) (0.015) (0.065) (0.042) Lives with one other 0.021 0.000*** 0.026 0.013 -0.054 (0.014) (0.000) (0.016) (0.065) (0.040) Count of ADL/IADLs 0.007*** 0.000** 0.007*** 0.023** 0.017** (0.002) (0.000) (0.003) (0.010) (0.008) N 2,610 8,561 1,851 282 135 In Table 11.7, we estimate logistic regressions point increase in the probability of receiving formal relating the probability of receiving formal care care. In India, women and more educated individu- among individuals with at least one ADL or IADL als are more likely to receive formal care. Although difficulty to socio-demographic factors and the significant, these estimated effects reflect tiny in- count of ADL/IADL difficulties. The prevalence of creases from an average probability of receiving formal care receipt is low in all countries: 3 percent formal care of less than 1 percent. Living in a rural in China, less than 1 percent in India, 2 percent in area is associated with a 14 and 2 percentage point Mexico, 15 percent in Poland, and 6 percent in Ro- decrease in the likelihood of receiving formal care mania. Overall, the estimates do not identify very compared to living in an urban area in Romania clear patterns across countries, partly due to the and China, respectively. Low income is negatively extremely low prevalence of formal care receipt in and significantly associated with a lower likelihood certain countries and, therefore, to the high selec- of receiving formal care in Mexico and Romania, tivity of the group receiving formal care. As with the but not in other countries. Holding other factors receipt of any care, the count of ADL and IADL dif- constant, living alone increases the probability of ficulties correlates positively and significantly with receiving formal care in China and India, relative the probability of receiving formal care. Specifical- to living with two or more other people, by 5 and 1 ly, in both China and Mexico each additional ADL percentage points, respectively. or IADL difficulty is associated with a 1 percentage A SUPPLY SIDE ASSESSMENT CURRENT LTC LANDSCAPE IN LMICS As shown in the previous section on demand, across LMICs and MICs, familial (or informal) care remains the primary form of care for older persons. As publicly financed LTC is limited, paid care is primarily paid for by older persons and their families out-of-pocket. There is generally limited availability of and access to formal home care services, community-based services (such as adult daycare centers), and institutional services (such as nursing and residential care facilities). In this section, we briefly review and summarize existing evidence on the avail- ability of formal LTC services and delivery systems in selected countries across Asia, Central and South America, Central and Southern Europe and Middle East, supplemented by some evidence from Sub-Saharan Africa. Availability of LTC services Asia. Asia is home to several LMICs that have the fragmented to meet the escalating care needs. largest and fast-growing older populations in the In China, the current LTC landscape, inclu- world, including China, India, and Indonesia. To sive of public and private actors, has been shaped varying degrees, formal LTC services are emerging by government policies, consumer needs, market in these countries to fill the void of family based forces, and tradition. It is characterized by a fast elder care. However, these services are still in the growing residential (institutional) care sector, rela- early stage of development, and, far too spotty and tively slow and limited development of home- and COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E C H A P T E R 11 357 community-based services, shortage of well-trained Mental Health Program, and National Program professional LTC workforce, weak quality regu- for Palliative Care (Scheil-Adlung 2015). Public lations, and a lack of organized financing (Feng et al. funding for LTC is limited, and as a result, out- 2020). Public LTC financing is minimal and largely of-pocket payment for LTC is the norm among the limited to supporting a narrowly defined group of few who can afford it, leaving the vast majority of welfare recipients and subsidizing the construction older people without access to LTC services. of residential care beds and operating costs. To In Indonesia, the government is responsible for address the gaps of unmet LTC needs among the LTC coordination and delivery, including home and majority of older people outside the means-tested community-based care (Asian Development Bank social welfare system, China’s primary strategy is to 2021). In 2016, the Indonesian government encourage entrance of the private sector in LTC ser- launched the National Strategic Plan, which stated vice provision, consistent with the transition of the that government-mandated community health government’s traditional role as direct provider to centers were responsible for providing primary purchaser and regulator of LTC services (Feng et al. health care and LTC for older people, including 2012). As of 2020, China had 8.21 million residential providing free health check-ups and organizing care beds registered, or 31 beds per 1,000 people social activities for older people (Dyer et al. 2019). ages 60 (Ministry of Civil Affairs 2021). However, LTC is not yet provided in an integrated More recently, China started piloting public and coordinated manner, the minimum require- social insurance LTC financing models, and ments for care provision are not met at many state- concurrently, programs for integrating healthcare funded community health centers, and local political and LTC services in selected locations across the commitment and resource availability have limited country, though evaluation of the impact and LTC delivery (Asian Development Bank 2021). viability of these pilot programs is lagging. The There are approximately 277 residential homes for ongoing Long Term Care Insurance (LTCI) pilots older people, with a capacity of 18,100 beds; of these signal China’s move toward using public social facilities, 3 are run by the central government, 71 by insurance as the core financing strategy for LTC local governments, and 189 are private (Asian for the broad population, following the same stra- Development Bank 2021). Only a few private tegy it adopts for financing healthcare. The shor- organizations provide institutional LTC and some tage of well-trained LTC workers and allied health private companies supply day care and equipment professionals is a major barrier for quality im- services, estimated to support no more than a few provement and for effective integration of medical thousand older people mainly in urban areas (Asian and LTC services (Feng et al. 2020). Development Bank 2021). In India, families assume the primary responsibility In Viet Nam, while the government has issued of caring for older people, which is codified in law, as in many policies to support older persons and ensures China. The Maintenance and Welfare of Parents and subsidies and health insurance for those aged 80 Senior Citizens Act (passed in 2007) obliges adult and over and for the vulnerable elderly, there is no children to provide “maintenance”—including food, com-prehensive LTC-based model or national clothing, medical attendance and treatment—to their integration of LTC service provision. While the parents if in need, and those who do not fulfill this Ministry of Health provides healthcare for the obligation can be prosecuted (Scheil-Adlung 2015). Few elderly in health-care facilities and communities, formal LTC services such as old age homes, day care LTC service provision mainly supports family care centers, residential facilities and domiciliary care with home-based services. Non-government orga- services are available, and these services are nizations (NGOs) and the private sector also predominantly provided by private and not-for-profit provide LTC, and paid home care is emerging for organizations, although a few public facilities also exist supporting older people without the means to pay (Rajagopalan et al. 2020). Currently, care homes and for private care (Van, Tuan, and Oanh 2021). residential facilities are largely unregulated, with limited Residential aged care, including care centers, is not data available on their numbers and characteristics widely available. In 2015-2016, of residential care (Harbishettar et al. 2021). Some publicly support-ed centers available, 36 percent were public, 36 LTC services exist but they are scanty and fragmented percent were NGOs or religious providers, and 27 under the auspices of multiple government programs, percent were private, of which 82 percent were such as the Integrated Program for Older Persons, licensed (Dyer et al. 2019). Nursing homes are National Program for Healthcare of Elderly, National mostly located in urban areas and mainly afforded 358 C H A P T E R 11 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E by the wealthy (Van, Tuan, and Oanh 2021). In Viet sioners(Lloyd-Sherlock et al. 2017). Private ser- Nam, the government provides social assistance vices are increasing but mostly limited to payments to older, poorer persons without close consumers with higher incomes (Inter-American family or retirement pensions; however, the number Development Bank 2020c). of ben-eficiaries is limited (Dyer et al. 2019). In Brazil, there is no federally mandated The LTC system in Thailand prioritizes coverage for the aging population in need of aging in place and home and community- LTC. Through a means-tested system, Brazil based care. The government has implemented, provides some public LTC services focused on alongside local authorities, an integrated home sheltering people in economic deprivation and community-based care model (further (Giacomin et al. 2021). Brazil’s LTC homes are described below). Residential LTC is available for scarce, public LTC expenditure per older adult is those with complex care needs and insufficient minimal, and it is estimated that at least caregiving support at home; services for 600,000 formal LTC workers are needed to fill dependent older persons are available at private the gap (Giacomin et al. 2021). nursing homes, private hospitals, government In Argentina, there is little public funding for residential homes, and homes for poor older LTC services. While the government does cover a persons supported by charitable organizations range of publicly operated services for older (Asian Development Bank 2020). In 2016, there persons, these services are fragmented and were 442 private facilities offering residential uncoordinated due to the overlapping of functions LTC services (Asian Development Bank 2020). and services provided by national, provincial, and Central and South America. Across countries mu-nicipal actors (Inter-American Development in Central and South America, comprehensive LTC Bank 2020a). Private schemes are estimated to cover policies and systems are essentially nonexistent. no more than eight percent of older people Public funding or direct provision of LTC services is with higher incomes who can afford access to scarce. The private market for senior care is services provided by private and civil society emerging in some countries, with access limited organizations such as LTC facilities, day centers, by ability to pay for such care. rehab centers, and home-care services (Dyer et al. In Mexico, there is little public funding for 2019). LTC services, and the responsibility for service Similarly, in Chile, a variety of publicly development rests with the states (OECD 2011). supported LTC services, including residential care Health care for older people is provided through facilities, day centers, and home care, are offered but the Mexican Health System, which also covers all a nationally organized LTC system is absent (Inter- other age groups (Gutiérrez Robledo, López Ortega, American Development Bank 2020b). Private and Arango Lopera 2012). The few government- services include those provided in residential care provided services for the elderly that do exist are settings, home care and support services, day extensions of poverty reduction or other social centers, and telecare services, with the non-profit service programs. There are few private (for profit sector, particularly religious organizations, playing a and not-for-profit) institutions offering services strong role in service provision (Inter-American such as adult daycare and institutional LTC for Development Bank 2020b). non-self-sufficient older people without family Central and Southern Europe and Middle (OECD 2011). East. Even though out of all MICs considered in the In Costa Rica, recent policy efforts have at- section on demand, Poland and Romania have the tempted to formalize a national LTC system (Matus- highest proportion of older people with ADLs and Lopez and Chaverri-Carvajal 2022). The government IADLs who use formal LTC services, traditional has focused on facilitating access to services, such as informal care by family members predominates in long-term residential services, day centers and home both countries. At the same time, in response to the care visits, using subsidies to families or non-profit growing care needs of the population, formal care service providers (Inter-American Development organized by the state or private institutions is Bank 2020c). In 2010, Costa Rica’s National Council progressively supplementing traditional informal for Older People established the Progressive Atten- care provision. In Poland, delivery and financing of tion Network for Integral Elder Care, aimed at formal social care services is mainly the respon- establishing community-based LTC networks na- sibility of local governments, which are obligated by tionwide to conduct activities such as training retired the central government to finance and provide teachers to act as unpaid community volunteer pen- social care services from their local budgets, covering COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E C H A P T E R 11 359 institutional, community, and home-based care has implemented a new national community- (while the authority representing the central based palliative care program in line with the government in local governments has supervisory National Turkish palliative care policy, making powers over institutional services). Institutional Türkiye the only country in the MENA region care for older people is split between the health (besides Israel) to have such policies (Ismail and and social welfare branches. Financing of insti- Hussein 2021). tutional medical care services provided to eligible In Greece, private formal (self-paid) care is individuals is the responsibility of the National widespread and is received by approximately 25 Health Fund, while the provision is organized by percent of people with needs (World Bank, 2021). various (public, as well as for- and not-for-profit) Public formal services are also available—the providers. The health and social brunches are not publicly-provided “Help at Home” care is received well integrated around the needs of old people or by 9 percent of elderly with care needs and is in terms of service provision or financing. increasing with time. Help at Home dispropor- Administrators and providers face incentives to tionally serves the rural population (as compared to transfer patients to the other part of the system private care). About 41 percent of seniors with care (European Commission 2021). Romania has needs utilize professional public or formal private significantly increased its LTC capacity in the last care (in most cases in addition to informal care) seven years, which allowed local authorities to and there has been an increase in the proportion of better identify, assess, and establish eligibility to seniors who rely on a combination of informal and support older people with care needs. Similar to professional self-paid care (World Bank, 2021). Romania, it continues to have a fragmented LTC The Arab region is characterized as having system. Both countries benefited from consi- primarily residual social welfare systems that rely derable European-level financial resources, which heavily on family or community-based social sup- have been catalytic in encouraging the deve- port, especially for social care which has long been lopment of care capacity and supporting expansion neglected in social policy in the region (Hussein into underserved communities. Most of this and Ismail 2017). Policy attention to LTC needs is expansion occurred through private provi-ders scant, given deeply rooted cultural norms entering the market (European Commission 2021). emphasizing the role of the family (particularly Similarly, in the MENA region, there is women) in the elderly care provision and support evidence indicating increased demands for a formal system. Although informal LTC continues to LTC market and the emergence of aged care dominate, there is evidence of increased use of economies within relatively unregulated structures formal care by older people, such as care homes (Hussein 2022). In Türkiye, as in most of the or home care services, which are mostly initiated MENA region, care for older people is primarily and covered by civil society and religious provided informally by the family and the organizations, sometimes subsidized by public community (Ismail and Hussein 2021). However, funds (Hussein and Ismail 2017). sociodemographic trends, such as changes in family Sub-Saharan Africa. In Sub-Saharan Africa, structures, migration, and increases in women’s most organized care is provided in urban areas, and formal employment, are undercutting the availa- the two most common models for LTC delivery are bility of such care. There exists LTC at home through charitable organizations (faith-based, civil provided by informally employed domestic and society or public welfare bodies) or private services migrant live-in care workers, funded either through for higher-income populations, provided mainly in cash-for-care schemes or out-of-pocket. Home care residential homes (World Health Organization is viewed as a culturally appropriate option while 2017). Institutional care is relatively new and often residential care is stigmatized. In recent years, the not available. In Kenya, there were approximately 16 Ministry of Family and Social Policy has piloted LTC facilities in 2017 and the main providers were some new elderly care interventions such as shared religious organizations (Dyer et al. 2019). In South living and elderly care centers, organized by the Africa, limited publicly funded LTC exists. Resi- state for groups of older people to live together dential care is provided mainly by NGOs or religious with support workers attending to their needs organizations, and only 2 percent by government during the day. Since 2010, the Ministry of Health (Dyer et al. 2019). 360 C H A P T E R 11 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E LTC delivery systems company that delivers the desired services or A variety of LTC provision and delivery models operate a government-built facility, or both are emerging in LMICs where some forms of for- (Feng et al. 2020). A similar approach is used mal LTC services exit. As described in the country for government purchasing of home and examples below, these models may serve as useful community-based services, which are mostly examples for other countries looking to develop and operated by private-sector providers. Under integrate feasible LTC models. the government-contractor model, resources for In China, government-run social welfare organizing in-home care services are allocated homes and residential care facilities continue to by various levels of government, including function as the traditional safety net by providing districts, subdistricts, and local communities. housing, full costs of living and LTC services for in- The services are provided to eligible older people dividuals who qualify as welfare recipients—those residing in these settings, with various levels of who have no ability to work, have no source of in- government initiating, funding, and supervising come, and have no families or relatives to support these services. The providers are typically them. In rural areas, these individuals are also engaged through a government procurement known as wubao (which literally translates as “five process. Most of them are non-profit. guarantees”), for whom the local government guar- In other MICs, policy efforts and private sector antees food, clothing, housing, medical care, and initiatives have put an emphasis on fostering the burial expenses (Feng et al. 2020). While tradition- development of home and community-based LTC ally these public facilities exclusively served welfare services—aptly so to meet older people’s needs and recipients meeting stringent eligibility criteria, over preference for such services. In Thailand, for example, time they have opened doors to private-paying indi- through a pilot program established in 2016 and viduals who have the means and who can no longer managed by the National Health Security Office and be cared for at home by family members. Indeed, local authorities, the government is working to increase nowadays private-paying customers make up the access to, and availability of, home and community- majority of the resident population in most govern- based services. The program operates through a care- ment-run residential care facilities in urban areas, management system, providing 2–8 hours of home- although the clientele of rural facilities continues to based care support a week, depending on need, be dominated by wubao elders (Feng et al. 2012). through caregivers with 70 hours of training who are Over the past 20 years, the number of residen- supervised by a care manager (Asian Development tial care facilities in the urban private sector has Bank 2020). Medical services, including preventive grown rapidly in China, in response to the surge of services and physiotherapy, are also available through consumer demands for LTC services and policy sup- the LTC pilot program and the Thailand universal port (Feng et al. 2020). This development has been health coverage package (Asian Development Bank driven by real estate developers, venture capitalists, 2020). Romania and Poland are also experimenting investors, and other businesses interested in tap- with expanding their community-based LTC services, ping China’s booming senior care market. The gov- with local authorities increasingly commissioning and ernment has leveraged various financial incentives purchasing services from the rapidly developing and (e.g., lump-sum subsidies for new constructions growing private for-profit and not-for-profit sector and recurrent subsidies for occupied beds) and oth- (European Commission 2019). er preferential policy treatments (e.g., tax breaks, Another community-based model of LTC pro- land allotment or leasing, and reduced utility rates) vision and delivery is driven by organizations of to incentivize entry of the private sector. Ranging older people. In Cambodia, China, India, Indonesia, from senior apartments to assisted living facilities Nepal, Sri Lanka, Myanmar, and Viet Nam, these are and retirement communities, these facilities pro- called Older People’s Associations (OPAs), which are vide various levels of personal care assistance and membership organizations led or managed by older professional services. Virtually all residents in these people to facilitate activities and deliver services for facilities are paying privately for the services. older people. By partnering with government service The Chinese national and local governments providers, OPAs provide a social protection function encourage various types of public-private partner- which complements existing mechanisms (HelpAge ship LTC service delivery models, whereby the gov- International n.d.). OPAs have been involved in ernment contracts with a competent private-sector organizing medical check-ups, conducting home COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E C H A P T E R 11 361 visits, and providing health education for older LTC policy planning can include many people. Additionally, OPAs play an important role goals such as analyzing demand and supply, in campaigning and promoting the interests of older identifying necessary changes for improvements people to policy makers (Petsoulas 2019). in the system and the reallocation of resources, In Kenya, a private nursing agency containing care costs, and improving the quality provides individualized, in-home care from of services. Described below are select country professionals to those who can afford to pay for examples. such care or who have medical insurance that Asia. In many countries in Asia and the Pacific, covers home-based care (Petsoulas 2019). While the first legal mention o f LTC was through laws for this type of private sector driven model is older people, and several subsequent national policies growing in popularity, gaps in access exist for or plans have used these laws as foundations (Walker older people without ability to pay or insurance. and Wyse 2021). For example, in Mongolia, the In Tanzania, HelpAge International imple- revised Law for the Elderly in 2017 was the building mented the Better Health for Older People in Africa block for the development of a draft Strategic Plan program (2014-2017), funded by the Department on Long-Term Care for Elderly led by the Ministry for International Development of the UK and aimed of Labor and Social Protection (Walker and Wyse to improve access to home-based services for poor 2021). In Singapore, LTC was originally included older people in need. The program, which within aging laws covering social welfare, and the supported approximately 4,500 older people, was current LTC strategy is part of the Successful Ageing delivered by 425 trained volunteers who were framework led by the Ageing Planning Office within selected in consultation with the local community the Ministry of Health (Walker and Wyse 2021). and supervised by registered nurses and clinical To assist in policy development, governments officers (Petsoulas 2019). Within this program, care utilize, to a certain extent, advisory groups and as- plans are developed to assist with ADLs and social sociations for input. In Cambodia, China, India, support programs are provided, such as programs In-donesia, Nepal, Sri Lanka, Myanmar, and Viet to aid in socialization with other older people and Nam, Older People’s Associations (OPAs) play an workshops from volunteers to learn about topics import-ant role in campaigning and promoting such as nutrition (Petsoulas 2019). the interests of older people to policy makers (Petsoulas 2019). The role of government policies In most countries of Asia and the Pacific, LTC is operated in a decentralized or partly decen- Government stewardship functions across LMICs tralized system, and responsibilities for different vary as they are a result of the interaction of elements of LTC are divided among national, numerous factors, such as political and economic provincial, city, and local levels (Walker and Wyse contexts, social and cultural factors, as well as 2021). A key question for policymakers in these existing health and social welfare system structures countries is whether to set up a standalone LTC (Walker and Wyse 2021). The main functions of system or cover LTC within the existing public the government in LTC include policy planning health care systems (UNESCAP 2018). and implementation, regulation through developing Central and South America. In much of and monitoring quality standards and fostering the Central and South America, aging policy has LTC market, and workforce development. Below, followed a progression from a charity- to a rights- we briefly re-view these themes drawing on the based approach (Calvo et al. 2019). In Argentina, limited experiences of LMICs, with occasional Chile, and Costa Rica, with continued progress and reference to those of HICs, as appropriate. success in the development of pensions as well as a As noted above, comprehensive LTC policies growing emphasis on older people’s rights, aging and systems are not well developed in virtually all policy has begun to slowly make its way up the LMICs. Many of these countries struggle with a policy agenda (Calvo et al. 2019). lack of a clear definition of the tasks, roles, and In Argentina, throughout the 2000s, the responsibilities of different ministries, agencies, National Directorate of Policies for Older Adults and organizations in caring for older people. As a (DI-NAPAM) was responsible for implementing result, the inappropriate use of acute care hospital and coordinating aging policies and supporting the services and emergency departments is high training of aging professionals. The creation of a among the elderly (Giacomin et al. 2021). Federal Council of Older Adults encouraged older 362 C H A P T E R 11 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E adults’ participation in policy design (Calvo et al. community. The most recent aged care plan 2019). Chile’s government followed a similar path. (2017) prioritized home-based care services with After success in implementing its national pension responsibilities at the municipal level (Ismail and system in the 1980s, Chile created the National Hussein 2021). Service for Older Persons (SENAMA), a public, Sub-Saharan Africa. LTC policy development is decentralized service located within the Social De- lagging in Sub-Saharan Africa, relative to other parts velopment Ministry, responsible for promoting the of the world. However, countries in this region have health, quality of life, rights, and autonomy of older begun to form a range of policy frameworks Chileans (Calvo et al. 2019). encompassing aspects of LTC, both directly and in- Costa Rica is in the midst of creating and directly, including older person focused conditions implementing a national LTC system and is one of in policies or constitutions, dedicated bills on older the first middle-income countries to do so. Th e LTC adults or, as in Tanzania and Zimbabwe, national system consists of three benefits, including c are strategies on healthy ageing (Aboderin 2019). services, cash for care (only in specific cases), and Examples of government mechanisms for caregiver training (Matus-Lopez and Chaverri- quality monitoring from HICs include inspection, Carvajal 2022). The Ministry of Human Deve- regulation, and/or reliance on market forces. lopment is responsible for establishing the guidelines England and Australia utilize inspection-based and organizing the system while it is implemented surveillance of LTC services against certain by different public institutions working in social, standards. The US and Canada utilize data health care, labor, and educational areas. The system measurement or public reporting as a driver of is funded through general tax revenues and market forces to maintain quality. Across these copayment and is centralized in that local govern- countries, the main responsibility for regulating ments do not participate in the design or financing quality remains with government (Mor, Leone, of the program. and Maresso 2014). Central and Southern Europe and Middle Governments can play a multitude of roles in East. Romania currently embarked on far-reaching ensuring availability and accessibility of services. reforms of LTC services with the main stated pri- These include directly providing care, acting as a ority to develop home care and day care services at purchaser of services from qualified providers, and a community level and to increase the availability playing a role in fostering LTC markets (Walker of public residential institutions. It is envisaged that and Wyse 2021). The capacity of governments to the share of funds from the state budget provided to foster markets and develop public-private part- local authorities for social services for older people nerships is dependent on their ability to develop would increase and the flow of funds change accord- contracts, set prices, and monitor and supervise ing to the principle that funding follows beneficia- private-sector providers (Block et al. 2009). ries. LTC reforms are part of a broader package to Governments can also assist providers through develop integrated community social services and subsidies and grants. ensure the availability of qualified social workers, It is also important to strike a balance between especially in rural areas (World Bank 2022). Poland the government function of regulation and is making major strides in improving the quality of regulation from market forces. In the US, the LTC (Ministry of Finance of Poland, 2021). government plays a large role in regulation within In the Middle East too, there has been some the LTC industry to maintain a minimum level of re-cent acknowledgement that policies and quality of care and to provide incentives to improve strategies should be put in place to address the quality, including regulations pertaining to receipt need for for-mal LTC provision (Ismail and of payment, staffing requirements, and the Hussein 2021, Hus-sein and Ismail 2017). Recent monitoring and enforcement of care standards developments such as cash-for-care schemes and (Grabowski 2008). Simultaneously, public reporting social assistance schemes are a result of these of health care quality data can inform consumer discussions. These efforts are framed mainly choice. For example, Nursing Home Compare is a within the traditional family-based care structure recognized resource that reports the ratings of with the goal of enabling the family to continue nursing homes in the US; however, it has been caring for older relatives. For example, Türkiye subject to scrutiny due to reliance on self-reported has developed plans to address LTC within data. As is the case in many European countries as the context of maintaining the family’s central well as in the US, public reporting mechanisms are role and ensuring care delivery at home and in the more often focused on institutional care. COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E C H A P T E R 11 363 Another critical function of the government in China, there is an urgent need for the government within LTC systems is to build and sustain work- to increase investment substantially in education and force development. The success of LTC systems is, to training programs not only to upgrade the skills of a large extent, dependent on the capacity of the direct care workers on the frontline but also to workforce and therefore investment in enhancing produce a cadre of multidisciplinary professionals, the skills of the workforce is imperative. The lack of such as geriatricians, pharmacists, therapists, dieti- a qualified and professional workforce in LTC is a cians, nurses, social workers, case managers, and LTC common challenge across all LMICs. As is the case facility administrators (Feng et al. 2020). DISCUSSION AND POLICY CONSIDERATIONS Through presented analysis and review, we identify several prominent findings and themes concerning the current LTC landscape across LMICs. Below, we discuss them and suggest potential policy options for LTC policymakers in these countries to consider. First, there is evidence of large gaps between ris- difference though is that most developed coun- ing LTC needs on the demand side and the lack tries provide some form of retirement incomes to of formal LTC provision on the supply side across finance elderly care and have policies to support LMICs, and it is imperative to increase govern- caregivers as well as home and community-based ment policy interventions to address these gaps. services to alleviate part of the caregiver burden. The prevalence of having difficulty with at least one In some developed countries, such as Germany, ADL or IADL and therefore needing help varies family caregiving has been integrated into the LTC across LMICs and MICs countries considered here, insurance system. Family caregiving is undeniably but nowhere is it below 15 percent of the 65 and an unreplaceable and effective method of caring older group. As data for more LMICs become avail- for the elderly but LMICs must adopt policies to able, future research should continue investigating support caregivers so that family caregivers can the demand for LTC to better understand the needs continue to fulfill this role without having to and the gaps in low-income settings. It is evident sacrifice their own needs. Policymakers in LMICs from the presented analysis, however, that LTC should develop and implement policies to increase pro-vision in LMICs has continually relied on the formal LTC service provision and support when family as primary caregivers. On a broad scale, this family care is insufficient or absent. has made sense and continues to make sense, given Because the provision of elder care by family the lack of formal LTC infrastructure and support members is a feature of the intergenerational social avail-able to the majority of older people and the contract deeply rooted in cultural values in most fact that family members know their loved ones and LMICs, public policies should not discourage it. their needs best. Some countries, including Algeria, Public policies should provide meaningful and Argentina, Brazil, Chile, China, India, Mexico, affordable options to families so that they can decide Russia, and Türkiye have gone as far as legalizing whether elderly care will be supplied within the family and man-dating adult children’s responsibility of or whether they will seek it in the market. This is taking care of their elderly parents (Scheil-Adlung especially important for the middle class because the 2015). The issue with such mandates is that they do poor can typically access last-resort and safety net not always take into account the capacity of the programs, while the rich can afford to pay for high- family caregivers and may result in caregiver end services. The middle-class, which represents the burden in the form of physical and emotional majority of the population, must often contend with health, social life, and financial status, particularly the “missing market” of affor-dable quality care. The for women. The enforce-ability of such laws is also absence of affordable care can lead to spending down questionable (Feng 2019). savings, the loss of market wages for care-givers, and The centrality of family caregiving is not exclusive an increased need for medical care for physical and to LMICs. Family caregiving is just as widespread in mental illnesses for both the caregivers and care developed countries (European Commission and recipients. This is why having a LTC market with a Zigante 2018), and indeed, it is an integral component continuum of care is important. Furthermore, of any LTC system (Villalobos Dintrans 2020). The government involvement in elderly care is necessary, 364 C H A P T E R 11 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E primarily through stewardship, including support to facilities in Viet Nam. Another concept is utilizing qualified service providers and subsidies to eligible government-owned buildings and other fixed assets care recipients. The actual provision of care can come to develop concessional arrangements with private from a combination of the private formal and semi- providers contracted to operate them as aged care formal sectors as well as from family members. facilities. Providers would set prices, establish fees, Ultimately, a hybrid system of informal and and collect revenues from their clients. The formal LTC services would be ideal for LMICs. government will use concession fees to in-fluence the While informal caregiving can have negative effects pricing policies of these providers. These concessions on the labor market, the income-generating capacity can be leveraged to develop government assets as of families, and the health of caregivers, informal institutional care facilities with nursing, re- caregivers also help contain the formal LTC system’s habilitation, palliative, or/and hospice services, which costs (Villalobos Dintrans 2020). When designing the may also provide day services and home-based care proper mix of formal and informal elements of LTC for nearby communities. This model is widely used provision, policy-makers should seek information for in Singapore and is increasingly visible in China. guidance by approaching questions such as: what Because the bricks and mortars of these facilities are kinds of needs exist in the population and what under the firm control of government ownership, kinds of services are in demand; how the system can this model has one distinct advantage of forestalling take into account families’ preferences; how much accurate estimate speculation for profiteering which money is available to finance the system; and how is a longstanding concern in countries, such as the the design of the LTC system affects caregivers’ US, where private equity firms are actively involved behaviors (Villalobos Dintrans 2020). Policy makers in LTC facility ownership (Braun et al. 2021). in LMICs should begin with a review of needs before Given the increasing role of the private sector in developing a clear agenda of priorities and strive to the LTC system, policymakers in LMICs should be create a balanced LTC system of mixed services that wary of the possible risk and unintended con- reflects older people’s preferences and needs. sequences coming with it. A trend observed in the Second, it is necessary to engage the private US and in many European countries is the sector in developing LTC services, markets, and privatization of LTC provision (Bergman et al. 2016, delivery systems in LMICs; at the same time, it is Stolt, Blomqvist, and Winblad 2011, Polivka and essential to strengthen government stewardship for Luo 2019), particularly in the nursing home clear guidance on rules of engagement as well as industry. Privatization involves the move to private quality assurance and the regulatory capacity to (most of-ten for-profit) delivery of services and enforce them. Most countries, including HICs and managerial practices, as well as an increasing LMICs, have opted to contract out LTC services fully emphasis on “responsibilization”—a transfer of care or partially to private non-profit, or for-profit orga- responsibilities to individuals and their families nizations. The global trend is a move away from (Armstrong, Arm-strong, and Bourgeault 2020). direct public provision of services. Where public Privatization can have benefits, including prices funding is involved, the public-private partnership is a being driven competitively in a free market commonly used mechanism for the engagement of economy rather than controlled by the government, private sector service providers in the LTC system, as resulting in higher spending but also greater is the case in China. investment in healthcare, providing potential for We suggest two concepts of LTC delivery growth, development, and improved qual-ity models for policymakers in LMICs to consider, (Donaldson 2018). The drawbacks of privatization which are expected to alleviate financing and supply can include higher costs, diminished access, less constraints for service provision. One is opening efficiency, the potential for lower quality of care, and existing public welfare homes and residential aged loss of public control over vital services (Polivka and care facilities to self-paying individuals and Luo 2019). Thus, it is crucial for the government to providing them with enhanced quality services. retain the supervising and stewardship responsibility These public facilities will continue to serve their while delegating the actual service provision and traditional clients (welfare recipients) and receive production activity to the private sector. government support to do so while strengthening Developing effective regulatory frameworks their revenue base through cross-subsidization from takes time and resources. In LMICs, regulatory private-paying clients. This model is widely used in oversight is often lacking and constitutes a low pri- China and has recently been implemented in some ority on the policy agenda. While the LTC infra- COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E C H A P T E R 11 365 structure and markets are emerging and growing the following steps and strategies. fast in some countries like China, regulation is lag- In the initial stages, financing from general tax ging behind. The development of regulatory frame- revenue is sensible. These public funds should first works can be helped with increased public financ- finance programs that provide a safety net for older ing of LTC services whereby the government, with persons in need and programs for disease preven- a greater stake in the reimbursement system, can be tion and healthy aging. Once resources increase for better positioned to demand regulatory compliance the financing of services to help older adults need- and quality assurance from providers (Feng 2019). ing assistance with ADLs and IADLs, these resourc- By playing an active role in purchasing services and es should be targeted at those with greater needs allowing the private sector to be service providers, in terms of frailty and disability. Income consider- the government ensures that the private sector has ations are also important, prioritizing those with some stake in ensuring quality of services as well. lower income. Gradually, eligibility expansion can Third, the lack of organized financing is a lead to coverage expansion. major impediment to both the development of An important issue in developing a robust pub- and access to affordable LTC services in LMICs; lic LTC financing system is the division of responsi- policymakers in these countries should take steps bilities between the central and local governments. aimed at establishing a systematic approach to A centralized system is best for ensuring horizontal public LTC financing, ideally following a broad- equity but can be less responsive to local condi- based social insurance model. In many LMICs, tions, needs, and traditions. A more decentralized new forms of elder care options outside the familial approach risks creating or institutionalizing dispar- sphere are emerging in the public sector and private ities across regions and individuals and may result market, but both access to and affordability for such in inefficiencies as each geographic unit designs and options remain limited, largely due to the absence manages its own system. On balance, a co-sharing of organized financing, particularly public financing arrangement between the national and local author- for LTC. As limited public support is available, in ities is likely to prove most efficient and equitable. virtually all LMICs eligibility is narrowly defined and On the supply side, public financing for the typically restricted to a small number of public wel- LTC system will also require supporting providers fare recipients who are poor and disabled, without through subsidies. Furthermore, procuring their family or otherwise among society’s most vulnerable services through competitive bid processes will members, through strict means-tests. As such, insuf- greatly improve the efficiency of both the overall ficient public funding for LTC results in high private system and of public expenditures. In designing expenditure in the form of out-of-pocket payments, financing of service providers, explicit linkages be- access gaps, and inequalities (Scheil-Adlung 2015). tween financing and delivery models—in-home, When considering financing options, policy community-based, and institutional—are critical. makers in LMICs should be aware that private or The gold standard is aging in place, but many coun- commercial LTC insurance plays a minimal (or tries face enormous residual challenges in contain- supplemental at best) role even in HICs with mature ing public expenditures because they initially fund- LTC systems, so this is unlikely to be a viable option. ed residential care facilities and beds, which heavily Instead, they should aim to develop a systematic ap- skewed their LTC systems toward institutional care proach to public LTC financing, ideally following a rather than home or community-based services. broad-based, universal-coverage social insurance Demand-side subsidies (e.g., vouchers and cash model similar to those adopted in Germany, Japan, allowances) are effective instruments for increasing and Korea (Feng and Glinskaya 2020). China’s LTC purchasing capacity of the poor and those with insurance pilots will also offer important lessons as greater needs. They are also compatible with pro- they develop. As with health care and income sup- moting aging in place, for example, if vouchers can port for older people, this approach recognizes that be redeemed for home- and community-based care. LTC needs represent a normal life risk that the vast Financial incentives to promote aging in place can majority of people cannot address on their own. also include subsidies for informal care provision The guiding principles for building towards such (e.g., respite care or vouchers that can be redeemed a comprehensive financing model are equity, fiscal by family caregivers). Overall, the role of demand- sustainability, and efficiency. Based on these princi- and supply-side financing must be balanced. Inter- ples, policy makers in LMICs could consider taking national experiences show that to develop an effi- 366 C H A P T E R 11 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E cient and equitable LTC market both are needed. sures to improve recruitment and retention in the The principle of “money following people” should LTC sector, including publicly funded training, in- be followed when channeling financing to provid- creases in wages and benefits, and improvements in ers, rather than subsidizing land, utilities, or beds. working conditions (Colombo et al. 2011). In Japan, LTC financing models should also encourage for example, providers receive subsidies for intro- integration in service delivery across health and so- ducing LTC equipment, such as lifts, that promote cial services. It would be premature to include LTC welfare and reduce the burden of care workers (Co- in the package of basic health services because its lombo et al. 2011). coverage remains shallow in most LMICs. Instead, Supporting informal caregivers. Monetary com- it is more realistic to initiate the development of pensation or financial support to informal caregiv- a basic package of LTC services, and pilot it with ers recognizes the value of informal care. Still, it external financing so as to develop a robust under- should not be the only policy for caregivers as total standing of the implied costs and benefits. reliance on informal caregiving may hinder profes- In addition, cost sharing and private spending sionalism or incentivize the creation of unregulated are important components of sustainable LTC fi- markets (Villalobos Dintrans 2020). Instead, these nancing in all countries. Private spending can play benefits should be combined with in-kind services, two important roles—providing additional resourc- such as respite care and labor policies, to allow es to publicly funded services and growing the mar- the participation of caregivers in the formal labor ket share served by the private sector. Both should market (Villalobos Dintrans 2020). Supporting be encouraged by the government. and increasing the capacity of informal caregivers Lastly, drawing on the experiences of HICs, should be a priority to address inequality and the policy makers in LMICs should consider multi- disproportionate burden on women (World Health pronged strategies to build and strengthen the Organization 2020a). The need for supporting fam- LTC workforce and to support family caregivers. ily caregivers for persons living with dementia is Some examples are discussed below. particularly acute, given the rapid increase of older Enacting policies to support training, including people with Alzheimer’s disease and related disor- creating career pathways to allow workers to advance ders in LMICs. to positions of increased responsibility and higher We conclude with a cautionary note that there wages. Many states in the US have implemented dif- never exists an ideal LTC system that works in all ferent workforce development policies tied to wage countries. However, to address impending aging increases, including increasing wages if the worker and LTC policy challenges, there is much for policy obtains a specific number of hours of training and makers in LMICs to learn from the LTC systems works in the field for a certain number of years. By in HICs. The World Health Organization has re- giving them added responsibility and autonomy, cently promoted an integrated continuum of LTC workers may be motivated to remain in the job or framework to guide all countries toward establish- encourage others to seek these positions (Stone and ing person-centered, primary health care driven, Wiener 2001). Germany has recently increased pub- and integrated delivery systems encompassing the lic funding for the third and last year of training of full range of acute, post-acute, and LTC services in older workers who want to change career into LTC, a continuum (World Health Organization 2021a). where it previously only paid for the first two years The care continuum should also include pallia- (Colombo et al. 2011). In Japan, LTC training is free tive care and end-of-life care. Person-centeredness for job seekers and is organized through the Public implies and requires the delivery of care and ser- Employment Services (“Hello Work”). It includes vices at or near the place where older people live, training at specialized private institutions or training consistent with their preference for aging in place. schools. Even though LTC trainees constitute around At present, such integrated LTC service provision 10 percent of total trainees, their employment in the and delivery models largely remain a concept in LTC sector is high (Colombo et al. 2011). most LMICs. Nevertheless, aided by cross-coun- Improving recruitment and retention of LTC try learning and a growing knowledge base from workers. Issues with staffing shortages and high international LTC research, leapfrogging can be turnover are not unique to LMICs. Many OECD achievable through the adoption and adaptation of countries have developed and implemented mea- international best practices. COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E C H A P T E R 11 367 REFERENCES 1. Aboderin, Isabella. 2019. “Toward a Fit-for-purpose Policy Ar- idad, Jeffrey Domingo, et al. 2021. “Dementia Incidence, Bur- chitecture on Long-term Care in Sub-Saharan Africa: Impasse den and Cost of Care: A Filipino Community-Based Study.” and a Research Agenda to Overcome it.” Journal of Long-Term Frontiers in Public Health 9 (May). https://doi.org/10.3389/ Care:119–126. fpubh.2021.628700. 2. Al-Janabi, Hareth, Fiona Carmichael, and Jan Oyebode. 2018. “In- 16. Donaldson, Benjamin S. 2018. “The Financial Ethics of Health- formal care: choice or constraint?” Scandinavian Journal of Caring care Privatization: United States vs. the United Kingdom.” Moral Sciences, 32(1), 157–167. https://doi.org/10.1111/scs.12441 Cents 7 (2):13-26. 3. Allotey, Pascale, Tamzyn Davey, and Daniel D. Reidpath. 2014. 17. Dyer, Suzanne M., Madeleine Valeri, Nimita Arora, Tyler Ross, Me- “NCDs in low and middle-income countries - assessing the ca- gan Winsall, Dominic Tilden, and Maria Crotty. 2019. Review of pacity of health systems to respond to population needs.” BMC International Systems for Long-Term Care of Older People: Report Public Health 14 Suppl 2:S1. doi: 10.1186/1471-2458-14-S2-S1. prepared for the Royal Commission into Aged Care Quality and 4. Armstrong, Pat, Hugh Armstrong, and Ivy Bourgeault. 2020. Safety: Research Paper 2. Adelaide, Australia: Flinders University. “Privatization and COVID-19: A Deadly Combination for 18. Edemekong, Peter F., Deb L. Bomgaars, Sukesh Sukumaran, Nursing Homes.” In Vulnerable: The Policy, Law and Ethics of and Shoshana B. Levy. 2020. “Activities of Daily Living.” PubMed. COVID-19, edited by Colleen Flood, Vanessa MacDonnell, Jane Treasure Island (FL): StatPearls Publishing. 2020. https:// Philpott, Sophie Thériault and Sridhar Venkatapuram, 447-462. pubmed.ncbi.nlm.nih.gov/29261878/. University of Ottawa Press. 19. European Commission 2019 Joint Report on Health Care and 5. Asian Development Bank (ADB). 2020. Country Diagnostic Long-Term Care Systems and Fiscal Sustainability – Country Doc- Study on Long-Term Care in Thailand. uments 2019 Update . https://economy-finance.ec.europa.eu/ 6. _____. 2021. Country Diagnostic Study on Long-Term Care in publications/joint-report-health-care-and-long-term-care-sys- Indonesia. tems-and-fiscal-sustainability-country-documents-2019_en. 7. Bergman, Mats A., Per Johansson, Sofia Lundberg, and Giancar- 20. European Commission. 2021. Long-term care report : trends, lo Spagnolo. 2016. “Privatization and quality: Evidence from el- challenges and opportunities in an ageing society. Volume I. Lux- derly care in Sweden.” J Health Econ 49:109-19. doi: 10.1016/j. embourg: Publications Office. jhealeco.2016.06.010. 21. European Commission, and V Zigante. 2018. Informal care in 8. Block, Miguel Angel, Adetokunbo Lucas, Octavio Gomez-Dan- Europe: Exploring formalisation, availability and quality. Euro- tes, and Julio Frenk. 2009. “3.3 Health policy in developing pean Commission, Directorate-General for Employment, Social countries.” In Oxford Textbook of Public Health, edited by Roger Affairs and Inclusion. Detels, Robert Beaglehole, Mary Ann Lansang and Martin Gul- 22. Feng, Quishi, Zhihong Zhen, Danan Gu, Bei Wu, Pamela W. liford. Oxford University Press. Duncan, and Jama L. Purser. 2013. “Trends in ADL and IADL dis- 9. Braun, Robert Tyler, Hye-Young Jung, Lawrence P. Casalino, ability in community-dwelling older adults in Shanghai, China, Zachary Myslinski, and Mark Aaron Unruh. 2021. “Association 1998-2008.” J Gerontol B Psychol Sci Soc Sci 68 (3):476-85. doi: of Private Equity Investment in US Nursing Homes With the 10.1093/geronb/gbt012. Quality and Cost of Care for Long-Stay Residents.” JAMA Health 23. Feng, Zhanlian. 2017. “Filial piety and old-age support in China: Forum 2 (11):e213817-e213817. doi: 10.1001/jamahealthfo- tradition, continuity, and change.” In Handbook on the Family rum.2021.3817. and Marriage in China, edited by Xiaowei Zang and Lucy Xia 10. Calvo, Esteban, Maureen Berho, Mónica Roqué, Juan Sebastián Zhao, 266-285. Edward Elgar Publishing. Amaro, Fernando Morales-Martínez, Emiliana Rivera-Meza, L 24. Feng, Zhanlian. 2019. “Global Convergence: Aging and Long- M Robledo, Elizabeth Caro López, Bernardita Canals, and Rosa Term Care Policy Challenges in the Developing World.” J Aging Kornfeld. 2018. 2019. “Comparative analysis of aging policy re- Soc Policy 31 (4):291-297. doi: 10.1080/08959420.2019.1626205. forms in Argentina, Chile, Costa Rica, and Mexico.” J Aging Soc 25. Feng, Zhanlian, and Elena Glinskaya. 2020. “Aiming Higher: Ad- Policy 31 (3):211-233. doi: 10.1080/08959420.2018.1465797. vancing Public Social Insurance for Longterm Care to Meet the 11. Carretero, Stephanie, Jorge Garcés, Francisco Ródenas, and Global Aging Challenge Comment on “Financing Long-term Vicente Sanjosé. 2009. “The Informal Caregiver’s Burden of De- Care: Lessons From Japan”.” Int J Health Policy Manag 9 (8):356- pendent People: Theory and Empirical Review.” Archives of Ger- 359. doi: 10.15171/ijhpm.2019.121. ontology and Geriatrics 49 (1): 74–79. https://doi.org/10.1016/j. 26. Feng, Zhanlian, Elena Glinskaya, Hongtu Chen, Sen Gong, Yue archger.2008.05.004. Qiu, Jianming Xu, and Winnie Yip. 2020. “Long-term care sys- 12. Chauhan, Shekhar, Shubham Kumar, Rupam Bharti, and Ratna tem for older adults in China: policy landscape, challenges, and Patel. 2022. “Prevalence and determinants of activity of daily future prospects.” The Lancet 396 (10259):1362-1372. living and instrumental activity of daily living among elderly in 27. Feng, Zhanlian, Chang Liu, Xinping Guan, and Vincent Mor. India.” BMC Geriatr 22 (1):64. doi: 10.1186/s12877-021-02659-z. 2012. “China’s rapidly aging population creates policy chal- 13. Clark, Robert L., Richard V. Burkhauser, Marilyn Moon, Joseph lenges in shaping a viable long-term care system.” Health Aff F. Quinn, and Timothy Smeeding. 2004. The Economics of an (Millwood) 31 (12):2764-73. doi: 10.1377/hlthaff.2012.0535. Aging Society. Wiley-Blackwell. 28. Ferri, Cluesa P., and K. S. Jacob. 2017. “Dementia in low-income 14. Colombo, Francesca, Ana Llena-Nozal, Jérôme Mercier, and Frits and middle-income countries: Different realities mandate tai- Tjadens. 2011. “Help Wanted? Providing and Paying for Long- lored solutions.” PLoS Med 14 (3):e1002271. doi: 10.1371/jour- Term Care.” OECD Health Policy Studies. Paris: OECD Publishing. nal.pmed.1002271. 15. Dominguez, Jacqueline, Leo Jiloca, Krizelle Cleo Fowler, Ma. 29. Giacomin, Karla Cristina, Paulo José Fortes Villa Boas, Marisa Fe De Guzman, Jhozel Kim Dominguez-Awao, Boots Nativ- Accioly Rodrigues da Costa Domingues, and Patrick Alexan- 368 C H A P T E R 11 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E der Wachholz. 2021. “Caring throughout life: peculiarities of https://doi.org/10.1007/978-3-319-69892-2_847-1 long-term care for public policies without ageism.” Geriatr 47. Li, Chaofan, Rui Zhou, Nengliang Yao, Thomas Cornwell, and Gerontol Aging 15:e0210009. doi: https://doi.org/10.5327/ Shuangshuang Wang. 2020. “Health Care Utilization and Un- Z2447-21232021EDITESP. met Needs in Chinese Older Adults With Multimorbidity and 30. Grabowski, David C. 2008. “The market for long-term care ser- Functional Impairment.” J Am Med Dir Assoc 21 (6):806-810. vices.” Inquiry 45 (1):58-74. doi: 10.1016/j.jamda.2020.02.010. 31. Gutiérrez Robledo, Luis Miguel, Mariana López Ortega, and Vic- 48. Lloyd-Sherlock, Peter, Anne M. Pot, Siriphan Sasat, and Fernan- toria Eugenia Arango Lopera. 2012. “The State of Elder Care in do Morales-Martinez. 2017. “Volunteer provision of long-term Mexico.” Current Geriatrics Reports 1 (4):183-189. doi: 10.1007/ care for older people in Thailand and Costa Rica.” Bull World s13670-012-0028-z. Health Organ 95 (11):774-778. doi: 10.2471/BLT.16.187526. 32. Gyasi, Razaq M., and David R. Phillips. 2020. “Aging and the 49. Lloyd-Sherlock, Peter. 2014. “Beyond Neglect: Long-term Care Rising Burden of Noncommunicable Diseases in Sub-Saharan Research in Low and Middle Income Countries.” International Africa and other Low- and Middle-Income Countries: A Call Journal of Gerontology 8 (2):66-69. for Holistic Action.” Gerontologist 60 (5):806-811. doi: 10.1093/ 50. Matus-Lopez, Mauricio, and Alexander Chaverri-Carvajal. 2021. geront/gnz102. “Population with Long-Term Care Needs in Six Latin American 33. Harbishettar, Vijaykumar, Mahesh Gowda, Saraswati Tenagi, Countries: Estimation of Older Adults Who Need Help Perform- and Mina Chandra. 2021. “Regulation of Long-Term Care ing ADLs.” Int J Environ Res Public Health 18 (15). doi: 10.3390/ Homes for Older Adults in India.” Indian J Psychol Med 43 (5 ijerph18157935. Suppl):S88-S96. doi: 10.1177/02537176211021785. 51. _____. 2022. “Progress Toward Long-Term Care Protection in 34. HelpAge International. n.d. “Older People’s Association (OPAs): Latin America: A National Long-Term Care System in Costa Strengths and Key Factors for Sustainability and Replication.” Rica.” J Am Med Dir Assoc 23 (2):266-271. doi: 10.1016/j.jam- HelpAge Briefing. Thailand: HelpAge International. https://www. da.2021.06.021. helpage.org/silo/files/older-people-s-associations-in-asia--stren 52. Ministry of Civil Affairs. 2021. “Statistical Communique on the gths-and-key-factors-for-sustainability-and-replication.pdf. Development of Social Services: 2020.” National Bureau of Sta- 35. Holmes, Wendy. 2021. Projecting the Need for and Cost of tistics of China. February 28, 2021. Long-Term Care for Older Persons. In ADB Sustainable Develop- 53. Ministry of Finance of Poland. 2021 “National Plan for Recon- ment Working Paper Series (No. 74): Asian Development Bank. struction and Resilience.” Government of Poland. April 30, 36. Hu, Jianhui. 2012. “Old-Age Disability in China: Implications for 2021. https://www.gov.pl/web/planodbudowy/kpo-wysla- Long-Term Care Policies in the Coming Decades.” Pardee RAND ny-do-komisji-europejskiej. Graduate School. 54. Mitra, Sohie, and Usha Sambamoorthi. 2014. “Disability prev- 37. Hussein, Shereen and Mohamed Ismail. 2017. “Ageing and Elder- alence among adults: estimates for 54 countries and progress ly Care in the Arab Region: Policy Challenges and Opportunities.” toward a global estimate.” Disabil Rehabil 36 (11):940-7. doi: Ageing Int 42 (3):274-289. doi: 10.1007/s12126-016-9244-8. 10.3109/09638288.2013.825333. 38. Hussein, Shereen. 2022. Ageing Demographics in the Middle 55. Mor, Vincent, Taziana Leone, and Anne Maresso (editors). 2014. East and North Africa: Policy Opportunities and Challenges. Regulating long term care quality: An international comparison. Report prepared for the World Bank. Cambridge, UK: Cambridge University Press. 39. Inter-American Development Bank (IDB). 2020a. Panorama of 56. Ndubuisi, Nweke E. 2021. “Noncommunicable Diseases Pre- Aging and Long-Term Care: Argentina.Washington, D.C.: In- vention In Low- and Middle-Income Countries: An Overview ter-American Development Bank. of Health in All Policies (HiAP).” Inquiry 58:46958020927885. 40. _____. 2020b. Panorama of Aging and Long-Term Care: Chile. doi: 10.1177/0046958020927885. Washington, D.C.: Inter-American Development Bank. 57. Organization for Economic Cooperation and Development 41. _____. 2020c. Panorama of Aging and Long-term Care: Costa (OECD). 2011. Mexico: Long-term Care. May 18, 2011. https:// Rica. Washington, D.C.: Inter-American Development Bank. www.oecd.org/mexico/47877877.pdf 42. Ismail, Mohamed, and Shereen Hussein. 2021. “An Evidence 58. Oyebode, Oyinlola, Ngianga-Bakwan Kandala, Peter J. Chilton, Review of Ageing, Long-Term Care Provision and Funding and Richard J. Lilford. 2016. “Use of traditional medicine in mid- Mechanisms in Turkey: Using Existing Evidence to Estimate dle-income countries: A WHO-SAGE study.” Health Policy and Long-Term Care Cost.” Sustainability 13 (11):6306. Planning, 31(8), 984-991. 43. Kazibwe, Joseph, Phuong B. Tran, and Kristi S. Annerstedt. 2021. 59. Petsoulas, Christina. 2019. “Thinking about Long-term Care in “The household financial burden of non-communicable diseas- a Global Context – A Literature Review.” Green Templeton Col- es in low- and middle-income countries: a systematic review.” lege, Oxford. Health Res Policy Syst 19 (1):96. doi: 10.1186/s12961-021-00732-y. 60. Polivka, Larry and Baozhen Luo. 2019. “Neoliberal Long-Term 44. Kedare, Jahnavi, and Chetan Vispute. 2016. “Research priorities Care: From Community to Corporate Control.” Gerontologist 59 for cognitive decline in India.” Journal of Geriatric Mental Health (2):222-229. doi: 10.1093/geront/gnx139. 3 (1):80-85. doi: 10.4103/2348-9995.181923. 61. Rajagopalan, Jayeeta, Saadiya Hurzuk, Faheem Arshad, Prem- 45. Lestari, Septi K., Nawi Ng, Paul Kowal, and Ailiana Santosa. kumar Raja, and Suvarna Alladi. 2020. “New country report: 2019. “Diversity in the Factors Associated with ADL-Related Dis- The COVID-19 Long-Term Care situation in India.” Interna- ability among Older People in Six Middle-Income Countries: A tional Long-Term Care Policy Network, CPEC-LSE, 30th May Cross-Country Comparison.” Int J Environ Res Public Health 16 2020. https://ltccovid.org/wp-content/uploads/2020/05/LTC- (8). doi: 10.3390/ijerph16081341. COVID-situation-in-India-30th-May-1.pdf. 46. Li, Jia, and Yajun Song. 2019. “Formal and Informal Care.” In 62. Scheil-Adlung, Xenia. 2015. “Long-term care protection for Encyclopedia of Gerontology and Population Aging, edited older persons: A review of coverage deficits in 46 countries.” by Danan Gu and Matthew E. Dupree, 1–8. Springer, Cham. In ILO Extension of Social Security Working Paper No. 50. Geneva, COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E C H A P T E R 11 369 Switzerland: International Labour Office. tivity limitations among older adults in 23 low and middle 63. Siop, Sidiah, Lois M. Verbrugge, and Tengku Aizan Hamid. 2008. income countries.” Sci Rep 10 (1):10442. doi: 10.1038/s41598- “Disability and Quality of Life Among Older Malaysians.” Pa- 020-67166-4. per presented at the 2008 Population Association of America 75. Wijesiri, H. S. Maliga S. K., Sudharshani Wasalathanthri, SH De (PAA) Annual Meeting, April 17-19, 2008. New Orleans. https:// Silva Weliange, and Chandrika N. Wijeyaratne. 2021 “The Preva- paa2008.populationassociation.org/papers/80962. lence and Correlates of Activity Limitations Among the Elderly 64. Stolt, Ragnar, Paula Blomqvist, and Ulrika Winblad. 2011. in Informal Caregiving Settings in Colombo District, Sri Lanka: “Privatization of social services: quality differences in Swed- A Community Based Cross-Sectional Study.” International Jour- ish elderly care.” Soc Sci Med 72 (4):560-7. doi: 10.1016/j. nal of Caring Sciences 13 (3):1568-1577. socscimed.2010.11.012. 76. World Bank. 2021. Technical report with an assessment of de- 65. Stone, Robyn I., and Joshua M. Wiener. 2001. “Who Will Care for mand for and supply of elderly care in Greece Part I: Overview Us? Addressing the Long-Term Care Workforce Crisis.” Washing- of demand for and supply of home-based care in Greece: an ton, D.C.: US Department of Health and Human Services. empirical investigation based on Survey of Health, Ageing and 66. Thakur, Rp, A. Banerjee, and Vb Nikumb. 2013. “Health prob- Retirement in Europe (SHARE). lems among the elderly: a cross-sectional study.” Ann Med 77. _____. 2022. Toward Developing a Long-Term Care Strategy Health Sci Res 3 (1):19-25. doi: 10.4103/2141-9248.109466. for Older People in Romania, mimeo. 67. United Nations Economic and Social Commission for Asia and 78. World Health Organization (WHO). 2017. Towards long-term the Pacific (UNESCAP). 2018. “Financing for Long-term Care care systems in sub-Saharan Africa. In WHO series on long-term in Asia and the Pacific.” Social Development Policy Brief No. care. Geneva, Switzerland: World Health Organization. 2018/01. Thailand: United Nations Economic and Social Com- 79. _____. 2020a. “Decade of Healthy Ageing 2021–2030.” Gene- mission for Asia and the Pacific. va, Switzerland: World Health Organization. https://www.who. 68. United Nations, Department of Economic and Social Affairs int/initiatives/decade-of-healthy-ageing. (UNDESA), Population Division. 2017. World Population Pros- 80. _____. 2020b. “Decade of Healthy Ageing: Baseline Report.” pects: The 2017 Revision. Geneva, Switzerland: World Health Organization. 69. United Nations. 2019. “World Population Prospects 2019.” Unit- 81. _____. 2021a. “Framework for Countries to Achieve an Inte- ed Nations, Department of Economic and Social Affairs, Popu- grated Continuum of Long-term Care.” Geneva, Switzerland: lation Division. https://population.un.org/wpp/. World Health Organization. 70. Van, Phan H., Khuong A. Tuan, and Tran T.M. Oanh. 2021. “Older 82. _____. 2021b. “Global Status Report on the Public Health Persons and Long-term Care in Viet Nam.” In Coping with Rapid Response to Dementia.” Geneva, Switzerland: World Health Population Ageing in Asia, edited by Osuke Komazawa and Ya- Organization. suhiko Saito. Jakarta: Economic Research Institute for ASEAN 83. World Health Organization, and World Bank. 2021. “Tracking and East Asia (ERIA). Universal Health Coverage 2021 Global Monitoring Report.” 71. Villalobos Dintrans, P. 2020. “Designing Long-Term Care Sys- CONFERENCE EDITION. tems: Elements to Consider.” J Aging Soc Policy 32 (1):83-99. 84. Zarzycki, Mikolaj, Val Morrison, Eva Bei,and Diane Seddon. doi: 10.1080/08959420.2019.1685356. 2022. “Cultural and Societal Motivations for Being Informal 72. Walker, Wendy, and Meredith Wyse. 2021. “Leadership and Gov- Caregivers: A Qualitative Systematic Review and Meta-synthe- ernance of Long-Term Care Systems in Asia and the Pacific”. ADB sis”. Health Psychology Review, 1–30. https://doi.org/10.1080/1 Brief No. 198. Manilla, Phillipines: Asian Development Bank (ADB). 7437199.2022.2032259. 73. Wang, Li-li. 2013. “A Study on the Demand,Supply and Utiliza- 85. Zhu, Haiyan. 2015. “Unmet needs in long-term care and their tion of Home-Based Care Services for the Elderly Based on the associated factors among the oldest old in China.” BMC Geriatr Theory of “Services Chain”.” Population Journal 35:49-59. 15:46. doi: 10.1186/s12877-015-0045-9. 74. Weber, Daniela and Sergei Scherbov. 2020. “Prospects of ac- 370 CHA P T ER 12 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E Social Protection and Jobs Chapter 12 Exploring the Labour Market Outcomes of the Risk Factors for Non-communicable Diseases: A Systematic Review Chapter 13 Productive Longevity: What Can Work in Low- and Middle-Income Countries? Chapter 14 Adequacy Pensions and Access to Health Care: Maintaining Human Capital during Old Age 12 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CH A P T ER 12 371 Exploring the labour market outcomes of the risk factors for non-communicable diseases A systematic review Debapriya Chakraborty a, Daphne Wu a, Prabhat Jha a a Centre for Global Health Research, Unity Health Toronto, University of Toronto, Canada 372 CHA P T ER 12 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E INTRODUCTION Globally, 41 million people die from non-communicable diseases (NCDs) every year (World Health Organiza- tion 2022a). This accounts for approximately 74 percent of all deaths (World Health Organization 2022a). Many NCDs, including hypertension, overweight/obesity, type 2 diabetes, and hyperlipidaemia, can be prevented by reducing the four major NCD risk factors — tobacco use, harmful use of alcohol, physical inactivity, and un- healthy diet (World Health Organization 2022b). NCDs not only increase the risk of premature mortality, but they also carry a tremendous financial burden. In the next two decades, NCDs are estimated to result in US$47 trillion output loss globally (Bloom et al. 2011). For example, as compared to people with healthy body weight, studies in the USA, Canada, and Australia have shown that those with obesity have 30 percent higher medical expenditures (Withrow and Alter 2011). In the USA, obesity has an annual cost of US$ 113.9 billion in direct medical costs and studies have shown that compared to children with healthy body weight, children with obesity incur US$ 19,000 incremental lifetime direct medical costs (Bleich et al. 2018; Finkelstein, Graham, and Malho- tra 2014). In the UK, cardiovascular diseases cost the healthcare system £ 7.4 billion a year and come with an overall economic burden of £ 15.8 billion per year (Public Health England 2019). NCDs disproportionately affect populations in low- ment. This review is tied to that agenda because it and middle-income countries (LMICs): of the total attempts to parse out the cause and effect of phe- deaths due to NCDs globally, about 77 percent oc- nomena that are, ultimately, interrelated. This work cur in LMICs (World Health Organization 2022a). builds on former studies that have looked at the dif- Using health expenditure data from 18 countries, ferent risk factors and their effects on labour mar- one study found that compared with households ket outcomes separately and attempts to bring them without NCDs, the risk difference for catastrophic together in a coherent narrative. spending among households with NCDs is 1.7 per- This is a unique piece of work because it looks at cent in lower-middle-income countries, 0.8 percent multiple risk factors of NCDs, their metabolic/phys- in upper-middle-income countries, and 7.5 percent iological consequences, and a wide range of labour in China (Murphy et al. 2020). NCDs also exacer- market outcomes, whereas previous studies have bate susceptibility to communicable diseases, as been focused on fewer exposures and outcomes. evidenced by the recent COVID-19 pandemic. This We also want to establish if there are any obvious leads to an increased burden on health systems that gender differences in the association of labour mar- are usually inadequate or fragmented in LMICs. ket outcomes with the known risk factors for NCDs In addition to the adverse health systems out- and comment on any policy implications that might comes, NCDs and their risk factors adversely affect result from the data derived. Therefore, this review the labour and employment sectors. Reduced work- focuses on some of the major behavioural risk fac- force participation, lower overall income, and ear- tors for NCDs (smoking, heavy alcohol consump- lier retirement are some labour market outcomes tion), key physiological changes resultant of the risk that have been observed. Productivity levels are low factors (overweight and obesity, hypertension, type and work absenteeism is high for employees suffer- 2 diabetes), major depressive disorder, and their ing from obesity and type 2 diabetes due to ensu- accompanying labour market outcomes by looking ing health issues (e.g., arthritis or fatigue) (Segal at the differential impacts on the population at all et al. 2021; Pedron et al. 2019). The Healthy Lon- life stages. Although preventing all-cause mortality gevity Initiative (HLI), based on the framework by is a critical public health issue, increasingly the fo- O’Keefe and Haldane (2024), is a strategy adopted cus must shift to a more fulsome approach, which is by the World Bank to take a lifecycle and whole of healthy longevity of the population. society approach to NCD prevention and manage- COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CH A P T ER 12 373 METHODS We conducted a systematic literature review using the Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) guidelines (Page et al. 2021). Search strategy longer). Studies focusing on the risk factors of obe- sity, such as sedentary behaviour or consumption Literature searches for all peer-reviewed studies of unhealthy foods and beverages, were excluded. published before July 2022 were performed in July Study designs included cohort and longitudinal, and 2022. The following databases were searched: MED- those excluded were systematic reviews, cross-sec- LINE (Ovid), Embase (Ovid), EconLit (EBSCO), tional studies, modelling studies, and qualitative EconPapers, and Cochrane Database of Systemat- studies not focused on intervention design or eval- ic Reviews. The search strategy included keywords uation. There were no restrictions placed on coun- relating to the adult population, labour market try, language of publication, race, or socioeconomic outcomes, and risk factors for non-communicable groups. To make it generalizable, we excluded stud- diseases. No limits were placed on language, loca- ies on specific populations such as patient popula- tion, or publication date. The search terms used for tions or populations in specific occupations. MEDLINE are included in the Annex. Selection process Eligibility criteria All titles and abstracts were screened in Covidence We studied labour market outcomes, such as em- by two independent reviewers (DC and DW) and ployment status, probability of being employed, conflicts were resolved after discussion. If there weeks in labour force, job tenure, job search du- were uncertainties, the paper(s) were included for ration, wages, productivity loss, disability pension, full text review. Full texts were reviewed by two in- early retirement, absenteeism, and sickness ab- dependent reviewers (DC and DW) and conflicts sence. The exposures of interest in this review are: were resolved by consensus. smoking (defined as having smoked more than 100 cigarettes or bidis during the lifetime; we excluded Data extraction and synthesis vaping, second-hand smoke and cigar smoking); binge alcohol consumption (defined as 5 or more Data were independently extracted from each study standard drinks on one occasion for men and 4 or by two reviewers (85% by DC and 15% by DW) in more standard drinks on one occasion for wom- Microsoft Excel. The key information retrieved was en (Centers for Disease Control and Prevention first author’s name, title of article, publication year, 2022)); heavy alcohol consumption (defined as 15 language of publication, country or setting, World or more standard drinks per week for men and 8 or Bank region and income group, period of surveil- more standard drinks per week for women (Cen- lance (number of years or actual years between ters for Disease Control and Prevention 2019)); which data was captured), study design, sample overweight (body mass index (BMI) > 25kg/m2) size, whether or not data was stratified by biological or obesity (BMI > 30kg/m2); hypertension (blood sex, whether socioeconomic status or ethnicity were pressure over 140mmHg/90mmHg; history of be- captured, exposures targeted, outcomes measured, ing diagnosed as hypertensive by a physician); type any biases or confounding factors identified, and 2 diabetes (history of being diagnosed with type 2 summary of the findings. The data extraction tool diabetes by a physician); and major depressive dis- was pilot tested on a small sample of sources before order (MDD) (diagnosed by a physician or experi- being revised and finalised for the data extraction encing persistent and intense symptoms of MDD process from each included source. Any discrepan- for a sustained period of time, usually 2 weeks or cies were resolved through discussion. 374 CHA P T ER 12 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E RESULTS The selection process resulted in 109 eligible studies that are included in this review. The majority of the studies are from Europe and Central Asia. The search results are discussed below. Studies included After screening five databases, 9,200 records were reviewing the full texts based on our inclusion and identified. Figure 12.1 shows the number of articles exclusion criteria, 131 were deemed eligible to be included and excluded in each step of the review. included in the review. After discussion within the After removing duplicates, 4,719 titles and abstracts review team (DC, DW, and PJ), studies published were screened. From these, 4,517 articles were ex- prior to 2000 were excluded due to concerns of suit- cluded, the majority of which examined NCD out- ability, and reviews were excluded because relevant comes of labour market exposures. 190 full texts studies within were captured in our scope or would were screened for eligibility (12 full texts were not be used in guiding the discussion of the results. 109 available through the institution libraries). After studies were included in this review. FIGURE 12.1 PRISMA flow diagram in search and selection of records COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CH A P T ER 12 375 Location of the studies All studies captured were published in English lan- one study was found from Latin America & the Ca- guage; 96.3 percent of them (105/109) were con- ribbean, while none were found from Middle East ducted in high-income countries and the rest (3.7% and North Africa, South Asia, and Sub-Saharan Afri- or 4/109) were conducted in upper-middle-income ca (Figure 12.2). A breakdown of the studies, by coun- countries. In terms of region, studies from Europe try, is provided in Supplementary Material, Table S1. and Central Asia constituted the highest proportion Eighty percent of the studies were conducted in only of studies (64%), followed by North America (25%), eight countries – USA, Sweden, Finland, Denmark, and East Asia and Pacific (10%) (Figure 12.2). Only Norway, Republic of Korea, UK, and Germany. FIGURE 12.2 Proportion of studies by World Bank Regions (World Bank 2022) Table 12.1 enumerates key demographic and econom- gadottir et al. 2019; Bockerman et al. 2018; Lang ic characteristics of the eight countries that reported and Nystedt 2018; Bockerman, Hyytinen, and Mac- the highest number of studies. They are all high-in- zulskij 2016, 2017; Ostby et al. 2016), but they were come countries with a current health expenditure, as collapsed into one broad category (Georgia State percentage of GDP, between 8 and 17 percent, and University 2022). Sixty-nine percent (75/109) of the universal health coverage (UHC) greater than 80 per- studies reported socioeconomic status of study par- cent. However, in these countries, the proportion of ticipants; ethnicity was reported by 24 percent. the population facing catastrophic health expendi- tures (spending on health greater than 10 percent of Number of studies on each exposure total household budget) ranges from 2 to 12 percent. and outcome Characteristics of included studies Tables 12.2 and 12.3 present the frequency of expo- sures and outcomes examined in the studies includ- Of the included studies, 64 percent (70/109) were ed. Many of the studies looked at a combination of longitudinal, followed by cohort studies (36/109 or exposures. High BMI (overweight or obesity) was 33%). Within the longitudinal studies, there were the most frequently studied (50/109 or 46%), fol- some twin studies (Laine and Hyytinen 2022; Hel- lowed by harmful alcohol use (29/109 or 27%). In 376 CHA P T ER 12 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E terms of labour market outcomes, income/earnings two studies (2%) reported that depression and (36/109 or 33%), sickness absence (35/109 or 32%), heavy episodic drinking have protective effects on and employment status (33/109 or 30%) were most labour market outcomes. However, the protective commonly studied. effects were not significant at 95% confidence. Table 12.4 presents a subset of the total num- Direction of outcome ber of studies looking at the combination (five or more observed) of the targeted exposures and the Out of the 109 studies included in this review, 100 different outcomes. The associations most frequent- (92%) reported adverse effects of the exposures, of ly studied were the effects of overweight/obesity on which 84 studies reported significant results with income/earnings (21/109 or 19%), followed by the p-value < 0.05 (Supplementary Material, Table S3). effects of overweight/obesity on sickness absence Seven studies (6%) did not report any detectable ef- (17/109 or 16%), and the effects of alcohol use on fects between the exposures and outcomes, while employment status (13/109 or 12%). TABLE 12.1  Total number of studies by countries + characteristics of 8 countries which produced the highest number of studies Country Number % Median year of Population GDP per Current Labour force Population with UHC Service of study and range (millions, capita, PPP health participation rate, household spending on Coverage studies 2021) (current in- expenditure female (% of female health greater than 10% Index (SDG (World ternational $, (% of GDP, population ages of total household budget 3.8.1) (2019, %) Bank 2021) (World 2019) (World 15+) (national (SDG 3.8.2, reported (World Health 2021c) Bank 2021a) Bank 2019) estimate, 2021) data) (%) (World Health Organization (World Bank 2021b) Organization 2022c) 2021) USA 25 22.9 2010 (2002-2018) 331.9 69,287.5 17 56 4 (2019) 83 Sweden 15 13.8 2015 (2001-2020) 10.4 59,324.0 11 63 6 (1996) 87 Finland 14 12.8 2017 (2002-2022) 5.5 55,006.6 9 57 7 (2016) 83 Denmark 9 8.3 2019 (2003-2021) 5.8 64,651.2 10 58 3 (2010) 85 Norway 9 8.3 2016 (2010-2021) 5.4 79,201.2 11 61 5 (1998) 86 Republic of Korea 7 6.4 2018 (2016-2020) 51.7 46,918.5 8 54 12 (2018) 87 UK 4 3.7 2009 (2007-2018) 67.3 49,675.3 10 59 2 (2018) 88 Germany 4 3.7 2014 (2013-2018) 83.1 57,927.6 12 56 2 (2010) 86 All others 22 20.2 TOTAL: 109 100 TABLE 12.2  Frequencies of risk factors stud- TABLE 12.3  Frequencies of outcomes identified in included studies ied in included studies Outcomes studied Number of studies = 109 Risk factors studied Number of studies = 109 Income/Earnings 36 Overweight or obesity 50 Sickness Absence 35 Alcohol use 29 Employment Status 33 Smoking 19 Pension 18 Depressive disorders 18 Other economic outcomes (work ability, job hunt duration, etc.) 11 Diabetes 17 Early retirement/Working life expectancy 9 Hypertension 3 Other social outcomes (education, marriage, etc.) 11 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CH A P T ER 12 377 TABLE 12.4  Frequency of studies associated with different combinations of exposures and outcomes Exposure Outcome Total number Number of studies that Number of studies that found no of studies found adverse effects or protective effects Income/Earnings 21 18 3 Sickness absence 17 16 1 Employment status 11 9 2 BMI Pension 7 7 0 Early retirement/Working life expectancy 5 5 0 Other economic outcomes (work ability, job hunt duration, etc.) 5 4 1 Employment status 13 12 1 Alcohol use Income/Earnings 8 7 1 Sickness absence 7 7 0 Smoking Sickness absence 6 6 0 Sickness absence 6 5 1 Depressive Employment status 5 5 0 disorders Pension 5 5 0 Diabetes Sickness absence 9 8 1 Impact of obesity Of the 21 studies looking at the association between Of them, 16 showed adverse effects and one showed BMI and income, 18 showed adverse effects and three either no or protective effects. In Belgium, high BMI showed either no significant or protective effects. was an independent predictor of sick leave – ranging Finnish women who are overweight earn 30% less (p from 31 percent (p < 0.001) increase for men to 47 < 0.05) in the long term and work fewer hours than percent (p < 0.001) increase for women (Moreau et those who were never overweight (Laine and Hyytin- al. 2004). In the UK, overweight and obesity were as- en 2022), and Taiwanese women with obesity earned sociated with a 13 percent (p < 0.05) and 51 percent up to 20 percent less (p < 0.05) when compared to (p < 0.05) increase, respectively, of long term sickness their healthy weight counterparts (Huang and Chen absence compared to healthy BMI (Ferrie et al. 2007). 2019). Swedish men with obesity incur an income In the USA, individuals with obesity experience a 4.6 penalty of 9.6 percent (p < 0.01) when compared percent (p < 0.001) higher rate of absenteeism com- to their healthy weight counterparts (Dackehag, pared with those with healthy weight (6.6 ± 1.1 sick Gerdtham, and Nordin 2015). We also found that the days with obesity vs. 3.2 ± 1.2 sick days with healthy impact of overweight and obesity on income differs weight) (VanWormer et al. 2012). In Germany, there by ethnicity. In the USA, Moreau et al. (2004) found was a significant (p < 0.01) gender difference in sick an 8 percent (white males) to 16 percent (Hispanic leave incidence (27% increase in women with obesity females) reduction in earnings with every two stan- compared to 15% decrease (not significant) in men), dard deviation increase in BMI (Slade 2017). Howev- and long-term absenteeism (41% increase in wom- er, amongst Black males, increase in BMI was found en with overweight compared to 16% decrease (not to be associated with an increase in wages earned significant) in men) (Reber, Konig, and Hajek 2018). (Slade 2017). In the Republic of Korea, the relation- Of the 11 studies that looked at the combination of ship between high BMI and income or earnings was BMI and employment status, nine showed adverse protective for men. Men with overweight or obesity results while two showed protective or no discernible earned 6.8 times higher monthly wages and were 1.46 effects. In Denmark, compared to those with healthy times more likely to be placed in professional jobs as weight, those with obesity and severe obesity were 18 compared to those without overweight or obesity and 27 percent, respectively, more likely to be unem- (Ahn, Kim, and Han 2019; H. Lee et al. 2019). ployed (Bramming et al. 2019). In Republic of Korea, Seventeen studies examined overweight/obesity employment probability decreased by 2.81 percent as the exposure and sickness absence as the outcome. (with job hunting experience) or 4.78 percent (with 378 CHA P T ER 12 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E no job-hunting experience) per week for women with on medically certified sick leave compared to their overweight and obesity (Ahn, Kim, and Han 2019). non-smoking, non-obese counterparts (Roos et al. All the included studies that looked at the relationship 2017). In Sweden, daily smoking was associated with a between BMI and pension, and between BMI and 30 percent long term sick listing (Vingard et al. 2005). early retirement or working life expectancy, showed adverse effects. There was 1.94 times to 3.45 times the Impact of depression probability of all-cause disability retirement associ- ated with severe obesity and obesity, respectively, in Of the six studies examining the association between Finland (Roos et al. 2013). In Sweden, the probability depressive disorders and sickness absence, five found of receiving disability pension increased from 1.36 adverse effects. In Norway, people with MDD were times (for overweight) to 3.04 times (for severe obe- 30 percent as likely to go on all-cause sick leave as sity) (M. Neovius, Kark, and Rasmussen 2008). In the those without MDD (Torvik et al. 2016). Five studies Republic of Korea, there was a 60 percent increased looked at the combination of depression and employ- risk of early exit from work due to unhealthy BMI (W. ment status and another five at depression and pen- Lee et al. 2018). Of the five studies that looked at BMI sion. All found adverse effects. In the Netherlands, and other economic outcomes (such as work ability, depression was associated with a higher risk of early job hunt duration, etc.), four found adverse effects. work exit (hazard ratios between 3.30 - 5.56) (De Breij et al. 2020). In Australia, depression leads to 25.6 per- Impact of harmful alcohol use cent - 34.2 percent increase in unemployment rates for men (Bubonya, Cobb-Clark, and Ribar 2019). In Of the 13 studies looking at the association of harmful the USA, those experiencing symptoms of depression alcohol use and employment status, 12 showed adverse are 60 percent more likely to face unemployment as effects. In Sweden, heavy alcohol use is associated with compared to those without depression (Whooley et a 28 percent - 48 percent increase in unemployment al. 2002). In Norway, depressed men and women are rates (Jorgensen et al. 2017). In Norway, heavy alcohol 2.4 to 2.9 times and 1.6 times, respectively, as likely to consumption was predictive of 17 percent - 32 per- go on disability pension as those without depression cent increased rate of unemployment (Kaspersen et al. (Lassemo et al. 2016; Knudsen et al. 2010). 2016). In Russia, a longitudinal study linked excessive drinking to up to three times the rate of unemploy- Impact of diabetes ment in working age men (Cook et al. 2014). Black men in the USA had a 47 percent lower probability Of the nine studies that looked at the association of of being employed in their mid-life if they were heavy type 2 diabetes and sickness absence, eight found drinkers (Sloan et al. 2011). Of the eight studies that adverse effects. In Denmark, men with diabetes looked at the association of alcohol use and income, (64%) were more likely to go on sick leave than seven found adverse effects. A Swedish cohort found women with diabetes (46%) when compared with that alcohol use disorder in men predicted a decrease those without diabetes (Nexo et al. 2020). In France of income by 0.24 standard deviations (Kendler et al. and Finland, there was a 33 percent increase in 2017). All seven studies that looked at the association work disability days for those with diabetes when between alcohol use and sickness absence showed ad- compared to those without (Virtanen et al. 2015). verse effects. In Finland, binge drinking was associat- ed with 21 percent - 62 percent increase in working Sex differences in association between days lost (Shiri et al. 2021; Ervasti et al. 2018). exposure and outcome Impact of smoking Most of the studies stratified their results by bio- logical sex (69/109 or 63%) (see Supplementary Of the eight studies looking at the association between Material, Table S2). There is evidence suggesting smoking and sickness absence, all found adverse ef- that biological sex may be an effect modifier in the fects. In Finland, smoking is associated with a 19 relationship between certain combinations of expo- percent (past smokers) to 30 percent (current smok- sures and outcomes. For example, of the 21 studies ers) increase in working days lost (Shiri et al. 2021). looking at the association between high BMI and Also in Finland, smoking and obesity together were income/earnings, 12 (or 57%) indicated overweight associated with a 2.23 and 2.69 times higher proba- and obesity affect income of females more adversely bility among women and men respectively, of going than that of males (Figure 12.3). COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CH A P T ER 12 379 FIGURE 12.3 Differences in outcomes identified by biological sex DISCUSSION Main findings Socioeconomic status (SES) and social determinants of health (SDHs) are strong predictors for NCDs (Marmot 2005; World Health Organization 2022d). Health equity is influenced by SDHs such as income and social pro- tection, unemployment and job insecurity, and working life conditions (World Health Organization 2022d). Generally, individuals in lower SES groups are more vulnerable to NCDs due to lack of monetary resources, which limits access to healthy and nutritious foods; lack of quality education that would teach people about health literacy and good dietary behaviours and the benefits of regular physical activity; and psychosocial factors arising from deprivation such as mental stress, control of life and finances, and social isolation (Marmot 2005). We were interested in exploring the contrasting reverse causality – how the incidence of risk factors and certain NCDs affect people’s labour market outcomes – to better understand the societal impacts of NCDs globally. This review includes several studies that have 2009), and the likelihood of obtaining disability shown statistically significant associations between pension (K. Neovius et al. 2008). There are a cou- risk factors of NCDs and common labour market ple of notable exceptions to this, especially amongst outcomes. We found that individuals exposed to black and Korean men where higher BMI was not overweight or obesity, diabetes, hypertension, de- penalised (Slade 2017; Ahn, Kim, and Han 2019; H. pressive disorders, excessive alcohol use, and ciga- Lee et al. 2019). Possible explanations for this are rettes are more likely to have lower rates of employ- that males with larger body sizes are idealised or ment, lower income, and higher rates of sickness less stigmatised compared to females. Additional- absence and disability pension. This association is ly, there are cultural and racial factors to consider corroborated by previous studies which found high where particular groups have history of widespread BMI to be a risk factor of low income/earnings poverty and therefore larger body sizes are associat- (Kim and Han 2017; Reiband, Heitmann, and So- ed with wealth (Slade 2017). rensen 2020), higher probability of unemployment Stigma could explain why those with over- (Reiband, Heitmann, and Sorensen 2020), greater weight or obesity earn less than those without. One sickness absence (Amiri and Behnezhad 2019; K. study found that women with persistent obesity Neovius et al. 2009; Schmier, Jones, and Halpern earned 20 percent less than their healthy weight 2006; Trogdon et al. 2008; van Duijvenbode et al. counterparts (Huang and Chen 2019). For exam- 380 CHA P T ER 12 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E ple, those with obesity are perceived to be undis- posures targeted and outcomes observed, which ciplined, unwilling to work, weak-willed, and lazy makes it a challenge to pool the results into a me- (Puhl and Heur 2009, 2010). This negative percep- ta-analysis. Causality between overweight/obesity tion is self-actualised in many instances after soci- and income/earnings were the most prevalent in ety has reinforced it with actions such as making the studies captured. Furthermore, women who suf- hiring decisions based off someone’s weight or their fered from overweight or obesity were more likely mental health diagnosis (Puhl and Brownell 2003; to earn less than men with overweight or obesity. Puhl and Heur 2009). Studies have shown that weight discrimination in the United States has sig- Strengths and limitations of this review nificantly increased (66%, p < 0.01) over a decade, from 7 in 1995/96 to 12 percent in 2004/06 (Andre- Some strengths of this study include the scale – it is yeva, Puhl, and Brownell 2008). A survey of over a global review with no limitations placed on lan- two thousand women with overweight or obesity guage or location, which casts a wide net and cap- found that 25 percent of them reported experienc- tures a myriad of NCD risk factors and labour mar- ing discrimination in their workplace due to their ket outcomes. Although no language restriction was weight (Puhl and Heur 2009). Further, 54 percent placed, only English articles were retrieved. This experienced stigma from their colleagues and 43 could be because of the databases that were used for percent experienced the same from their employers the search. Due to the heterogeneity of the studies, or supervisors (Puhl and Heur 2009). In the USA variability in study design, methodology, and differ- and other Western settings, studies have shown that ing effect size measurements, we were unable to un- differential treatment, derogatory humour and pe- dertake a formal meta-analysis. Additionally, there jorative comments, being passed over for promo- is the potential for publication bias since all stud- tions, or even being fired are some examples of ex- ies were in English and most of the studies (96.3%) periences of weight stigma in the workplace (Puhl were conducted in high-income country settings. and Brownell 2003; Puhl and Heur 2009). In gen- eral, cultivation theory suggests that thinner wom- Conclusions en are idealised more so than men, and this could explain the gender differences when it comes to the All the longitudinal or cohort studies included in stronger association for women with BMI and in- this review that looked at plausible causal relation- come/earnings than for men. ship between NCD risk factors and labour market Furthermore, smoking is a strong predictor outcomes were from high-income and upper-mid- of being a recipient of a disability pension (Amiri dle-income countries – particularly the USA; north- and Behnezhad 2020) and sickness-related absence ern European countries like Sweden, Finland, Den- (Weng, Ali, and Leonardi-Bee 2013), and depres- mark, and Norway; and South Korea – where there sion was shown to be associated with sickness ab- has been a long tradition of registry-based epidemi- sence (McIntyre, Liauw, and Taylor 2011). These ological studies. Based on these studies at least, we results indicate that prioritising NCD prevention find that individuals with overweight or obesity, di- should not only be a health sector priority but be abetes, hypertension, depressive disorders, excessive at the top of the agenda for the economic and fiscal alcohol use, and cigarette use are more likely to have growth sectors, as well. This is why the HLI frame- lower rates of employment, lower income, and high- work, which represents a multifactorial agenda of er rates of sickness absence and disability pension. healthy longevity, inclusive growth, and wellbeing Going forward, more studies done outside of the throughout a person’s life course through a com- context of high-income countries looking at how the bination of behavioural and environmental policy growing NCD burden is affecting the employment implementations, is key in prevention and effective sector, with particular attention to curbing the risk management of NCDs (O’Keefe and Haldane 2024). factors as a key prevention priority, are required. Ad- Even among the studies that displayed statis- ditionally, prevention of these adverse labour market tically significant results, there is variability in the outcomes requires multi-sectoral efforts. quality of the study and the combinations of ex- COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CH A P T ER 12 381 REFERENCES 1. Ahn, R., T. H. Kim, and E. Han. 2019. “The Moderation of Obesi- org/10.1111/add.12329. ty Penalty on Job Market Outcomes by Employment Efforts.” 15. Dackehag, M., U. G. Gerdtham, and M. Nordin. 2015. “Productiv- Int J Environ Res Public Health 16 (16). https://doi.org/10.3390/ ity or discrimination? An economic analysis of excess-weight ijerph16162974. penalty in the Swedish labor market.” Eur J Health Econ 16 (6): 2. Amiri, S., and S. Behnezhad. 2019. “Body mass index and risk of 589-601. https://doi.org/10.1007/s10198-014-0611-7. sick leave: A systematic review and meta-analysis.” Clin Obes 9 16. De Breij, S., J. Macken, J. Y. Qvist, D. Holman, M. Hess, M. Huis- (6): e12334. https://doi.org/10.1111/cob.12334. man, and D. J. H. Deeg. 2020. “Educational differences in the 3. -_____. 2020. “Smoking and disability pension: a systematic influence of health on early work exit among older workers.” review and meta-analysis.” Public Health 186: 297-303. https:// Occup Environ Med 77 (8): 568-575. https://doi.org/10.1136/ doi.org/10.1016/j.puhe.2020.04.013. oemed-2019-106253. 4. Andreyeva, Tatiana, Rebecca M. Puhl, and Kelly D. Brownell. 17. Ervasti, J., M. Kivimaki, J. Pentti, J. I. Halonen, J. Vahtera, and M. 2008. “Changes in Perceived Weight Discrimination Among Virtanen. 2018. “Changes in drinking as predictors of changes Americans, 1995–1996 Through 2004–2006.” OBESITY 16 (5): in sickness absence: a case-crossover study.” J Epidemiol Com- 1129-1134. https://doi.org/10.1038/oby.2008.35. munity Health 72 (1): 61-67. https://doi.org/10.1136/jech- 5. Bleich, Sara N, Kelsey A Vercammen, Laura Y Zatz, Johannah 2017-209777. M Frelier, Cara B Ebbeling, and AnnaPeeters. 2018. “Interven- 18. Ferrie, Jane E., Jenny Head, Martin J. Shipley, Jussi Vahtera, Mi- tions to prevent global childhood overweight and obesity: chael G. Marmot, and Mika Kivimäki. 2007. “BMI, Obesity, and a systematic review.” The Lancet Diabetes & Endocrinology 6 Sickness Absence in the Whitehall II Study.” OBESITY 15 (6): (4): 332-346. https://doi.org/https://doi.org/10.1016/S2213- 1554-1564. 8587(17)30358-3. 19. Finkelstein, Eric Andrew, Wan Chen Kang Graham, and Rahul 6. Bloom, D.E., E.T. Cafiero, E. Jané-Llopis, S. Abrahams-Gessel, L.R. Malhotra. 2014. “Lifetime Direct Medical Costs of Childhood Bloom, S. Fathima, A.B. Feigl, T. Gaziano, M. Mowafi, A. Pandya, Obesity.” Pediatrics 133 (5): 854-862. https://doi.org/10.1542/ K. Prettner, L. Rosenberg, B. Seligman, A.Z. Stein, and C. Wein- peds.2014-0063. stein. 2011. The Global Economic Burden of Noncommunicable 20. Georgia State University. 2022. “Literature Reviews: Types of Diseases. Geneva: World Economic Forum. Clinical Study Designs.” Last Modified March 11. Accessed No- 7. Bockerman, P., A. Hyytinen, J. Kaprio, and T. Maczulskij. 2018. “If vember 10, 2022. https://research.library.gsu.edu/litrev. you drink, don’t smoke: Joint associations between risky health 21. Helgadottir, B., L. Mather, J. Narusyte, A. Ropponen, V. Blom, behaviors and labor market outcomes.” Soc Sci Med 207: 55-63. and P. Svedberg. 2019. “Transitioning from sickness absence to https://doi.org/10.1016/j.socscimed.2018.04.039. disability pension-the impact of poor health behaviours: a pro- 8. Bockerman, P., A. Hyytinen, and T. Maczulskij. 2016. “Devil in spective Swedish twin cohort study.” BMJ Open 9 (11): e031889. disguise: Does drinking lead to a disability pension?” Prev Med https://doi.org/10.1136/bmjopen-2019-031889. 86: 130-5. https://doi.org/10.1016/j.ypmed.2016.03.008. 22. Huang, C. Y., and D. R. Chen. 2019. “Association of weight 9. ---. 2017. “Alcohol Consumption and Long-Term Labor Mar- change patterns in late adolescence with young adult wage ket Outcomes.” Health Econ 26 (3): 275-291. https://doi. differentials: A multilevel longitudinal study.” PLoS One 14 (7): org/10.1002/hec.3290. e0219123. https://doi.org/10.1371/journal.pone.0219123. 10. Bramming, M., M. B. Jorgensen, A. I. Christensen, C. J. Lau, K. 23. Jorgensen, M. B., L. C. Thygesen, U. Becker, and J. S. Tolstrup. K. Egan, and J. S. Tolstrup. 2019. “BMI and Labor Market Partic- 2017. “Alcohol consumption and risk of unemployment, sick- ipation: A Cohort Study of Transitions Between Work, Unem- ness absence and disability pension in Denmark: a prospec- ployment, and Sickness Absence.” Obesity (Silver Spring) 27 (10): tive cohort study.” Addiction 112 (10): 1754-1764. https://doi. 1703-1710. https://doi.org/10.1002/oby.22578. org/10.1111/add.13875. 11. Bubonya, M., D. A. Cobb-Clark, and D. C. Ribar. 2019. “The 24. Kaspersen, S. L., K. Pape, G. A. Vie, S. O. Ose, S. Krokstad, D. Gun- reciprocal relationship between depressive symptoms and nell, and J. H. Bjorngaard. 2016. “Health and unemployment: 14 employment status.” Econ Hum Biol 35: 96-106. https://doi. years of follow-up on job loss in the Norwegian HUNT Study.” org/10.1016/j.ehb.2019.05.002. Eur J Public Health 26 (2): 312-7. https://doi.org/10.1093/eur- 12. Centers for Disease Control and Prevention (CDC). 2019. “What pub/ckv224. is Excessive Alcohol Use?”. Last Modified December 30. Ac- 25. Kendler, K. S., H. Ohlsson, K. J. Karriker-Jaffe, J. Sundquist, and cessed November 8, 2022. https://www.cdc.gov/alcohol/ K. Sundquist. 2017. “Social and economic consequences of onlinemedia/infographics/excessive-alcohol-use.html#:~:- alcohol use disorder: a longitudinal cohort and co-relative text=Heavy%20drinking%3A%20For%20women%2C%20 analysis.” Psychol Med 47 (5): 925-935. https://doi.org/10.1017/ heavy,drinks%20or%20more%20per%20week. S0033291716003032. 13. _____. 2022. “Binge Drinking.” Last Modified January 6. Accessed 26. Kim, T, and E Han. 2017. “Entry Body Mass and Earnings: Once November 8, 2022. https://www.cdc.gov/alcohol/fact-sheets/ Penalized, Ever Penalized?” Biodemography Soc Biol 63 (4): 332- binge-drinking.htm#:~:text=Binge%20drinking%20is%20de- 346. https://doi.org/10.1080/19485565.2017.1403302. fined%20as,a%20severe%20alcohol%20use%20disorder. 27. Knudsen, A. K., S. Overland, H. F. Aakvaag, S. B. Harvey, M. Ho- 14. Cook, S., B. L. DeStavola, L. Saburova, and D. A. Leon. 2014. topf, and A. Mykletun. 2010. “Common mental disorders and “Acute alcohol-related dysfunction as a predictor of em- disability pension award: seven year follow-up of the HUSK ployment status in a longitudinal study of working-age study.” J Psychosom Res 69 (1): 59-67. https://doi.org/10.1016/j. men in Izhevsk, Russia.” Addiction 109 (1): 44-54. https://doi. jpsychores.2010.03.007. 382 CHA P T ER 12 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E 28. Laine, L. T., and A. Hyytinen. 2022. “Temporary and persistent work for impact pathways between non-communicable dis- overweight and long-term labor market outcomes.” Int J Health eases, human capital and healthy longevity, economic and Econ Manag 22 (2): 181-203. https://doi.org/10.1007/s10754- wellbeing outcomes,” in Unlocking the Power of Healthy Lon- 021-09315-4. gevity: Compendium of Research for the Healthy Longevity 29. Lang, E, and P Nystedt. 2018. “Blowing up money? The earnings Initiative. Washington D.C.: World Bank. penalty of smoking in the 1970s and the 21st century.” J Health 42. Ostby, K. A., N. Czajkowski, G. P. Knudsen, E. Ystrom, L. C. Gjerde, Econ 60: 39-52. https://doi.org/10.1016/j.jhealeco.2018.05.003. K. S. Kendler, R. E. Orstavik, and T. Reichborn-Kjennerud. 2016. 30. Lassemo, E., I. Sandanger, J. F. Nygard, and K. W. Sorgaard. 2016. “Does low alcohol use increase the risk of sickness absence? A “Predicting disability pension - depression as hazard: a 10 year discordant twin study.” BMC Public Health 16 (1): 825. https:// population-based cohort study in Norway.” Int J Methods Psy- doi.org/10.1186/s12889-016-3502-2. chiatr Res 25 (1): 12-21. https://doi.org/10.1002/mpr.1473. 43. Page, Matthew J, Joanne E McKenzie, Patrick M Bossuyt, Isa- 31. Lee, H., R. Ahn, T. H. Kim, and E. Han. 2019. “Impact of Obesity belle Boutron, Tammy C Hoffmann, Cynthia D Mulrow, Larissa on Employment and Wages among Young Adults: Observa- Shamseer, Jennifer M Tetzlaff, Elie A Akl, Sue E Brennan, Rog- tional Study with Panel Data.” Int J Environ Res Public Health 16 er Chou, Julie Glanville, Jeremy M Grimshaw, Asbjørn Hrób- (1). https://doi.org/10.3390/ijerph16010139. jartsson, Manoj M Lalu, Tianjing Li, Elizabeth W Loder, Evan 32. Lee, W., J. H. Yoon, J. W. Koo, S. J. Chang, J. Roh, and J. U. Won. Mayo-Wilson, Steve McDonald, Luke A McGuinness, Lesley A 2018. “Predictors and estimation of risk for early exit from work- Stewart, James Thomas, Andrea C Tricco, Vivian A Welch, Penny ing life by poor health among middle and older aged workers Whiting, and David Moher. 2021. “The PRISMA 2020 statement: in Korea.” Sci Rep 8 (1): 5180. https://doi.org/10.1038/s41598- an updated guideline for reporting systematic reviews.” BMJ 018-23523-y. 372: n71. https://doi.org/10.1136/bmj.n71. 33. Marmot, Michael. 2005. “Social determinants of health in- 44. Pedron, Sara, Karl Emmert-Fees, Michael Laxy, and Lars equalities.” The Lancet 365 (9464): 1099-1104. https://doi. Schwettmann. 2019. “The impact of diabetes on labour market org/10.1016/s0140-6736(05)71146-6. participation: a systematic review of results and methods.” BMC 34. McIntyre, Roger S., Samantha Liauw, and Valerie H. Taylor. 2011. Public Health 19 (25): 1-13. “Depression in the workforce: the intermediary effect of med- 45. Public Health England. 2019. “Health matters: preventing ical comorbidity.” Journal of Affective Disorders 128: S29-S36. cardiovascular disease.” Last Modified 14 February. Accessed https://doi.org/10.1016/s0165-0327(11)70006-4. November 8, 2022. https://www.gov.uk/government/publi- 35. Moreau, M., F. Valente, R. Mak, E. Pelfrene, P. de Smet, G. De cations/health-matters-preventing-cardiovascular-disease/ Backer, and M. Kornitzer. 2004. “Obesity, body fat distribu- health-matters-preventing-cardiovascular-disease. tion and incidence of sick leave in the Belgian workforce: the 46. Puhl, R. M., and K. D. Brownell. 2003. “Psychosocial origins of Belstress study.” Int J Obes Relat Metab Disord 28 (4): 574-82. obesity stigma: toward changing a powerful and pervasive https://doi.org/10.1038/sj.ijo.0802600. bias.” obesity reviews 4: 213-227. 36. Murphy, A., B. Palafox, M. Walli-Attaei, T. Powell-Jackson, S. Ran- 47. Puhl, R. M., and C. A. Heur. 2009. “The stigma of obesity: a re- garajan, K. F. Alhabib, A. J. Avezum, K. B. T. Calik, J. Chifamba, view and update.” Obesity (Silver Spring) 17 (5): 941-64. https:// T. Choudhury, G. Dagenais, A. L. Dans, R. Gupta, R. Iqbal, M. doi.org/10.1038/oby.2008.636. Kaur, R. Kelishadi, R. Khatib, I. M. Kruger, V. R. Kutty, S. A. Lear, 48. ---. 2010. “Obesity Stigma: Important Considerations for Pub- W. Li, P. Lopez-Jaramillo, V. Mohan, P. K. Mony, A. Orlandini, A. lic Health.” American Journal of Public Health 100, no. Framing Rosengren, I. Rosnah, P. Seron, K. Teo, L. A. Tse, L. Tsolekile, Y. Health Matters (6): 1019-1028. Wang, A. Wielgosz, R. Yan, K. E. Yeates, K. Yusoff, K. Zatonska, 49. Reber, K. C., H. H. Konig, and A. Hajek. 2018. “Obesity and sick- K. Hanson, S. Yusuf, and M. McKee. 2020. “The household eco- ness absence: results from a longitudinal nationally represen- nomic burden of non-communicable diseases in 18 countries.” tative sample from Germany.” BMJ Open 8 (6): e019839. https:// BMJ Glob Health 5 (2): e002040. https://doi.org/10.1136/bm- doi.org/10.1136/bmjopen-2017-019839. jgh-2019-002040. 50. Reiband, H. K., B. L. Heitmann, and T. I. A. Sorensen. 2020. “Ad- 37. Neovius, K., K. Johansson, M. Kark, and M. Neovius. 2009. “Obe- verse labour market impacts of childhood and adolescence sity status and sick leave: a systematic review.” Obes Rev 10 (1): overweight and obesity in Western societies-A literature 17-27. https://doi.org/10.1111/j.1467-789X.2008.00521.x. review.” Obes Rev 21 (8): e13026. https://doi.org/10.1111/ 38. Neovius, K., K. Johansson, S. Rossner, and M. Neovius. 2008. obr.13026. “Disability pension, employment and obesity status: a system- 51. Roos, E., M. Laaksonen, O. Rahkonen, E. Lahelma, and T. Lalluk- atic review.” Obes Rev 9 (6): 572-81. https://doi.org/10.1111/ ka. 2013. “Relative weight and disability retirement: a prospec- j.1467-789X.2008.00502.x. tive cohort study.” Scand J Work Environ Health 39 (3): 259-67. 39. Neovius, M., M. Kark, and F. Rasmussen. 2008. “Association be- https://doi.org/10.5271/sjweh.3328. tween obesity status in young adulthood and disability pen- 52. Roos, E., T. Lallukka, E. Lahelma, and O. Rahkonen. 2017. “The sion.” Int J Obes (Lond) 32 (8): 1319-26. https://doi.org/10.1038/ joint associations of smoking and obesity with subsequent ijo.2008.70. short and long sickness absence: a five year follow-up study 40. Nexo, M. A., J. Pedersen, B. Cleal, and J. B. Bjorner. 2020. “In- with register-linkage.” BMC Public Health 17 (1): 978. https://doi. creased risk of long-term sickness absence, lower rate of return org/10.1186/s12889-017-4997-x. to work and higher risk of disability pension among people 53. Schmier, J. K., M. L. Jones, and M. T. Halpern. 2006. “Cost of obe- with type 1 and type 2 diabetes mellitus: a Danish retrospec- sity in the workplace.” Scand J Work Environ Health 32 (1): 5-11. tive cohort study with up to 17 years’ follow-up.” Diabet Med 37 https://doi.org/10.5271/sjweh.970. (11): 1861-1865. https://doi.org/10.1111/dme.14203. 54. Segal, A. B., M. C. Huerta, E. Aurino, and F. Sassi. 2021. “The 41. O’Keefe, Philip, and Victoria Haldane. 2024. “Towards a frame- impact of childhood obesity on human capital in high-in- COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CH A P T ER 12 383 come countries: A systematic review.” Obes Rev 22 (1): e13104. rome H. Markovitz, Karen Matthews, and Stephen B. Hulley. https://doi.org/10.1111/obr.13104. 2002. “Depressive Symptoms, Unemployment, and Loss of 55. Shiri, R., A. Hiilamo, O. Rahkonen, S. J. W. Robroek, O. Pietilainen, Income: The CARDIA Study.” Arch Intern Med. 162: 2614-2620. and T. Lallukka. 2021. “Predictors of working days lost due to sick- 66. Withrow, D., and D. A. Alter. 2011. “The economic burden of ness absence and disability pension.” Int Arch Occup Environ Health obesity worldwide: a systematic review of the direct costs of 94 (5): 843-854. https://doi.org/10.1007/s00420-020-01630-6. obesity.” obesity reviews 12: 131–141. https://doi.org/10.1111/ 56. Slade, P. 2017. “Body mass and wages: New evidence from j.1467-789X.2009.00712.x. quantile estimation.” Econ Hum Biol 27 (Pt A): 223-240. https:// 67. World Bank. 2019. “Current health expenditure (% of GDP).” doi.org/10.1016/j.ehb.2017.07.001. Accessed November 4, 2022. https://data.worldbank.org/indi- 57. Sloan, F, P.R. Costanzo, D. Belsky, E. Holmberg, P.S. Malone, Y. cator/SH.XPD.CHEX.GD.ZS. Wang, and S. Kertesz. 2011. “Heavy drinking in early adulthood 68. _____. 2021a. “GDP per capita, PPP (current international $).” and outcomes at mid life.” J Epidemiol Community Health 65 (7): Accessed November 4, 2022. https://data.worldbank.org/indi- 600-5. https://doi.org/10.1136/jech.2009.102228. cator/NY.GDP.PCAP.PP.CD. 58. Torvik, F. A., T. Reichborn-Kjennerud, L. C. Gjerde, G. P. Knudsen, 69. _____.. 2021b. “Labor force participation rate, female (% of fe- E. Ystrom, K. Tambs, E. Roysamb, K. Ostby, and R. Orstavik. 2016. male population ages 15+) (national estimate).” Accessed No- “Mood, anxiety, and alcohol use disorders and later cause-spe- vember 4, 2022. https://data.worldbank.org/indicator/SL.TLF. cific sick leave in young adult employees.” BMC Public Health 15: CACT.FE.NE.ZS. 702. https://doi.org/10.1186/s12889-016-3427-9. 70. _____.. 2021c. “Population, total.” Accessed November 9, 2022. 59. Trogdon, J. G., E. A. Finkelstein, T. Hylands, P. S. Dellea, and S. https://data.worldbank.org/indicator/SP.POP.TOTL. J. Kamal-Bahl. 2008. “Indirect costs of obesity: a review of 71. _____. 2022. “The World by Income and Region.” World De- the current literature.” Obes Rev 9 (5): 489-500. https://doi. velopment Indicators. Accessed September 30, 2022. https:// org/10.1111/j.1467-789X.2008.00472.x. datatopics.worldbank.org/world-development-indicators/ 60. van Duijvenbode, D. C., M. J. Hoozemans, M. N. van Poppel, and the-world-by-income-and-region.html. K. I. Proper. 2009. “The relationship between overweight and 72. World Health Organization (WHO). 2021. “UHC Service Cov- obesity, and sick leave: a systematic review.” Int J Obes (Lond) 33 erage Index (SDG 3.8.1).” THE GLOBAL HEALTH OBSERVATORY. (8): 807-16. https://doi.org/10.1038/ijo.2009.121. Last Modified November 19. Accessed November 4, 2022. 61. VanWormer, J. J., J. A. Linde, L. J. Harnack, S. D. Stovitz, and R. https://www.who.int/data/gho/data/indicators/indicator-de- W. Jeffery. 2012. “Weight change and workplace absenteeism tails/GHO/uhc-index-of-service-coverage. in the HealthWorks study.” Obes Facts 5 (5): 745-52. https://doi. 73. _____.. 2022a. “Noncommunicable diseases.” Last Modified 16 org/10.1159/000345119. September. Accessed September 30, 2022. https://www.who. 62. Vingard, E., P. Lindberg, M. Josephson, M. Voss, B. Heijbel, L. Al- int/news-room/fact-sheets/detail/noncommunicable-diseas- fredsson, S. Stark, and A. Nygren. 2005. “Long-term sick-listing es#:~:text=Noncommunicable%20diseases%20(NCDs)%20 among women in the public sector and its associations with kill%2041,%2D%20and%20middle%2Dincome%20countries. age, social situation, lifestyle, and work factors: a three-year 74. _____. 2022b. “Noncommunicable diseases: Risk factors.” Ac- follow-up study.” Scand J Public Health 33 (5): 370-5. https://doi. cessed November 15, 2022. https://www.who.int/data/gho/ org/10.1080/14034940510005860. data/themes/topics/noncommunicable-diseases-risk-factors. 63. Virtanen, M., M. Kivimaki, M. Zins, R. Dray-Spira, T. Oksanen, J. E. 75. _____.. 2022c. “Population with household expenditures Ferrie, A. Okuloff, J. Pentti, J. Head, M. Goldberg, and J. Vahtera. on health greater than 10% of total household expenditure 2015. “Lifestyle-related risk factors and trajectories of work dis- or income (SDG 3.8.2) (%).” THE GLOBAL HEALTH OBSER- ability over 5 years in employees with diabetes: findings from VATORY. Last Modified September 13. Accessed November two prospective cohort studies.” Diabet Med 32 (10): 1335-41. 4, 2022. https://www.who.int/data/gho/data/indicators/ https://doi.org/10.1111/dme.12787. indicator-details/GHO/population-with-household-expendi- 64. Weng, S. F., S. Ali, and J. Leonardi-Bee. 2013. “Smoking and tures-on-health-greater-than-10-of-total-household-expendi- absence from work: systematic review and meta-analysis of ture-or-income-(sdg-3-8-2)-(-). occupational studies.” Addiction 108 (2): 307-19. https://doi. 76. _____. 2022d. “Social determinants of health.” Accessed No- org/10.1111/add.12015. vember 28. https://www.who.int/health-topics/social-deter- 65. Whooley, Mary A., Catarina I. Kiefe, Margaret A. Chesney, Je- minants-of-health#tab=tab_1. 384 CHA P T ER 12 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E ANNEX 12.1 Search strategy TABLE 12A.1 Search strategy for Ovid (MEDLINE) Sample Results: 1375, Database: All Ovid Medline <1946 – July 21, 2022> 1 exp alcohol-related disorders/ or alcoholism/ 119218 2 alcoholism.ti,ab,kf. 30918 3 alcoholic*.ti,ab,kf. 74549 4 exp drinking/ 14671 5 exp drinking behavior/ or exp alcohol drinking/ 82568 6 (body weight and measures).mp. [mp=title, abstract, original title, name of substance word, subject heading word, floating 16664 sub-heading word, keyword heading word, organism supplementary concept word, protocol supplementary concept word, rare disease supplementary concept word, unique identifier, synonyms] 7 body weight/ or overweight/ or obesity/ 391203 8 exp body fat distribution/ or exp body mass index/ or *body weight/ or *sagittal abdominal diameter/ or *waist circumfer- 182219 ence/ or *waist-height ratio/ or *body surface area/ or *waist-hip ratio/ 9 (obesit: or obese).ti,ab,kf. 356663 10 Weight Prejudice/ 113 11 exp Smoking/ 158043 12 (cigaret* or cigarette* or cigar* or pipe or smoke or smoking or tobacco).ti,ab,kf. 356387 13 exp Depressive Disorder/ 118465 14 (depressed or depression).ti,ab,kf. 455150 15 (“mood disorder:” adj3 depression).ti,ab,kf. 1442 16 exp Diabetes Mellitus/ 482733 17 (diabetes or diabetic).ti,ab,kf. 720699 18 or/1-17 2248946 19 exp economic status/ 429 20 Employment/ec, sn [Economics, Statistics & Numerical Data] 9285 21 personal earnings.tw. 23 22 Personal wealth.tw. 37 23 financial security.ti,ab,kf. 610 24 financial status.ti,ab,kf. 1222 25 *income/ or *pensions/ or *remuneration/ 13030 26 income.ti,ab,kf. 156936 27 salary.ti,ab,kf. 5086 28 net worth.af. 189 29 19 or 20 or 21 or 22 or 23 or 24 or 25 or 26 or 27 or 28 176716 30 18 and 29 37006 31 18 and 25 1375 13 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAP T ER 13 385 Productive Longevity What can work in low- and middle-income countries? Sara Johansson de Silva a and Indhira Santos b a Human Capital Project, World Bank b Human Development, Eastern Europe and Central Asia, World Bank 386 CHAP T ER 13 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E INTRODUCTION The world’s population is aging at a dramatic speed, raising concerns for how to maintain income and productiv- ity growth, improve the welfare of older people, and help them contribute to society. By 2050, one in six members of the world’s population will be at least 65 years of age, compared to fewer than one in 10 today. Nearly four out of five of these seniors will be living in what are currently low- and middle-income countries (LMICs) (United Na- tions 2019). Aging will imply shrinking working populations relative to dependents, which will reduce economic growth if other factors—particularly productivity and employment rates—do not change. Aging is also likely to require higher public and private expenditures for health and long-term care (LTC) services. These demographic tides prompt a need for policies that can sustain welfare levels and ensure that welfare is equitably distributed across generations and socioeconomic groups. Labor market policies that help extend productive working lives and increase the labor force participation and productivity of older workers – “productive longevity” – are part of this important agenda. This chapter identifies a set of labor market-related policies that can address key con- straints to productive longevity in LMICs and provides some preliminary evidence on their effectiveness. Policies supporting aging populations in remaining (Acemoglu and Restrepo 2022). healthy, skilled, and economically active have positive Supporting productive longevity can also help impacts across generations. Seniors who are in good individuals increase their healthy years of living. At physical and mental health and remain in the labor the individual level, there is substantial evidence that force help maintain the size of the workforce and pre- working in older age, in the right circumstances, can vent labor dependency ratios – the ratio of the pop- create a virtuous circle of healthy living and produc- ulation not working to that of the population work- tive work. Research focused on high-income coun- ing – from increasing. Harnessing the productivity tries (HICs) shows that opportunities for voluntary of these mature workers boosts the pool of human part-time paid work, as well as volunteering activities, capital, that is, the health and skills that ensure pro- can contribute to strengthening physical and mental ductive work. The gains can be substantial: estimates health at older age, including for people over 80 years suggest that Greece could increase GDP by a stagger- of age (Staudinger et al. 2016). The Gallup World Poll ing 23 percent by increasing employment levels of the shows that working full-time or voluntarily part-time 55+ age group to those of New Zealand (PwC 2018). in older age, as opposed to being retired, is correlated And productive longevity does not appear to crowd with higher levels of reported happiness.1 out younger workers from labor markets, what econ- Policies focusing on productive longevity are omists refer to as the “lump of labor fallacy.” Looking part of a broad policy agenda addressing aging and at Organization for Economic Cooperation and De- welfare challenges. First, actions across many differ- velopment (OECD) countries, higher employment ent policy areas will be needed to boost growth, in- rates for seniors in fact coincide with higher employ- clusive job opportunities, and social protection, while ment rates for youth (Böheim and Nice 2019, Gruber managing macroeconomic and fiscal conditions. et al. 2009, and Munnell and Wu 2013), and there is These include policies relating to education and life- some evidence from the United States and Japan that long learning, health and care systems, labor migra- higher senior employment is correlated with higher tion and labor market regulations, social protection wages for younger workers (Kalwij et al. 2010). At systems, taxation policy, innovation policy, and more. firm level, mixed-age teams have been shown to have Second, policies also need to target all generations: higher productivity than age-homogenous teams, promote longer working lives for those able to work supporting the notion that older and younger work- productively, invest in younger generations to raise ers are, on average, complements rather than substi- productivity over entire working lives, and enhance tutes (Göbel and Zwick 2013). Conversely, if seniors social protection for those who can no longer work. leaving employment deplete the market of relevant This is particularly the case in LMICs, where human skills, skill shortages may lead firms to automate pro- capital levels are lower and intergenerational human duction processes to a higher degree, which could capital gaps across generations larger, compared to instead lead to job destruction across generations HICs. Investing at scale in low-productivity mature 1 https://www.brookings.edu/blog/up-front/2014/03/28/why-aging-and-working-makes-us-happy-in-4-charts/ COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAP T ER 13 387 workers would likely be very costly compared to in- we look at would be relevant for people aged 55 years vesting in children and youth. Today’s adults are to- and above. Policy initiatives may differ depending on morrow’s elderly, and early investment in skills devel- the particular age group involved, however. opment is associated with a strong positive impact on There are specific growth and employment health, which in turn contributes to later retirement.2 opportunities associated with aging. Low aggre- While recognizing the need for broad policy gate labor demand in the formal sector, including packages, the chapter is focused on policies that en- for older workers, is often a critical problem to able and incentivize older workers in LMICs to re- productive employment growth in many LMICs. main active, and that can raise the productivity of Aggregate demand-side constraints limit the reach their work further. As will be discussed in more de- of supply-side policies aiming at incentivizing pro- tail below, in LMICs, many older people, often even ductive employment, for older workers and others. the majority, work for themselves or their families in This said, the nature of labor demand is changing the informal sector, in low-productivity employment globally, affecting demand for mature workers, and and without social security. Raising their productiv- will also impact labor markets in LMICs, albeit to ity and providing protection are critical challeng- a varying degree. Aging can itself generate demand es. Given largely underdeveloped social protection for products and services, thus generating jobs in systems in low-income countries (LICs) and many the so-called “silver economy.” middle-income countries (MICs), policies relating to The remainder of the chapter is organized as taxation, labor regulations, and formal social protec- follows: The next section describes the labor policy tion programs have limited reach. For the group of agenda for aging, highlighting the commonalities relatively better-off workers, often in the formal sec- and diversity of policy needs and options between tor, different labor market policies and social security HICs and LMICs as well as within LMICs. The fol- system incentives and barriers can matter, however. lowing section presents some stylized facts regard- A focus on this group is likely to yield greater results ing the nature of work for older adults. The next for this set of policies, as it is arguably less expensive section after this presents a conceptual framework and easier to extend working lives than renew or start around the main constraints to longer productive working lives at mature ages, particularly after a long working lives. Based on this framework, the pro- period of absence from the labor market. In what fol- ceeding section looks at the available evidence on lows, we will refer to “mature workers” approaching what works, drawing where possible on evidence retirement as the target age group. We do not consid- from LICs and MICs. The final section concludes er a precise age cut-off to define old age, seniority, el- with some meta-lessons and points to the important derly or mature workers. Typically, the interventions research agenda that lies ahead. A GLOBAL POLICY AGENDA FOR AGING Aging is a global phenomenon, and all countries will need to consider how to optimize policy to address the impact of older populations. Old-age dependency ratios are already high in many LMICs (particularly in East Asia and Eastern and Central Europe) and increasing rapidly in others (Latin America and the Middle East and North Africa). The share of persons above 65 years of age reaches 18 percent both in rich East Asian countries and in the OECD but is below 5 percent in Sub-Saharan Africa, the Middle East and North Africa, and South Asia. (Figure 13.1). In most of Latin America and Eastern Europe and Central Asia, fertility rates have already dropped below replacement levels, which means that the population will also be shrinking. Whereas most countries in South Asia and Sub-Saharan Africa still have relatively young populations, old-age populations are increasing rapidly in abso- lute numbers, and many Southeast Asian countries have already undergone a significant demographic transition in a very short period. Based on current trends, fertility rates will be below replacement levels in almost all Asian countries within 20 years.3 Many of these countries risk experiencing a slowdown in growth and thus “getting old before getting rich.” As the demographic window of opportunity is closing, sustaining welfare levels will require significant reforms, also in countries where social protection and skills-development systems are still developing. 2 Allel et al. (2021) show that, in Mexico, more access to formal education during childhood and adolescence is associated with a long-term positive effect on health and later retirement. 3   https://www.csis.org/analysis/will-many-developing-countries-get-old-they-get-rich 388 CHAP T ER 13 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E Globally, women’s life expectancy at birth exceeds affect old-age productivity and employment de- that of men by nearly five years, and the gap per- cisions. On the other hand, many workers in the sists into older ages. The gender longevity gap is informal sector—given the lack of access to social highest in Latin America, Europe, and Eastern and protection and overall low levels of private sav- South-Eastern Asia. This gender gap suggests that ings—arguably already work too late in life, caught in productive aging policies need to pay specific atten- a vicious circle of ill-health and low productivity and tion to skills and labor market constraints for wom- straining work, leading to even worse physical and en, and health constraints for men. mental health. The priorities and space for reforms will differ Increasing the productivity of low-productivity between less and more advanced economies, reflect- mature workers in LMICs would likely be difficult, ing different contexts of aging and productive work. given health and education gaps and lower life Most analysis and evidence on aging societies and expectancy. As access to education and health has the relationship with longer working lives is based expanded over the past decades, inter-generational on HICs. Much can be learned from these experi- gaps in education and health are also significant. In ences, but there are important differences between a country like Viet Name, which has undergone a less and more advanced economies that affect prior- remarkable upskilling effort in past decades, over 60 ities as well as the available and desirable set of pol- percent of those over age 60 have only completed icy instruments that can support longer and more primary education, whereas the vast majority of productive working lives. These include, inter alia, young people have at least higher secondary labor market structure, access to social protection, education (World Bank 2021a). In many LMICs, human capital levels across generations, norms, human capital levels in the younger generations, and the technical capacity and political space for while increasing, remain insufficient and need to reform. Nevertheless, countries that are now build- increase substantially, however, raising the urgency ing up their education, labor, and social protection of investing in children and youth who make up the systems can also learn not to repeat errors made by future workforce. more advanced economies. The experience from Despite these caveats, there are strong reasons HICs and some MICs shows that the cost of putting to form productive longevity policies for workers who in place inadequate solutions can be substantial. have higher human capital levels and resources to Unsustainable pension systems or insufficient at- retire. This group always exists: in many LICs, likely tention to skills development for older generations represented by public sector workers; in MICs, a have long-term repercussions that can be very dif- larger share is likely to come from the formal pri- ficult (politically and technically) to undo. Poorer vate sector. For this group, the policy agenda could countries need to avoid costly mistakes, such as more closely follow that of advanced economies and entrenching costly future pension payment con- focus on extending working lives. Although early tingencies in social protection design, or ignoring investment is critical, fostering skills development human capital investments, the effects of which will throughout working lives may be more efficient manifest themselves in decades to come. than has previously been considered, not least The dual labor market structure of most LMICs because a person over her lifetime generally spends presents a particular dilemma for policymakers. The much more time at work than in school. Research International Labor Organization (ILO) estimates shows that returns to work experience may explain that, on average, about 70 percent of the employed as much of the income gap between HICs and workforce is informally employed in developing LMICs as gaps in education: workers in LMICs countries.4 On the one hand, high informality and develop less human capital on the job than do low-productivity work mean that policies focusing workers in HICs (Jedwab et al. 2021). In addition, on formal workers and formal firms will be less ef- the hiring rate of mature workers may in fact be fective in changing aggregate outcomes, despite higher in LMICs than in HICs.5 This higher dyna- promising evidence in some areas that policy can mism suggests significant pay-offs to investing in human capital and adequate employment transition. 4   Based on statistics for 61 developing countries, from Elgin et al. 2021. 5   Annex Table 13A.1 shows that in several middle-income countries (Brazil, Colombia), the hiring rate of 55-64 year olds is much higher (34 and 24 percent respectively) than in EU27 or OECD countries (6 and 8 percent). COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAP T ER 13 389 FIGURE 13.1 Demographic changes differ across regions, but aging is coming everywhere 19 17 17 226 218 12 176 156 8 120 6 5 94 3 3 48 Source: World Development Indicators (left), UN 2019 (right). LMICs hence face a more comprehensive policy discuss the evidence for measures that aim to extend agenda than HICs, with both opportunities and risks. working lives though policy instruments that apply A significant part of the challenge of productive aging mostly to formal work as well as for measures that in LMICs is clearly about investing in children and can support more productive employment in both youth and increasing productivity and boosting so- the formal and informal sectors. In doing so, we rec- cial protection and health support among older work- ognize that the profile and context for mature workers ers, often in informal, low-productivity employment. today is different from those of the past and certainly But there are also opportunities for developing pol- likely to differ from those of tomorrow. Policy evi- icies addressed to mature workers. In this paper, we dence reported should be considered in this light. WORK IN OLD AGE: SOME STYLIZED FACTS This section presents some stylized facts regarding the nature of work among mature workers, looking at par- ticipation patterns and trends and the development of human capital - in the form of access to good health and the means to adapt to a labor market with rapidly changing skills needs – elements that determine the ability to work into old age and influence worker productivity.6 The nature of work among older adults Old-age labor force participation is generally lower, up through country income groups, participation the higher a country’s average income level. In the rates fall systematically. In lower middle-income average LIC, more than half of the 65+ population countries (LMICs), 33 percent of the 65+ popula- is active in the labor market (Figure 13.2). Moving tion is active, and in upper middle-income coun- 6   A complete assessment of labor market conditions for mature workers is outside of the scope of this paper. However, several World Bank country-level or regional studies provide more comprehensive overviews of labor market conditions for older workers including sector and occupations analysis, for example for Thailand (World Bank 2021b), Viet Nam (World Bank 2021a), East Asia and the Pacific (World Bank 2016), Malaysia (World Bank 2020), Russia (Levin 2015). 390 CHAP T ER 13 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E tries (UMICs) and HICs, this rate falls to 19 and 13 a need to work. In Thailand, for example, 31 percent percent, respectively. Across income groups, Euro- of people above age 60 report that labor income is pean countries are outliers, with significantly lower their main source of income, which is much higher participation rates than other regions. High labor than the share reporting pensions (5 percent) or el- force participation in poor countries largely reflects derly allowance (17 percent) – (World Bank 2021b). FIGURE 13.2 Labor force participation for older workers falls with a country’s income Source: Estimates based on Staudinger et al. 2016. Refers to 2013 data. The pattern of participation over the life cycle ipation is very low, the pattern for male participation is similar across country income levels, despite very is nonetheless similar to that of both richer and poor- different levels of participation. Figure 13.3 presents er countries, and even among women, there is an in- three cases as an illustration: low-income Liberia, mid- verse u-shape between participation and age. Globally, dle-income Morocco, and the high-income Republic patterns of age-related change in the balance between of Korea. Across these very different countries, labor paid work, unpaid work (including care for others), force participation falls after age 55, for both women personal care (for oneself), and leisure is surprisingly and men. In Morocco, where female labor force partic- consistent across countries (Ferranna et al. 2022). FIGURE 13.3 Labor force participation patterns are similar over the life cycle across richer and poorer countries Source: Estimates based on ILO data. Refers to 2019 data. COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAP T ER 13 391 Gender-gaps among mature workers appear as 6 percent by 2050 (World Bank 2022a). In a more important in MICs than in HICs. Based on a study of mature workers in Argentina, 70 percent comparison of a set of middle-income countries with of persons receiving a benefit and working reported the European Union (EU) and OECD, mature doing so because their pensions were insufficient to workers in the developing world (i) have higher cover needs (Barrientos 2011). In Malaysia, 40 per- gender employment gaps and (ii) exit the labor force cent of the population are not covered by pensions, later, but (iii) with earlier exit for women relative to and among those covered, many continue to work in men (Annex 13.1). The gender gap in labor force old age as their contribution history is insufficient to participation rates for workers aged 65+ is below 10 ensure adequate pensions (World Bank 2020). In a percentage points in Canada, Germany, and the set of Latin American countries, 18 percent of those United Kingdom, but above 30 for Chile, Malaysia, who receive a pension still work, and their working Mexico, and Türkiye (World Bank 2020). For LICs, hours approach full-time and are only marginally the gender gaps are not likely to be quite as dramatic. lower than those of persons who do not receive a The poorer the country, the higher the chances that pension (Eclac/ILO 2018). In a global study involv- women also have to work as they age. Lacking social ing consultations with older persons in LMICs, one- protection, many mature workers in LMICs third of participants reported working for pay. What continue working to secure their livelihoods. Data is more, twice as many were willing to work for pay on social insurance in LICs are scarce, but what is if given the opportunity (UNFPA 2012). available suggests that only a few percent of the Better-off workers are also considering continu- population are covered. In MICs, rates of coverage ing working later in life, whether to top up retirement are in double digits but rarely exceed 30 per-cent income or for other benefits, such as social connec- (Figure 13.4). In some countries, particularly in tions. According to a 2012 Eurobarometer Survey rapidly transforming economies such as Viet Nam, focused on Active Aging, one-third of Europeans informality is significantly more prevalent among would like to continue working after retirement age mature workers than among younger adults (Lam and two-thirds would welcome opportunities for par- and Elsayed 2021). Even for those covered, adequacy tial retirement (European Commission 2012). Labor of pension benefits is often low, raising incentives to supply elasticities seem to increase significantly over continue to work informally. Consider Georgia, the lifecycle (based on data from the United States) – where the universal pension amounts to 17 percent (French and Jones 2012), showing that incentives can of the average wage and is projected to fall to as little matter greatly for participation. FIGURE 13.4 In MICs, pension coverage remains low Source: Rutkowski 2018, reproduced in World Bank 2022b. Note: The figure shows participation rates in contributory pension schemes from the 1990s to the 2010s. The dashed horizontal line represents no change in rates of contribution over time. 392 CHAP T ER 13 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E Old-age labor force participation and educa- returns to work experience are significant, at around tion interact differently across countries. For HICs, 2 percent globally, and that education cannot easily levels of education are a strong determinant for la- substitute for experience. There is significant varia- bor force participation at older ages, likely reflecting tion across countries, however. First, as mentioned that education gives access to fulfilling work oppor- above, there are higher returns to experience for tunities and is not associated with physically strain- workers in HICs compared to LMICs. The authors ing labor (Boissonneault et al. 2020). In poorer hypothesize that this gap occurs as workers in HICs countries, however, more highly educated persons are more likely to be able to accumulate human cap- are more likely to have access to pensions and/or ital over time; however, the higher returns could accumulated private savings that make retirement also reflect more access to formal and regulated possible, whereas less-educated persons are more employment with automatic wage increases that are likely to be obliged to work in older age. unrelated to human capital or productivity, or to a Globally, self-employment is more common higher likelihood of working in firms that are richer among older workers than other age groups. Across in physical capital which can also boost productiv- countries, the 55-64 age group often has the highest ity among complementary factors of production. rate of self-employment in the working age popu- Second, in economies transitioning out of state-led lation (15-64). Compared to EU and OECD coun- economic systems (former communist economies), tries, mature workers in MICs are more likely to be skill sets have become obsolete more rapidly than in self-employed.7 In East and Southeast Asian coun- other countries, and so returns to experience have tries, self-employment rates are nearing 90 percent been lower for specific older cohorts.8 for older rural men and women; even in urban areas, In advanced economies, the labor market par- the share of the self-employed increases as of age 60, ticipation of older workers has increased in the reflecting that the self-employed are more likely to past few decades. In 1996, the average worker in an stay in the labor market, whereas employees, at least OECD country could expect to exit the labor force from the formal sector, retire (World Bank 2016). at approximately 62 years of age; in 2016, the age In HICs, self-employment can represent “opportu- of exit had increased by two years, to 64 (Boisson- nity” entrepreneurship, a desire to use experience neault et al. 2020). Between 2002 and 2017, the par- to run a business or lead a more flexible working ticipation rate for workers between 55 and 74 years life. In the United States, the rate of new entrepre- increased by 9 percentage points across OECD neurs is significantly higher in the age group 55-64 countries, reversing a decline that had taken place than in the age group 20-34 (0.37 vs 0.23 percent). since the 1970s (Geppert et al. 2019). These trends The share of older adults among new entrepreneurs have occurred across OECD countries, across gen- also increased from 15 percent in 1996 to 26 per- ders, and across socioeconomic groups. cent in 2018, and older adults are more likely than In contrast to trends over the past 20 years, the other age groups to be opportunity driven (Farlie et COVID-19 pandemic has resulted in a significant al. 2019). However, higher rates of self-employment withdrawal of older workers from the labor market, among older workers in LMICs often signal “push however. Data from Australia, Austria, China, and factors,” such as lack of wage-work opportunities, the United Kingdom show that older workers were with higher informality and overall, more precar- worse hit in the labor market than younger work- ious employment. In Thailand, for example, infor- ers by the COVID-19 pandemic, both in terms of mality rates are below 50 percent for workers aged quality of work, and in loss of work (Pit et al. 2021). 35-44, but they are nearly 80 percent for those aged In the United Kingdom and the United States, the 55-64 and over 90 percent for workers aged 65 or largest cuts in total employment were due to old- more (World Bank 2021b). er workers (55 and over) who dropped out of the Earnings generally increase with age, all else labor market during the height of the pandemic. equal, but patterns vary and may depend on age co- Many of these workers have not (yet) returned to hort characteristics. A recent global study finds that work and may never do so - once out of the labor 7   See Table 13A.1 8   Jedwab et al. 2021. Chernina and Gimpelson (2022) show, for Russia, that when time, cohort and experience effects are separated, the cohort and experience factors work in opposite directions, reflecting the massive depreciation of skills. COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAP T ER 13 393 force, older workers are more likely to remain in- Human capital and productive longevity active than their younger counterparts (Pizzinelli and Shibata 2022). A study of interstate informal Health and skills – human capital – are central to sector migrant workers in India showed a strong productive longevity. The increase in labor force and positive correlation between age and income participation observed in OECD countries in the losses during the COVID-19 pandemic (Guha et past decades can largely be attributed to higher life al. 2020). These outcomes stand in contrast to the expectancy and educational attainment, whereas Global Financial Crisis, which did not result in a changes to statutory retirement age accounted for sustained reduction in older workers’ participation less than 10 percent of changes (Geppert et al. 2019). in labor markets (OECD 2013). Before the COVID-19 pandemic, healthy life The changing nature of work may be contrib- expectancy had been increasing globally, prolong- uting to more age-friendly jobs. Technology is en- ing the physical and mental ability to work. A strong abling automation of many tasks, but also raises relationship can be found between health and work demand for skills that are (still) complementary to in old age.9 In the past two decades, health-adjusted technology, such as creativity, problem solving, and or healthy life expectancy (HALE)10 has increased socio-emotional competencies (World Bank 2019). by 5.4 years, from 58.3 in 2000 to 63.7 in 2019 Globally, fast-aging countries are seeing higher (Figure 13.5). Healthy additional life years have in- rates of automation, driven by skills gaps related to creased across all country income groups, although specific cohorts (manual production skills) – (Ace� - there are still gaps between advanced and develop- moglu and Restrepo 2022). At the same time, in the ing countries. By 2015, a woman aged 60 living in United States, in parallel to the automation process, an UMIC, could expect to live nearly 17 addition- jobs have become more “age-friendly”: less physi- al years in good health, all else equal, an increase cally demanding, more flexible, or involving auton- by 21 months compared to 2000. LMICs and LICs omous and team work, for example. These jobs are have also seen increases in HALE, albeit somewhat not necessarily taken by mature workers, however, smaller, and women aged 60 can now expect to live although they are more adapted to their needs and in good health for another 14 years. There are sig- preferences (Acemoglu et al. 2022). nificant gender gaps, however: men can expect two FIGURE 13.5 Healthy life expectancy has been increasing across the globe Source: Estimates based on WHO statistics. 9   For example, Giles et al. 2012. 10   HALE is the average number of years that a person can expect to live in “full health” by taking into account years lived in less than full health due to disease and/or injury. https://www.who.int/data/gho/indicator-metadata-registry/imr-details/66 394 CHAP T ER 13 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E fewer years of healthy life than women in HICs and The aging process need not prevent older peo- UMICs. These statistics point to a gendered work, ple from working productively and learning new life, and health challenge: whereas women are more things; depending on the task, aging may even be likely to live longer than men, especially in HICs helpful. Aging inevitably involves some cognitive and MICs, they are less likely to participate in work, decline. However, both the timing and extent of de- especially in MICs. cline vary significantly across individuals, and the Working longer can in fact create a virtuous overall rate of decline can, as mentioned, be mod- circle of health and work at older age, but the out- erated. Moreover, not all cognitive abilities neces- come likely depends on the nature of work and em- sarily decline during aging, at least not before 80 ployment conditions. Age-related physical and psy- years of age, and some even improve at older ages chological changes can be moderated with physical (Spotlight 13.1). The ability to develop new skills is and intellectual activity and other lifestyle factors critical to retaining productivity, however. Where- (Crawford et al. 2009). Work can bring cognitive as improved health is increasing the potential years stimulation and social interaction, which contribute of work, the shelf-life of skills is falling.11 The evi- to continued mental and physical agility. Social in- dence on capacity to learn at older age suggests that teractions at work, which are likely to demand con- older adults’ learning process is different from that stant adaptation, have been shown to be particularly among younger people, but not necessarily less ef- important to maintain brain capacity (Staudinger et ficient (Picchio 2021; Thomas et al. 2020). In fact, al. 2016). However, prolonging work in (predomi� - workers can continue to contribute significantly to nantly low-skill) jobs that provide limited stimula- workplace productivity at older ages, drawing on tion, jobs that cause undue stress, and jobs that are experience and capacity for judgement. Whereas physically straining with high risk of occupational they may be somewhat slower to complete some disorders, is bad for healthy longevity (Staudinger tasks, they compensate for lack of speed by execut- et al. 2016, Prinz et al. 2018). The overall employ� - ing them better (Backes-Gellner et al. 2011). ment situation also matters, as job insecurity in old Mature workers may also have stronger so- age can in fact lead to deteriorating mental health cio-emotional skills that are in increasing demand. (Gutierrez and Michaud 2019). Contrary to common perceptions, mature work- Beyond health, skills investments are also criti- ers score higher than younger colleagues on broad cal as demographic patterns shift, since the nature of concepts of job performance, including on reduc- jobs is also changing, raising the importance of con- ing counterproductive work behaviors, workplace tinuous skills development. Across the globe, there aggression, and tardiness, all of which are import- is substantial evidence that jobs are becoming more ant factors affecting overall productivity. Moreover, intense in cognitive, analytical tasks. Technological mature workers are not necessarily less innovative change is resulting in higher demand for technolo- in the workplace than younger workers (Ng and gy-complementary skills, both higher-order cogni- Feldman 2008). A meta-study found that mature tive skills as well as socio-emotional skills (World workers generally function equally well or even Bank 2019). Accessing productive work, hence, in- slightly better than young workers on emotional creasingly requires a more complex set of skills and competencies, which are specifically important for adaptive capacity. Evidence from the EU shows that the growing services sector (and arguably for the the productivity decline often observed among old- broader economy, as well) – (Doerwald et al. 2016). er workers is more likely due to skills obsolescence A study of German workers in the services sector than to age itself; in Mexico, wages of college-edu- showed that older workers were more engaged with cated workers have fallen due to a fall in wages of their work, more confident in their abilities to do mature rather than young workers, pointing to a the job, and less prone to burnout than younger similar potential effect (Mayhew and Rijkers 2004). employees (Johnson et al. 2017). 11   https://www.weforum.org/agenda/2017/07/skill-reskill-prepare-for-future-of-work/ COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAP T ER 13 395 SPOTLIGHT 13.1 Ok, Boomer! The aging brain and implications for work Brain plasticity - the ability of the nervous system to change its activity in response to intrinsic or extrinsic stimuli by reorganizing its structure, functions, or connections – declines with age. However, the evidence from neuroscience research on the impact of the aging brain and cognitive decline is mixed, at least for ages up until the 70s. The brain changes, but different functions combine and collaborate and can, in some circumstances, compensate for the reduction in brain plasticity. For example, the ability to receive information declines relatively early, but the capacity to direct information and act on the information (orienting and executive functions) can in fact increase, well into the 70s. Changes in cognitive skills are hence not uniform over a lifetime. Whereas some skills such as verbal memory and perceptual speed peak relatively early in adulthood, other skills such as verbal ability or auditory comprehension, which draw on orienting and executive functions, peak much later, often after 50 years of age. As a result, the aging brain may in fact retain a high degree of functionality. There is conclusive evidence that promoting mental and physical health as well as continuous brain exercise helps in slowing cog- nitive decline. Individuals who follow healthy diets and practice physical training, social activities, and education and training see slower decline in cognitive capacity and higher subjective well-being. This creates a positive feedback loop, as a workplace may be an effective mechanism to retain cognitive status with Intellectual engagement and cognitive stimulation. Source: Veríssimo et al. 2022; Desjardins and Warnke 2012. WHAT HOLDS BACK PRODUCTIVE LONGEVITY? What can be done to promote longer and more productive working lives? Several factors can constrain mature workers’ participation in the labor market and their productivity at work. This section presents a framework high- lighting constraints that can affect mature workers’ incentives and ability to work and do so productively, employ- ers’ incentives to hire or retain them and invest in their human capital, and matching between demand and supply (Figure 13.6). Throughout the remainder of the chapter, we try to identify LMIC dimensions where possible. FIGURE 13.6 A policy framework for longer and more productive working lives Source: Original figure for this publication. 396 CHAP T ER 13 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E Institutional, market, and behavioral failures create for groups with higher inactivity rates (and thus a constraints for mature workers and the firms that may larger labor supply reserve), including women and employ them. Institutions and policy and regulatory low-skilled workers, and for populations with high- frameworks affect incentives and opportunities for ly elastic labor supply, which includes many older workers to engage, and for firms to hire them. Exam- workers (Meghir and Phillips 2010). ples include labor regulations and social protection, Social welfare systems, including the level and including pensions. Market failures include missed- structure of pension benefits as well as unemploy- out externalities from more skills development or bet- ment or disability pensions, can further affect in- ter public services (including care services), imperfect centives to work. In many countries, formal sector information about workers’ ability, or the public-good workers enjoy relatively generous early retirement nature of some investments in health or education. rules, creating incentives to retire early (Bussolo et Finally, behavioral constraints include social norms al. 2015, World Bank 2016). In East Asia and Pacif- around who works, or internalized ageism – prejudice ic countries, retirement ages in national pension or discrimination on the grounds of a person’s age. schemes are relatively low compared with life ex- pectancy at age of retirement (World Bank 2016). Supply-side constraints: The design of other social welfare programs also Incentives and ability to work productively matters. If mature workers are eligible for pro- longed unemployment benefits, or if the require- To extend working lives with productive employ- ments for job search in unemployment or disabil- ment, workers need to (i) have the incentives to ity-benefit programs are waived for them, are too work, with institutions and regulatory frameworks lax, or are weakly enforced, alternative pathways that allow for it and do not discourage it, including to leaving the labor force without jeopardizing re- in ways that accommodate changing preferences; tirement benefits open (Ahlqvist and Boren 2017; and (ii) the skills and health necessary to be employ- Kyyrä and Ollikainen 2008). Such loopholes can able, particularly in the context of a rapidly changing hence reduce the impact of pension reforms aimed world of work. We discuss these two areas in turn. at increasing retirement ages.13 Labor regulations, tax systems, and social wel- Pension income, whether through contributo- fare systems can significantly influence or outright ry or noncontributory systems, can be an import- dictate opportunities for work beyond retirement, ant factor in determining workers’ participation with both positive and negative effects. Mandatory choices in middle-income countries. Analysis of the retirement ages, legal impediments to partial retire- impact of the early retirement age in Central Eu- ment, or unfavorable regulations for part-time work rope and the Western Balkans shows that pension or flexible work arrangements or pay may outright receipts are significantly and negatively associated prohibit or make it less attractive for workers to with labor force participation (Bussolo et al. 2015, continue. China’s retirement rules still oblige men Gragnolati et al. 2011). In Argentina, widening the to retire at 60 and women at 55, or 50 for specific, eligibility for noncontributory pensions lowered la- physically more burdensome jobs (OECD 2019a). bor force participation rates by 5 percentage points Almost half of the world’s countries, the vast ma- and increased informal work (Bosch and Guajardo jority LMICs, have limits on the use of fixed-term 2012). The South African social pension program contracts.12 On the other hand, employment pro- has been shown to reduce employment among the tection legislation favoring tenure may allow work- elderly (Baird et al. 201814). These experiences ap- ers to stay on the job and retain their wages. Taxa- pear consistent with historical evidence from more tion of earned income in addition to pensions and advanced economies. The introduction of a social second-earner taxation tend to rapidly increase pension from age 70 in the United Kingdom, some the implicit tax rate on work with age (Butrica et 100 years ago, lowered labor force participation al. 2006). Taxation appears particularly important rates by 6 percent, even with a very low adequacy 12   Data from World Bank Employing Workers Database. The median limit is three years (36 months) for cumulative contracts. 13   Pension reform during recession in Germany resulted in alternative pathways pushed by employers - unemployment and partial retire- ment (Lorenz et al. 2020). 14   The paper shows that cash transfers in general do not lower aggregate labor force participation, however. COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAP T ER 13 397 of benefits (Gisecke and Jäger 2021). As systems the sample, whereas unpaid work was highest when expand, it is important to make sure that programs people were in their late 30s (likely related to child- remain incentive-compatible. bearing and rearing), and again increased after age Information gaps due to limited financial lit- 60 (Ferranna et al. 2022). In countries with signifi- eracy can also impact retirement choices for those cant outmigration, whether international or intra- covered by pension systems. Understanding the ex- national, grandparents may be the main caregivers pected level of benefits and how that will impact wel- in so-called “skipped-generation” households con- fare in older age is a prerequisite for making rational sisting of only older people and children (UNFPA labor market decisions. Surveys in Sweden and Ire- 2012). Studies from Brazil, Mexico, and Thailand land show that knowledge of future pension benefits show that grandmothers’ care obligations affect is at best limited, and that the most economically both their participation and hours worked nega- vulnerable groups know the least about their future tively (Attanasio et al. 2022, Paweenawat and Liao benefits (Elinder et al. 2020, Barrett et al. 2015). 2021).15 Since families are often formed relatively In addition to rethinking pension systems, flexi- early in life, many of these grandmothers in LMICs ble work arrangements may be needed to accommo- often have significant potential working years ahead date work preferences that change with age. There is when they leave the workforce, reduce hours, or evidence that mature workers need and value flexible move to less-demanding activities. Spouse employ- work arrangements, including both those related to ment is also generally positively correlated: to retire flexible time schedules and changes in work roles is a joint decision, made with a view to spending and assigned tasks, and that providing such arrange- time together (Giles et al. 2012). A study from Rus- ments serves to delay retirement (Atkinson and San- sia showed that having a working wife is associated diford 2016). Particularly in HICs but also in some with a 13 percentage point increase in the probabili- middle-income countries, COVID-19 has given rise ty that a pension-age husband will still be employed to structural changes in the organization of work, (Levin 2015). In urban China, a working spouse is with a rapid move towards more home-based and associated with as much as 40 percent and 58 per- flexible solutions for many white-collar jobs, as well cent increases in the probability of men and women as a shift to independent work, that over the long run working, respectively (World Bank 2016). may benefit working conditions for mature workers Lack of safe and reliable transportation systems (McKinsey Global Institute 2021). affects mature workers’ access to productive oppor- Family and the availability of care services also tunities. More care responsibilities, lower stamina matter for labor market participation, especially for and physical strength, and related vulnerability to women. Grandmothers all over the world hold a crime means that mature workers depend heavily central responsibility for the care of their grandchil- on quality transportation services. A survey of se- dren, and hence support the labor market partici- nior workers in the United States suggested that lack pation of the parents of these children. Towards the of reliable transportation was a somewhat serious or end of working life, many are also caring for elderly very serious obstacle to working for 82 percent of parents or even their spouse. In Europe, one-third respondents (NCST 2011). Research on the impact of grandparents provide regular care and women of safe, affordable, reliable, and effective transport who do are also less likely to be working (Backhaus systems for elderly is lacking. However, limited ac- and Barslund 2021). Care obligations are even more cess to safe transportation is estimated to be a major prevalent in LMICs, where three in four elderly constraint for women’s participation in labor mar- people live in the same household with members of kets, reducing their participation probability by 17 younger cohorts, compared to one in four in HICs, percent (ILO 2017). Underdeveloped transport sys- and where formal care services are often lacking tems in LMICs are likely to similarly reduce access in supply and quality (UN 2007). A study of time to work for the older population. use allocation in a set of OECD countries, China, Psychosocial factors may generate additional and India, showed that the time share allocated to disincentives for work. If individuals internalize so- paid work fell successively as of age 45-50 across cial norms and stereotypes around age versus capa- 15   A study from Mexico also showed that female labor participation of a mother was not negatively correlated with more children, but that co-habitation with grandmothers increased in response to more children (Schmieder 2021). 398 CHAP T ER 13 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E bilities, responsibilities, and work, their incentives of training monotonically falls with age across all and even capacity to work may be reduced. Individ- countries. Low incidence of training is partly due uals who self-identify as being older risk considering to underdeveloped labor market and career guid- that they are too old for certain jobs, for changing ance information systems and services and limited jobs, for promotion, or for developing their skills and fragmented markets for training (including for further, and that they should not crowd out the firm-provided training, discussed below). However, younger generation (Vickerstaff and Van der Horst even where subsidized adult training programs are 2021). There is evidence, for example, that “health provided, take-up is often low (World Bank 2019), pessimistic” individuals prefer exiting the labor force and there is some evidence that the mature workers earlier than others, even when actual health levels who would be most in need of upskilling – those are not different from other workers. Unsurprising- with lower levels of education and lower wages – ly, those with higher education and income appear may also be the least interested in or able to par- to have more positive self-perceptions of aging than ticipate in training (Generation 2021). Workers people with lower incomes and education (WHO may lack information on the value of training, what 2021). Gender and social class also influence own skills to acquire, or what training opportunities are perceptions of optimal exit age (van der Horst 2019). available. Retraining at older ages may also seem Beyond issues of regulations, incentives, and more challenging, discouraging take-up. Workers other barriers, mature workers often underinvest in may choose to not invest in training if they will skills development in ways that hamper their labor not receive fair remuneration for their higher pro- market prospects and lower the opportunity cost ductivity (Almeida et al. 2012). Lack of training is of retiring. Although skills upgrading is essential arguably particularly problematic in LMICs, where to remaining relevant in labor markets, participa- human capital among workers is lower overall and tion in adult learning is low overall, especially for intergenerational human capital gaps are high. mature workers and in LMICs. Figure 13.7 shows, Since skills development is often cumulative and for EU member countries as well as five European skills demand can change quickly, the skills gap is MICs, that (i) training is lower in the European likely to compound over time. MICs than in the EU HICs, but (ii) the incidence FIGURE 13.7  Participation in formal or non-formal training, by age Source: Estimates based on Eurostat data. Refers to 2016 data. COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAP T ER 13 399 Poor health is also reducing both participation tivity among Brazilian firms showed that wages and productivity among mature workers. Good were in fact unrelated to productivity; a study in health is a critical asset for productive work and a the Netherlands found a similar disconnect be- central component of human capital. Healthy old- tween wage levels and performance (Rocha 2017, er people are more likely to participate in the labor Dohmen 2003). In many East Asian countries, the market, although again, in poorer countries, those seniority wage-setting practice is strong, and per- in poor health may still be obliged to work to the formance-based pay practices are not implemented extent they can. Health is not the focus of this chap- (World Bank 2016). Wages that rise above produc- ter and is not reviewed in detail here. It is worth tivity levels risk lowering the competitiveness of noting, however, that health interventions may be mature workers. Analysis of firm-level data in East included among low-hanging fruit for addressing Asian countries shows that seniority wage practices productivity constraints in low- and middle-in- indeed are strongly and negatively correlated with come settings. One area is nutritional deficiencies hiring workers in the age group 50-64.16 occurring as seniors need fewer calories but more Formal regulations also matter for labor costs nutrient-rich meals with age; other areas are broad- and may shift these above mature workers’ produc- er lifestyle choices that influence later development tivity levels. First, social contributions increase with of non-communicable diseases (NCDs). The com- wages. This can contribute significantly to raising pression of morbidity and the extension of good labor costs, as in some countries social insurance health into later years permits longer returns to contributions are very high – e.g., 45 and 36 per- continuous investment in education. cent of wages in Argentina and Brazil, respectively.17 Especially in the presence of asymmetric informa- Demand-side constraints: tion, the cost of hiring older workers may therefore Incentives and ability to hire and invest in be higher (OECD 2019b). If productivity does not mature workers increase in parallel with social contributions, unit labor costs for mature workers increase. A survey This section looks at employers’ incentives to retain of UK employers suggested that pension costs were or hire mature workers and adapt working situations the main deterrent to hiring mature workers – to increase their productivity. Factors that constrain more important than the limited time frame for re- relevant opportunities include informal and formal turn on investments in skills, or any perceived lack institutions governing employment and wages, mar- of skills (Auer and Fortenu 2000). Similarly, em- ket failures that restrict the quantity and quality of ployment protection legislation (EPL) can increase training, and psychosocial and normative barriers to the costs of labor adjustments, as many regulations employers’ recognizing the potential of older workers. imply higher costs with longer job tenure. Global- Regulations and informal institutions around ly, the average number of weeks’ pay required as wage setting may raise labor costs for mature work- severance in the case of dismissal increases from ers above productivity and hence lower their attrac- three for workers with one year of tenure, to 21 for tiveness to employers. There is significant evidence workers with 10 years of tenure. In Indonesia, a that job tenure or seniority leads to higher wages worker with 10 years of service receives an average in ways that may be disconnected from produc- severance payment equal to 22 months of wages.18 tivity increases. In OECD countries, 10 additional Strict EPL may be one of the reasons for not hiring years spent in a job with the same employer in- mature workers (Bussolo et al. 2015). crease wages by nearly 6 percent on average, all else “Ageism” and social norms, including those equal (OECD 2019b). Wages appear to rise even reinforcing age stereotypes, also make the hiring of more sharply with tenure in at least some LMICs, older workers more difficult. The assumptions that compared to HICs (Donovan et al. 2020). A study mature workers are frail, stubborn, less capable of of seniority wage setting and firm-level produc- learning, and technologically challenged can result 16   At the same time, seniority wages may have other underlying rationales, as firms may purposedly set wages below productivity levels for younger workers and above productivity levels for older workers, to foster firm fidelity (Zwick 2009). 17   https://iuslaboris.com/insights/comparing-social-security-contributions-rates-across-latin-america/ 18   Estimates based on Employing Workers Database, https://www.worldbank.org/en/research/employing-workers 400 CHAP T ER 13 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E in age-stereotyped hiring practices and workforce between being older and the chances of being called management practices that reduce the productivity for an interview, and callback rates began to decline of mature workers—for example, if mature workers for workers already in their early 40s (Carlsson and feel insecure about their jobs and thus invest less in Eriksson 2019). A similar approach for the United skills (Jin and Baumgartner 2019). There is signifi- States provided robust evidence of age discrimina- cant evidence of biased perceptions and manage- tion against older women, but not against older men ment practices with respect to older workers with- (Neumark et al. 2015). In Deloitte’s 2018 Global Hu- in firms. A survey by Generation 21 across Brazil, man Capital Trends report, 20 percent of business India, Italy, Singapore, Spain, the United Kingdom, leaders viewed older workers as a competitive disad- and the United States showed that 38 percent of vantage, and in countries such as the Netherlands, managers considered mature workers unwilling Russia, and Singapore, this percentage was far high- to take up new technology, 27 percent considered er; some 15 percent also felt the older employees to them unable to learn new skills, and 21 percent an- be blocking rising (young) talent (Deloitte 2018). swered that mature workers did not work well with Firms—particularly small and microenterpris- other generations. Yet, when hiring workers, the es— generally invest little in skills development for same managers valued experience above all (Gen- their workforces, despite the importance of skills eration 2021). In countries lacking age discrimi- for firm productivity. Individuals’ low investment nation legislation, practices such as defining upper in their skills development is thus mirrored in un- age limits for job vacancies – and to some extent derinvestment by employers. The World Bank’s en- mandatory retirement ages – outright discriminate terprise surveys show that only one-third of firms against mature workers (OECD 2019b). A survey in offer training to their workforce, and those that do, EU countries found that a majority of respondents invest in a minority of their workers (Figure 13.8a thought that being over 55 years of age put job ap- and b). Across country income groups, small and plicants at a disadvantage (European Commission micro-firms are the least likely to offer training, and 2015). An experimental study in Sweden showed firms in LMICs are less likely to offer training com- significant and independent negative correlation pared to firms in HICs, for each size class of firms. FIGURE 13.8 Access to training through employers is limited, especially in LMICs a. Average share of firms providing training and b. Average share of firms providing training, by firm share of workers receiving training (in firms where size and income group offered), by income group Source: Estimates based on World Bank Enterprise Surveys. COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAP T ER 13 401 Several market failures may help explain firms’ workers in the labor market may be costly and underinvestment in skills beyond formal education.19 lengthy, particularly for older workers. Firms may underinvest because training to some Public and private employment services may extent is a public good, and they risk losing the re- have limited capacity to serve older workers. Pub- turns to investment if other firms “poach” the skilled lic employment services exist in most developing workers, and more generally due to high worker countries (nearly 80 percent in 2020), and a ma- turnover. Managers may lack accurate information jority provide both job placement, counseling, and about what skills will be needed, the value of train- training services.20 However, official retirement ing for productivity, earnings, and employability, or age is often set as the upper age limit for program what training opportunities are available (Almei- eligibility, excluding older workers. Moreover, ser- da et al. 2012). Financial institutions that likewise vices such as job search support or training may lack understanding about training and the impact be poorly adapted to older workers. Caseloads for on productivity can restrict credit for training pur- employment service case workers often exceed the poses. Coordination failures arise when firms can- levels recommended by the ILO (1:100), suggest- not find productive workers and therefore cannot ing limited capacity to provide tailored services increase pay, whereas workers underinvest in skills to a group with specific training and other needs. because they will not be rewarded, and the training The increase in e-government services in high- and offer remains underdeveloped. In principle, the time middle-income countries needs careful design and horizon for reaping the benefits of skills investments outreach efforts to reach mature workers. A study by firms is shorter for mature workers, which may of employment services’ self-service tools in the also create disincentives to invest in their skills. United States found that these features were least Firms may also underinvest in other comple- used by mature workers precisely because of lower mentary strategies that are valuable to increase the digital skills (D’Amico et al. 2009). The risk of ex� - productivity of older workers. Relevant technology clusion may be even higher in LMIC settings, given tools, more ergonomic working conditions, health lower levels of education and skills. Digital gaps are literacy, and other health interventions, as well as likely to have worsened the impact of COVID-19 organizational changes such as multi-age teams or on mature workers’ employability as employment flexible working arrangements may be needed. As support often moved online. in the case of skills, firms may lack the knowledge Lack of job and career search skills among ma- or resources for such investments (Söderbacka et ture workers may also be holding back matching al. 2020, EU-OSHA 2016). opportunities. Digital skills are increasingly im- portant in identifying and applying for job opportu- Intermediation: Effective matching nities (as well as increasingly important for the job services for mature workers itself). However, globally, digital literacy is lower in the older generation (ITU 2021) and the gaps are Even when mature workers have the necessary skills likely to be higher in LMICs, where age-related gaps and health, and firms are prepared to hire them, in literacy and education are also higher. the search and matching process between jobs and 19   See the framework presented in Almeida et al. 2012. 20   Estimates based on World Bank Unemployment Protection and Employment Services Database. 402 CHAP T ER 13 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E PRODUCTIVE LONGEVITY: WHAT COULD WORK IN LMICS? This section presents evidence for interventions directed at individuals and families (supply side), at firms (de- mand side), and at employment services in charge of labor market intermediation. Across these dimensions, we identify policies that could raise incentives and remove barriers to accessing work, and those that could foster productivity by raising human capital and especially skills (there is limited evidence on health). Policies aimed at increasing labor force participation are largely focused on formal sector workers, for whom regulatory reform is relevant. Formal sector workers are more likely to have resources (and thus incentives) to retire, but at the same time are more able to continue working and with higher productivity, than workers in informal, low-productiv- ity jobs. Human capital policies, by contrast, are also relevant for the informal sector. In identifying relevant literature, we have given pri- rica have also reformed their pension systems. The ority to studies that present results for LICs or MICs objective has primarily been to increase financial and to rigorous evaluations, to the extent possible. sustainability, and labor market impacts have not However, the policy agenda focusing on productive been systematically studied, however. longevity is relatively recent and the evidence on Pension reforms require careful attention to policy effectiveness limited, especially for LMICs. the entire social insurance system. As discussed Where evidence for mature workers in LMICS is above, increasing retirement age may also result in scarce, we focus on the impact of similar interven- a more intensive use of alternative pathways into in- tions for other groups, or for HICs, and discuss po- activity, such as unemployment, disability support, tential differences that would need to be considered or partial retirement – especially for low-wage and when drawing lessons from these experiences. less healthy workers, and especially during reces- sions. In the studies of Germany and Austria cited Supply-side interventions: above, for example, some mature workers transi- Improve incentives and remove barriers to tioned into other forms of social insurance sup- work while investing in skills development port. Countries have consequently embarked on a parallel set of reforms to close loopholes, including This section discusses evidence on regulatory re- abolishing the right of older unemployed persons forms to improve work incentives and make work to waive job search requirements when accessing pay for mature workers, along with improvements unemployment benefits (Ardito 2021; Lorenz et al. in services and skills development aimed at increas- 2020; Staubli and Zweimuller 2013). ing the ability and capacity of mature workers to Relaxing partial retirement rules is a common- work longer and more productively. ly proposed reform to retain workers, but when de- signing and implementing such policies, it is also Reforming social insurance and labor taxation important to consider potential negative impacts. Studies show that flexible working arrangements Reforms to social insurance focused on increasing re- giving workers choice and control are linked to bet- tirement age and reducing work disincentives in the ter health and wellbeing (Graham 2014). When sur- design of pension schemes can play a role in extend- veyed, workers report favoring arrangements that ing working lives for formal sector workers. Across would allow them to adjust the timing and speed OECD countries, comprehensive social security of labor market exit according to their preferences reforms of contributions, benefits, and retirement (OECD 2017, Henkens et al. 2021, Munnell and ages are planned or have already been implemented Walters 2019). However, partial retirement options (Boissonneault et al. 2020). Raising retirement ages in OECD countries do not appear to increase total has resulted in postponed retirement and higher working hours, largely because they also afford the employment rates among mature workers. Raising opportunity to retire earlier than planned (OECD retirement ages in Austria and Germany, for exam- 2017). There is also some evidence that partial re- ple, resulted in increases in employment by nearly 10 tirement regulations can have a negative effect on percentage points for the affected groups (Riphahn hiring older workers or be used as an alternative and Schrader 2021; Staubli and Zweimuller 2013). pathway for retrenchment (Busch et al. 2021). Par- Many middle-income countries in East Asia and the tial retirement may be incompatible with job and Pacific, Europe and Central Asia, Latin America and task bundling, especially for high-skill jobs, and the Caribbean, and the Middle East and North Af- may raise the cost of mature workers, which in turn COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAP T ER 13 403 reduces incentives for hiring (Lorenz et al. 2020, labor supply, typically women, less-educated work- Munnell and Walters 2019). ers, and mature workers (Meghir and Phillips 2010). Pension reforms carry high political costs, OECD countries (Australia, Denmark, the Nether- pointing to the importance of getting the design lands, Sweden, and the United Kingdom, among oth- right from the outset. In Europe, several countries ers) have introduced tax credits, in some cases with have reversed earlier reforms by postponing, lim- the credit increasing with the age of the worker. Tax iting, or outright reversing existing or planned in- credit reform linked to pension reform increased par- creases in early or normal retirement ages (Fouejieu ticipation among mature workers in Sweden, except et al. 2021). More generally, the experience of both for self-employed and higher-earning workers, among advanced and developing economies shows that whom the “slack” may be smaller (Laun 2017). design flaws in social protection systems can take many years to untangle and may carry very high fis- Improving access to quality services cal and political costs. Countries seeking to expand social protection systems therefore should incorpo- Subsidizing child- or other forms of care for family rate lessons learned to ensure sustainability from members can improve labor market participation the outset and avoid painful reforms later. and earnings outcomes. Affordable childcare or Incentives to work increase as people receive in- long-term care for sick and elderly persons can re- formation on future pension income. Evidence from lieve mature workers from informal care burdens the United States shows that the framing of public and help them manage personal and professional messages regarding pensions matters for when ben- commitments. In Rio de Janeiro, Brazil, a lottery efits are claimed. Low-cost interventions (informa- was used to allocate public center-based daycare to tional brochures, invitation to a short web-tutorial) households. Access to daycare centers was associat- informing older workers about expected pension ben- ed with sustainable increases in household income, efits increased labor force participation one year later due to grandparents’ increased labor income. Grand- by 4 percentage points relative to the control group parents’ earnings, likelihood of employment, hours (Liebman and Luttmer 2015). Pension reforms have worked, and social security contributions increased routinely been accompanied by financial education measurably (Attanasio et al. 2017). This positive as- information campaigns in the OECD (OECD 2008). sociation between care services and participation is No equivalent research is yet available for LMICs, but consistent with research on the impact on women’s there is evidence that information campaigns geared labor market outcomes of different forms of child- towards changing jobs and career decisions, for exam- care provision in Latin America, Russia, and the EU, ple financial education or career guidance, can impact where studies have found strong effects on activation, individual and family choices (OECD 2008).21 employment, and hours worked (Mateo Diaz and Labor regulations can also allow for greater Rodriguez-Chamussy 2016, Del Boca and Locatelli flexibility in work arrangements to accommodate 2006, World Bank 2017). A quasi-experimental study varying and changing needs among mature work- in Spain focused on the implementation of succes- ers. For example, allowing for more part-time work, sive care-subsidy reforms, involving cash subsides as temporary work, or fixed-term contracts, can in fact well as care services. It concluded that changes in care serve mature workers looking to adapt their working subsidies had a stronger effect on early retirement schedules. Between 2007 and 2019, over 100 coun- decisions than reforms tightening options for early tries undertook significant reform of employment retirement (Costa-Font and Vilaplana-Prieto 2022).22 regulations, including on fixed-term contracts, sev- Safer transportation helps vulnerable groups erance payments, annual leave, and working-time access jobs. In Lima, Peru, an improved urban arrangements, but there is no evidence regarding the transport system led to significant gains in employ- impact on employment for mature workers (Kuddo ment and earnings for women (but not for men). 2018 and World Bank 2019). This demonstrates that infrastructure investments Tax reform can reduce disincentives to work. This that make it more convenient and safer to use pub- is especially the case for workers with a larger reserve lic transport can generate important labor market 21   For example, see Jensen 2010, 2012, Alzúa et al. 2021, OECD 2021. Information on trade-specific earnings, for example, can have substan- tial effects on women’s choices to enter comparatively high-earning trades traditionally reserved for men (Gassier et al. 2022). 22   Cash subsidies can have both positive and negative effects on caregivers’ labor force participation, as income effects may reduce participation. 404 CHAP T ER 13 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E impacts for groups that are more vulnerable to per- and impacts can grow over time, suggesting that design sonal safety threats during travel, that have care ob- and targeting are key (Card et al. 2018). ligations that compel them to work closer to home, Training that is not adapted to older workers’ and that rely heavily on efficient transport modes needs could reduce the value of training. Some studies (Martinez et al. 2019). Better transportation is sim- have found that the training of older workers has ilarly likely to improve mobility for mature workers lower impacts on key dimensions such as learning, who depend on public transportation but who may career development, earnings, or job security, and that also be vulnerable when commuting. older employees can be less ambitious in their training participation (Zwick 2011). The generally low uptake Strengthening human capital of adult training programs is at least partial evidence that many current programs do not fulfill needs. A Subsidizing skills training, and addressing skills-infor- likely reason is that many training programs are not mation asymmetries, is a common strategy for increas- adapted to older workers’ capabilities and learning ing workers’ employability, but requires careful design. styles: notably, they do not provide informal and self- With the right training methods, adults and mature paced learning and work-integrated approaches or workers can learn new skills to good effect (Knowland make use of older workers’ cognitive strengths, such and Thomas 2014). Tailoring programs to serve the as the capacity to direct and act on information needs of older workers will be essential. First, the de- received (Spotlight 13.2) – (Picchio 2021). In coun- sign of training interventions needs to consider train- tries with high informality, training programs that do ing modalities (see below), motivation among individ- not factor in the opportunity cost of attending uals and employers for taking up training offers, and training for workers’ earnings are likely to be less the need to build capacity in the training sector (Glick effective, as are programs that fail to ac-knowledge et al. 2015). Second, the deadweight and substitution how high-stress life circumstances may affect workers’ losses associated with training can be high (McKenzie mental capabilities (Saraf et al. 2018). 2017a) and, on average, many training programs fail to Countries are using different instruments to have any impact. However, results vary tremendously, increase individual access to, and incentives for, life- SPOTLIGHT 13.2 Gray matter – training for the aging brain The way in which the brain learns – develops new skills – differs over the lifecycle. As discussed in Spotlight 13.1 above, the aging brain gets better at using prior experiences and knowledge to find solutions to new problems, and it can even do this much faster than a younger brain. Hence, adapted training methods and support can foster an effective learning process at older age. Older employees get higher returns from informal and directly relevant training and from training contents that build on their ability to apply and use previously acquired knowledge and experiences. Learning programs should therefore exploit rules and regularities and allow for self-paced, job-related, and work-integrated training modes. Source: Picchio 2021, Thomas et al. 2020, Knowland and Thomas 2014. long learning, including voucher schemes and train- effective information and guidance, and a focus on ing subsidies. With individual learning schemes (in- supporting high-quality training matter significantly cluding vouchers, individual learning accounts, and (OECD 2019c). Specific outreach efforts are needed training savings accounts), individuals choose their to target low-skill groups, as well as employers, and own options for skills development (Spotlight 13.3). encourage them to co-invest in skills development Voucher-based systems can help stimulate training, (OECD 2019c, Vodopivec et al. 2019). For example, an especially for workers in non-traditional or less sta- information campaign on a training program in Ger- ble forms of employment. Voucher systems have had many succeeded in stimulating uptake only for work- mixed success, however, especially for hard-to-serve ers below age 45, suggesting that older workers were groups such as low-skilled workers. Experience shows considered less interesting due to limited pay-off time that simple design, adequate and predictable funding, (tenure) or lower ability (Van den Berg et al. 2018). COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAP T ER 13 405 SPOTLIGHT 13.3 Individual learning accounts to foster adult learning Individual learning accounts (ILAs) provide people with direct support through training entitlements in individual accounts and are recommended by the European Union (EU) Council as a policy to facilitate the “twin transition” to a more digitally capable and green Union. The training entitlements offered through ILAs are associated with the individual and as such fully transferrable across transitions between jobs, from job to learning, from employment to unemployment, and between activi- ty and inactivity. Thus, they can be accumulated and used over a lifetime. To deliver higher skills accumulation, these accounts need to be embedded in a broader framework addressing other skills constraints, including the availability of high-quality and relevant training opportunities, information and guidance services, paid training leave, and skills recognition frameworks. To achieve adequate funding, cost-sharing arrangements involving funds from the public sector, private sector, and social partners can be considered. Source: European Union Council 2022 and OECD 2019d. Mature workers’ motivation and aspirations can al. 2018, Brooks et al. 2018). There is evidence that be strengthened with targeted training and organi- older people continue to identify role models in their zational approaches. Experience from entrepreneur- working life just as younger people do; that these ship and employment programs in LMICs shows role models could be significantly younger than the that trainings focused on psychosocial, or “mindset,” workers themselves; and that being able to identify components have contributed to successful labor role models played a significant part in worker com- market (or business) outcomes, sometimes more so mitment and satisfaction (Gibson and Barron 2003). than technical training.23 These trainings focus on, Giving mature workers the opportunity to partici- inter alia, future-oriented and proactive thinking and pate in teams comprising people with varied back- a “can-do” attitude. This type of training, when deliv- grounds and skills could help provide such role mod- ered adequately, could be helpful to mature workers els and thus raise motivation and aspirations; and who may need or want change in their work in order would support higher overall productivity (Göbel to stay engaged longer. Training programs including and Zwick 2013). Generation, a non-profit organi- relatable role models have been successful, for exam- zation focused on fostering productive labor market ple in Chile and Kenya. The exposure to a role model transitions, offers a wide range of support including was, in fact, more important than the information mentoring and networking (Spotlight 13.4). or training the role model provided (Lafortune et SPOTLIGHT 13.4 Generation – facilitating labor market transitions over the working life The nonprofit organization Generation runs programs to close emerging skills gaps by upskilling and reskilling workers, including mid-career workers. Working alongside employers and trainees, Generation aims to foster career changes that benefit employees and employer. In Singapore, for example, Generation has teamed up with Microsoft to train 1000+ unemployed people, two in five of whom were mid-career workers – for entry-level jobs in the tech industry. Students receive technical as well as job-search skills training, work with mentors, and participate in employer networking events. The evidence suggests that Generation helps create sustainable job placements and helps individuals access higher-paying jobs, resulting in income growth for individuals. Source: Generation.org. 23   See for example Anderson-MacDonald et al. 2016, Bloom et al. 2018, 2010, Campos et al. 2017, Glaub et al. 2014, McKenzie and Puerto 2017. 406 CHAP T ER 13 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E Given that mature older adults learn better on allow for flexible solutions that do not unduly limit the job, encouraging mature apprenticeship pro- labor demand. The evidence on such reforms and grams could also be considered. The evidence on employment effects is generally mixed, however. A the impact of apprenticeships and internships sug- broad conclusion is that when extreme forms are gests that classroom training combined with on- avoided, the impact on employment is modest the-job training may have stronger impact on labor (Betcherman 2015, World Bank 2022b). The evidence market outcomes than skills training or on-the-job of payroll tax cuts on employment is likewise varied. training alone (J-PAL 2023). When coupled with Some cross-country studies point to substantial in- high-quality trainers, apprenticeships can be effec- creases in employment (Latin America) and tive in raising employment (Crépon and Premand reductions in informality (Central and Eastern 2018, Hardy et al. 2019). However, the majority of Europe and Latin America). Other country-level programs and evidence are focused on youth with studies of payroll tax cuts, however, show both low to limited relevance for mature workers. no employment effects (Argentina, Chile), and significant employment effects (Colombia, Türkiye) – Demand-side interventions: (Pagés 2017). As a guideline, policy makers could Incentives and remove barriers to hiring strive for a “flexicurity” approach which combines and investing in more workers broader reforms to severance pay and other benefits with permanent contracts that reduce labor costs Policies directed to potential employers focus on while increasing effective protection of workers in increasing incentives to hire mature workers by ad- case of employment shocks (OECD 2019b). For dressing regulatory constraints that raise labor costs, example, reforms need to be combined with policies and on initiatives to foster firm-provided training to that strengthen unemployment benefit systems, increase skills and thereby productivity levels. which remain underdeveloped in many LMICs (ILO 2021). Reduction in social contributions needs to be Closing the gap between labor costs paired with increased financing of social protection and productivity through general taxation (Packard et al. 2019). Wage subsidies are intended to help address Regulatory reform, information, and incentives can both high labor costs and information asymmetries help reduce gaps between mature workers’ earn- about worker competence but require strong ings and their productivity. Japan, the Republic of targeting and timing to deliver benefits. Wage Korea, and Singapore offer examples of countries subsidies reduce the cost of labor for some time that have implemented policies to reduce senior- (compensating for the fact that a worker may be ity wage practices. Korea has actively promoted, more unproductive in the beginning of employ- disseminated, and subsidized performance-based ment) and help employers reduce hiring risks by compensation models. As part of a broader effort providing them with the opportunity to evaluate to foster the hiring of senior workers, Japan and skill levels more comprehensively than is possible in Singapore have introduced financial grants to in- the hiring process. Evidence from North Macedonia centivize firms to incorporate worker performance on wage subsidies (not age-specific) suggests that, into their wage and personnel system (OECD when well designed and targeted, such subsidies can 2019b). Public sector reform can lead the way by help disadvantaged groups overcome information softening seniority wage practices, fully or partially constraints or externalities (Armand et al. 2020). Yet moving towards performance-based pay. they often suffer from low uptake and low sustain- Similarly, governments can reform social con- ability over the long run (McKenzie 2017a). tributions and employment protection legislation Experience from Finland and Belgium suggests that to lower costs while ensuring adequate social pro- wage subsidies specifically targeted to older workers tection support. This can include reforming systems may delay early retirement but carry such for severance pay and different entitlements that of- deadweight losses that they are not cost effective. ten increase automatically with tenure. Labor regu- Subsidized work appears simply to displace lations in several countries have undergone reform unsubsidized work, with little net gain in in the past decades, as policymakers attempt to bal- employment of older workers (Vodopevic et al. ance the need to protect workers with the need to 2019, Boockmann 2015). COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAP T ER 13 407 Addressing ageism ment in updating and upgrading workers’ skills, in- Age anti-discrimination laws (ADL) exist in many cluding among mature workers. Leveraging private developing countries, but weak enforcement may sector support is important to increase available re- render them less effective. An overview from 2008 sources for training, to increase private sector own- showed that laws that prohibit discrimination in the ership of the skills agenda, and to garner the specif- workplace due to age were in place in about 50 coun- ic knowledge and insights of firms regarding their tries (Ghosheh 2008), many of them developing current and future skill needs (UNESCO 2022). countries. High informality or weak enforcement re- Governments in many LMICs have implemented dif- duces the impact of these regulations, however. The ferent incentives, including private sector-financed general evidence on labor law enforcement in devel- approaches (e.g., training funds) and budget-funded oping countries is not conclusive, largely due to the financial incentives, but with mixed results. lack of studies addressing this issue (Ronconi 2019). Training funds have been a common approach Social campaigns to provide information and to provide resources to invest in training, especially raise awareness have proved successful in influenc- for small and medium enterprises (SMEs), but such ing stakeholders toward employing older workers. funds require strong governance, transparency, In Germany, a relatively low-cost local marketing and solid links to the private sector to work well. campaign used banners, interviews, job fairs, and In about half of the world’s developing countries, information brochures to debunk myths and inform there is a national training skills or training fund employers (and others) about the value of mature in place, most of them financed through a payroll workers. The intervention helped increase retention levy on employers.25 These training funds have re- of mature workers (Homrighausen and Lang 2019). ceived criticism on several grounds, however. They More generally, raising awareness about implicit ste- tend to mainly benefit large firms and may repre- reotypes can have significant payoffs: an interven- sent large deadweight loss (subsidizing training tion addressing implicit stereotypes about immi- that would have taken place without support) at the grants among Italian schoolteachers raised school price of higher tax rates. Training funds have also grades for immigrant children (Alesina et al. 2018). sometimes been used to support the general bud- Research undertaken by the World Bank shows that get, or for cross-subsidizing training in informal, social media campaigns can be highly effective in non-contributing firms. Finally, training has often influencing norms and can allow for very granular been delivered through public institutions that are targeting.24 Similarly, “edutainment” approaches less connected with the demand side. Many training that use entertainment media to convey education- funds are now being reformed to align contribu- al narratives, aspirational role models, and “new” tions better with benefits, for example by including norms to a large population may work: evidence government/donors as contributors, or strength- from TV series seeking to improve sexual health ening governance, transparency, and private sector and attitudes toward gender-based violence (GBV) voice in the use of funds (UNESCO 2022). in Sub-Saharan Africa showed strong impacts on Financial incentives to foster training have so far norms and practices (Banerjee et al. 2019a, 2019b). had limited impact on older workers’ employment in HICs. In Germany, a firm-level subsidy program tar- Promoting firm investment in geting workers ages 45 and above in SMEs was found workers’ human capital to have improved job stability and employment by prolonging working life; programs with long dura- Firms can help extend productive working lives by tion had a stronger effect on both jobs and earnings providing mature workers with access to continu- (Singer and Toomet 2013). For tax incentives, both ous skills development, complementary technology, substitution and deadweight losses appear to be sig- and working environments adapted to their needs, nificant, however. Direct subsidies, often structured along with targeted support that reduces health as co-financing arrangements, may be more effec- risks in old age. tive insofar as they facilitate the targeting of specific It is critical but difficult to increase firms’ invest- groups of firms and employees (Picchio 2021). 24   https://blogs.worldbank.org/voices/using-social-media-change-norms-and-behaviors-scale. 25   Estimates based on World Bank Employing Workers 2020. 408 CHAP T ER 13 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E Low-cost information campaigns (within com- national occupational health service to this effect panies, and to companies) may go some way towards (OECD 2019b). Occupational health and safety raising awareness and increasing firms’ own train- programs have been shown to improve risk factors ing initiatives or uptake of public skills programs, for injuries and chronic illnesses, many of which in- especially among SMEs. In many OECD countries, crease with age (Anger et al. 2015). This can include governments run campaigns and provide employers both access to health care and support for healthy with toolkits and guidance material on how to best lifestyles – diet, exercise, and social interaction, address skills and health issues (OECD 2019b). which also affects labor market participation as well Complementary investments in organization- as productivity (Prinz et al. 2018). However, there is al change or infrastructure can also have significant no conclusive evidence of positive impacts of such productivity payoffs. Adapted technology, ergonomic interventions on productivity or labor force partici- office equipment, mixed-age teams, and other adjust- pation, largely due to lack of quality studies (Poscia ments can help workers stay healthy and engaged. A et al. 2016). Some positive examples on which to much-cited example of the significant potential pay- build do exist, however. In the United States, an oc- offs of such policies is the auto manufacturer BMW’s cupational health intervention involving access to Dingolfing plant in Germany, where a small invest- physicians was associated with a significant reduc- ment in ergonomics and work organization (work- tion in both sickness absence and early retirement station rotation) resulted in a 7 percent productivity (Crawford et al. 2009). Interventions could also be increase, exceeded quality targets, and reduced ab- adapted to the informal sector context. A study of senteeism (Spotlight 13.5) – (EBRD 2020). informal sector workers in Thailand showed that Interventions focusing on workplace health safe working practices improved older workers’ should help increase worker retention and increase work ability, suggesting that targeted information/ mature workers’ productivity; however, there is lit- promotion programs could help increase their pro- tle evidence as to what works to encourage these. ductivity (Thanapop and Thanapop 2021). The United Kingdom targets SMEs with a specific SPOTLIGHT 13.5 Firms with age-sensitive management practices Centrica, a British multinational utility company, has long since removed age limits from its apprenticeship program. It addition- ally offers targeted “returnships”—featuring professional assignments, coaching, and mentoring—to professionals aged 50+ who have taken a career break but are interested in returning to work. The German auto manufacturer BMW staffed a production line whose age composition mimicked its expected future workforce (that is, older) and then collaborated with workers to adapt job environment and work rotation to their needs. Changes included ergonomic adaptation, light strengthening exercises, and job rotation to reduce worker strain. The cost of investment in adaptive measures, at 20,000 euros, was negligible, but resulted in a productivity increase of 7 percent, bringing the line on par with produc- tion lines featuring younger workers. In Achmea, a large Dutch financial services company, career advice is provided to an employee every five years after they have reached 45 years of age, and the company provides additional paid leave to mature workers for study. Berner Ltd., a Finnish chemical manufacturer, has implemented a program involving four pillars: training of management; improvement of the work community and environment; maintenance and enhancement of individual work ability; and the creation of a labor reserve. As a result, retirement owing to incapacity has almost been eradicated, and the average retirement age has increased. Success factors include external support from occupational health care providers and individual career planning and performance reviews. Source: EU OSHA 2016, Hannon 2016, Loch et al. 2010. COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAP T ER 13 409 Fostering senior entrepreneurship workers to jobs. Most European public employment Government or NGO-based programs targeting se- services have no specific policies for older workers nior entrepreneurship (start-up as well as existing (European Network of Public Employment Ser- entrepreneurs) are emerging across HICs, but there vices 2019). A first step is to change the age criteria is little evidence on impacts. Evidence from the Unit- for program eligibility, often capped at retirement ed States suggests that high-growth entrepreneurs, age, so that older workers are not automatically ex- even in tech-intensive industries, are middle-aged cluded from different forms of active labor market rather than young and, in many contexts, entrepre- support. A second step would be to fully capacitate neurship programs appear to work best for partic- the employment service workforce in assisting old- ipants with more experience (Azoulay et al. 2020). er workers. This will require information to rectify Yet, despite the importance of senior entrepreneur- ageist views among staff, identifying specific con- ship, there is little evidence on what interventions straints among mature workers (mobility, digital lit- work to encourage senior-led business start-up and eracy, training needs and modalities, ageist norms, growth, especially not from LMICs, where most and others), and stronger capacity within employ- entrepreneurship programs target youth. There is, ment services to reach out to employers for the ben- however, significant evidence on entrepreneurship efit of mature workers. Strengthened employment programs in LMICs that have successfully provid- services could provide tailored information on job ed financial support, business advice, training and opportunities and reskilling opportunities for old- market access – often in combination – to micro- er workers, as well as training in job search skills, entrepreneurs (Jayachandran 2020). Many entrepre- including in digital skills that are often lacking but neurs face a multitude of constraints (aspirations, increasingly essential for the job search. skills, information, finance, networks) and are best Comprehensive approaches involving employ- served by comprehensive programs, although some ers are needed to improve job matching. In Germa- potential high-growth firms may mostly be limited ny, “regional employment pacts” under the “Initia- by credit constraints (McKenzie 2017b). Given se- tive 50plus” were created with employment services niors’ specific needs, interventions could focus on and the private sector. These successfully included removing constraints including (often) lower tech- targeted efforts to reactivate and reintegrate ma- nical skills, reduced tolerance for financial risk, ture workers with more and better jobs. The pacts and shorter time horizon; interventions might also have reached out to firms and other stakeholders provide support to foster entrepreneurial mind-sets and have included targeted profiling, assessments, (Isele and Rogoff 2014). Programs in HICs have fo- specific training measures, internships, and infor- cused on digital skills, connect seniors with younger mation campaigns. These programs have primarily co-investors, provide targeted counseling and en- benefitted younger seniors (average age 54), how- trepreneurship training, and target micro-finance ever (OECD 2019b). programs to this group, much as is done for women Employment services could also help workers or other groups with special needs and constraints move out of arduous occupations to support lon- (Isele and Rogoff 2014). Role models may again ger working lives, although the evidence suggests prove useful for inspiration and knowledge transfer this is a challenging task (OECD 2019b). Belgium, (Omrani amd Ludivine 2019). for example, provides a subsidy to workers who transition into a less arduous occupation with the Intermediation interventions: same employer and lose income as a result (OECD Strengthening the capacity of employment 2018). Retraining and job search assistance can services to improve matching also be specifically targeted to groups particularly likely to exit the workforce early due to working Public and private intermediation services can be conditions (OECD 2019c). strengthened to improve the matching of older 410 CHAP T ER 13 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CONCLUSIONS: SOME META-LESSONS FOR PRODUCTIVE LONGEVITY The world’s population is aging rapidly in both poor and rich countries, calling for broad policy reform to main- tain welfare levels across generations. This chapter has focused on the set of policies that may help increase “productive longevity” by maintaining mature workers in the active labor force and increasing their productiv- ity. These objectives can be considered “low-hanging fruit”: it is likely considerably easier to assist and convince active workers to stay in the labor market, with incremental efforts to strengthen human capital and tweak incentives, than to activate and train people out of long-term unemployment or inactivity towards the end of the potential working life. As our review shows, the research agenda is only beginning to develop, especially for low- and middle- income countries, and the evidence is still thin in many areas. With this important caveat in mind, a few concluding observations follow. All countries, rich and poor, will need to put pro- across the entire population. Poorer countries face ductive longevity on the policy agenda. Many MICs different and perhaps starker intergenerational pol- are already dealing with aging, even shrinking, icy tradeoffs than do richer countries. Many low- populations; both LICs and MICs will face rapidly er-income countries’ older populations are char- growing numbers of elderly in the future, and the acterized by low levels of human capital and large shift from young to older populations can happen gaps relative to younger generations. Significantly quickly. A wide range of policy instruments are raising productivity levels of the mature workforce needed to protect welfare levels in an aging world. through skills investments may be very costly, espe- Prolonging working lives (for a particular set of cially in countries with limited resources for invest- workers) and increasing the productivity of mature ment. Some countries thus may focus on investing workers (across the board) is part of this agenda. heavily in the human capital of children and youth The optimal policy mix will differ depending (early childhood development, schooling) as well on the country context. Mature workers, especial- as younger adults (life-long learning). For the old- ly in poorer countries, are a heterogeneous group, er population, policy will center on strengthening with many workers, often the majority, working human capital with health interventions and stra- informally and in low-productivity occupations. tegic, demand-led skills development, supporting Given high informality and weak enforcement, the businesses operated by mature workers, and regulatory reform to social insurance, labor regu- providing social safety nets. lations, and taxation will only affect a smaller seg- Even with these constraints, countries can ment of workers, as will initiatives directed only at work to identify “win-win” policies that benefit all formal firms. In this dual labor market, however, generations. For example, providing childcare and it is this (relatively small) group of formal sector long-term care for the elderly, thus alleviating the workers that can reasonably be expected to work care burden, can stimulate labor market participa- longer. For the remainder of the mature workforce, tion and the move to more productive work in both policy needs to focus on how to avert old-age pov- older and younger workers, in particular women. erty by providing social protection and raising the Facilitating access to work, for example with safer productivity of their work. transportation systems, will have benefits for vul- Poorer countries can learn from richer coun- nerable groups in general. tries’ past mistakes and design social protection and Productive longevity does not affect all groups regulatory frameworks adequately for an aging pop- similarly, and more nuanced analysis is also need- ulation. Many LMICs are now developing their so- ed. This chapter is a first attempt at raising policy cial insurance systems and putting in place taxation areas and mapping out the associated evidence at and employment regulatory frameworks. Even with an aggregate level. Even so, the need for a differen- low coverage rates, poorly designed systems can be tiated approach shines through, perhaps especially extremely costly. By acting preemptively, countries from a gender perspective. Women live longer than can avoid repeating mistakes that have taken HICs men but generally have lower pensions. They are decades to unravel with reforms that are proving dif- less likely to be in paid work, have lower average ficult and politically costly. salaries, and face more obligations to provide infor- The impact and cost effectiveness of produc- mal care. There is even some evidence that wom- tive longevity policies need to be weighed against en are worse affected by ageism. At the same time, the payoff to investing in raising human capital there is some evidence from the United States that, COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAP T ER 13 411 although age-friendly jobs in the country have in- under development. Where available, the evidence creased, lower-skilled male workers have been left is often mixed, pointing to the challenges associ- out of employment gains (Acemoglu et al. 2022). ated with this policy agenda. Much less is known Looking in detail at how gender, skills, location, about what works for a developing-country context and other factors affect productive longevity will be characterized by, inter alia, higher informality and essential to develop adequate policies. less-developed social protection systems, higher More evidence is needed as to what works for shares of self-employment, lower skill and pro- productive longevity. As this review shows, most ductivity levels, smaller firms, and different norms evidence for what works in terms of prolonging around co-habitation and care responsibilities. To working lives is based on HICs, which have faced identify the best responses to the aging challenge, aging pressures for a longer time. Even in these a comprehensive research agenda will need to be countries, the knowledge base is still very much developed, and soon. REFERENCES 1. Acemoglu, Daron, and Pascual Restrepo. 2022. “Demographics IZA Journal of Labor Economics, 10(3). and Automation.” The Review of Economic Studies, 89(1):1–44. 12. Atkinson, Carol, and Peter Sandiford. 2016. “An Exploration 2. Acemoglu, Daron, Nicolaj Søndergaard Mühlbach, and An� - of Older Worker Flexible Working Arrangements in Small- drew Scott. 2022. “The Rise of Age Friendly Jobs.” NBER Work- er Firms: Older Worker FWA.” Human Resource Management ing paper Series, No. 20463. Boston: National Bureau of Eco- Journal, 20(1): 12-28. nomic Research. 13. Attanasio, Ozario, Richardo Paes de Barros, Pedro Carneiro, 3. Ahlqvist, Victor, and Erling Borén. 2017. “Leaving Already? The David Evans, et al. 2017. “Impact of Free Availability of Public Swedish Unemployment Insurance as a Pathway to Retire- Childcare on Labour Supply and Child Development in Brazil.” ment.” Masters Dissertation, Lund University. 3ie Impact Evaluation Report 58. New Delhi: 3ie. 4. Alesina, Alberto, Michela Carlana, Eliana La Ferrara, and Paolo 14. Auer, Peter, and Mariàngels Fortenu. 2000. “Ageing of the La- Pinotti. 2018. “Revealing Stereotypes: Evidence from Immi- bour Force in OECD Countries: Economic and Social Conse- grants in Schools.” NBER Working Paper Series, No. 25333. Bos- quences.” Employment Paper 2000/2. Geneva: International ton: National Bureau of Economic Research. Labour Organization. 5. Allel, Kasim, Ana León, Ursula Staudinger, and Esteban Calvo. 15. Azoulay, Pierre, Benjamin Jones, J. Daniel Kim, and Javier Mi- 2019. “Healthy Retirement Begins at School: Educational Differ- randa. 2020. “Age and High-Growth Entrepreneurship.” Ameri- ences in the Health Outcomes of Early Transitions into Retire- can Economic Review: Insights, 2(1):65-82. ment. Ageing and Society, 41(1): 137-157. 16. Backes-Gellner, Uschi, Martin R. Schneider, and Stephan Veen. 6. Almeida, Rita, Jere Behrman, and David Robalino (eds). 2012. 2011. “Effect of Workforce Age on Quantitative and Qualitative “The Right Skills for the Job? Rethinking Training Policies for Organizational Performance: Conceptual Framework and Case Workers.” Human Development Perspectives. Washington, Study Evidence.” Organization Studies, 32(8):1103-1121. D.C.: World Bank. 17. Backhaus, Andreas, and Mikkel Barslund. 2021. “The Effect of 7. Alzúa, Maria, Soyolmaa Batbekh, Altantsetseg Batchuluun, et Grandchildren on Grandparental Labor Supply: Evidence from al. 2021. “Demand-Driven Youth Training Programs: Experi- Europe.” European Economic Review, 137(C). mental Evidence from Mongolia.” World Bank Economic Review, 18. Baird, Sarah, David McKenzie, and Berk Ozler. 2018. “The Effects 32(3): 720-744. of Cash Transfers on Adult Labor Market Outcomes.”World Bank 8. Anderson-MacDonald, Stephen, Rajesh Chandy, and Bilal Zia. Policy Research Working Paper Series, No. 8404. Washington, 2016. “Pathways to Profits Identifying Separate Channels of D.C.: World Bank. Small Firm Growth through Business Training.” World Bank Pol- 19. Banerjee, Abhijit, Eliana La Ferrara, and V. Orozco Olvera. 2019a. icy Research Working Paper Series, No. 7774. Washington, D.C.: “The Entertaining Way to Behavioral Change: Fighting HIV with World Bank. MTV.” World Bank Policy Research Working Paper Series, No. 9. Anger, W. Kent, Diane Elliot, Todd Bodner, Ryan Olson, et. al. 8998. Washington, D.C.: World Bank. 2015. “Effectiveness of Total Worker Health Interventions.” Jour- 20. ______. 2019b. “Entertainment, Education, and Attitudes nal of Occupational Health Psychology, 20(2):226–247. Toward Domestic Violence.” AEA Papers and Proceedings, 10. Armand, Alex, Pedro Carneiro, Federico Tagliati, and Yiming 109:133–37. Xia. 2020. “Can Subsidized Employment Tackle Long-Term Un- 21. Barrett, Alan, Irene Mosca, and Brendan J. Whelan. 2015. “How employment? Experimental Evidence from North Macedonia.” Well-informed are Pension Scheme Members on their Future IZA Discussion Paper Series, No.13478. Bonn, Germany: Insti- Pension Benefits? Evidence from Ireland.” Journal of Aging and tute of Labor Economics. Social Policy, 27(4): 295-313. 11. Ardito, Chiara. 2021. “The Unequal Impact of Raising the Retire- 22. Barrientos, Armando. 2011. “Participation and Earnings of Older ment Age: Employment Response and Program Substitution.” People in Argentina: Nice Work if You Can Get It?” The Journal of 412 CHAP T ER 13 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E Development Studies, 47(7):1061-1079. sitions? Direct and Indirect Effects of a Subsidized Apprentice- 23. Betcherman, Gordon. 2015. “Labor Market Regulations : What ship Program.” Policy Research Working Paper Series, no. 8561. Do We Know About Their Impacts in Developing Countries?” Washington, D.C.: World Bank. The World Bank Research Observer, 30(1): 124-153. 41. D’Amico Ronald, K. Dunham, Annelies Goger, Charles Lea, et 24. Bloom, Nicholas, Aprajit Mahajan, David McKenzie and D. John al. 2009. “Findings from a Study of One-stop Self-Services: A Roberts. 2010. “Why Do Firms in Developing Countries Have Case Study Approach.” ETA Occasional Paper, 2011-19. Wash- Low Productivity?” American Economic Review: Papers & Pro- ington, D.C.: U.S. Department of Labor, Employment and ceedings, 100(2): 619–623. Training Administration. 25. _____. 2018. “Do Management Interventions Last? Evidence 42. Del Boca, Daniela, and Marinela Locatelli. 2006. “The Determi- from India.” World Bank Policy Research Working Paper, No. nants of Motherhood and Work Status: A Survey.” IZA Discus- 8339. Washington, D.C.: World Bank. sion Paper Series, No. 2414, Bonn, Germany: Institute of Labor 26. Böheim, René, and Thomas Nice. 2019. “The Effect of Early Re- Economics. tirement Schemes on Youth Employment.” IZA World of Labor 43. Deloitte. 2018. “The Rise of the Social Enterprise: 2018 Deloitte 2019, 70(2). Global Human Capital Trends.” Deloitte Insights. 27. Boissonneault, Michael, Jaap O. Mulders, Konrad Turek, and 44. Desjardins, Richard, and Anne Warnke. 2012. “Ageing and Skills: Yves Carriere. 2020. “A Systematic Review of Causes of Recent A Review and Analysis of Skill Gain and Skill Loss Over the Lifes- Increases in Ages of Labor Market Exit in OECD Countries.” PLoS pan and Over Time.” OECD Education Working Paper Series, No. ONE 15(4): e0231897. 72. Paris: OECD Publishing. 28. Boockmann, Bernhard. 2015. “The Effects of Wage Subsidies for 45. Doerwald, Friederike., Sussane Scheibe, Hannes Zacher and Older Workers.” IZA World of Labor 201, 189. Nico Van Yperen. 2016. “Emotional Competencies Across 29. Bosch, Mariano and Jarret Guajardo. 2012. “Labor Market Im- Adulthood: State of Knowledge and Implications for the Work pacts of Non-Contributory Pensions: The Case of Argentina’s Context.” Work, Aging and Retirement, 2(2). Moratorium.” IDB Working Paper Series, No. 366. Washington, 46. Dohmen, Thomas. 2003. “Performance, Seniority and Wages: D.C: Interamerican Development Bank. Formal Salary Systems and Individual Earnings Profiles.” IZA 30. Brooks, Wyatt, Kevin Donovan, and Terence R. Johnson. 2018. Discussion Paper Series, No. 935. Bonn, Germany: Institute of “Mentors or Teachers? Microenterprise Training in Kenya.” Labor Economics. American Economic Journal: Applied Economics, 10(4): 196-221. 47. Donovan, Kevin, Will Jianyu Lu, and Todd Schoellman. 2020. “La- 31. Busch, Fabian, Robert Fenge, and Carsten Ochsen. 2021. “Do bor Market Dynamics and Development.”Yale Economic Growth Firms Hire More Older Workers? Evidence from Germany.” CE- Center Discussion Paper Series, No. 1079. Yale University. Sifo Working Paper Series, No. 9219. Munich, Germany: Munich 48. European Bank for Reconstruction and Development (EBRD). Society for the Promotion of Economic Research – CESifo. 2020. “Economic Inclusion for Older Workers: Challenges and 32. Bussolo, Maurizio, Johannes Koettl, and Emily Sinnott. 2015. Responses.” London: European Bank for Reconstruction and “Golden Aging: Prospects for Healthy, Active, and Prosperous Development. Aging in Europe and Central Asia.”Washington, D.C.: World Bank. 49. Economic Commission for Latin America and the Caribbean/ 33. Butrica, Barbara, Richard W. Johnson, Karen E. Smith, and Eu- International Labor Organization (ECLAC/ILO). 2018. “Employ- gene Steuerle. 2006. “The Implicit Tax on Work at Older Ages.” ment Situation in Latin America and the Caribbean: Labour Washington, D.C.: The Urban Institute. Market Participation of Older Persons: Needs and Options.” 34. Card, David, Jochen Kluve, and Andrea Weber. 2017. “What Geneva: International Labour Organization. Works? A Meta Analysis of Recent Active Labor Market Pro- 50. Elinder, Mikael, Johannes Hagen, Mattias Nordin, and Jenny gram Evaluations.” Journal of the European Economic Associa- Save-Söderbergh. 2020. “Who Lacks Pension Knowledge, Why tion, 16 (3): 894-931. and Does it Matter? Evidence from Swedish Retirement Savers.” 35. Campos, Francisco, Michael Frese, Markus Goldstein, Leonardo Institute for Evaluation of Labour Market and Education Policy, Iacovone, et al. 2017. “Personality vs. Practices in the Making Working Paper Series, No. 2020:24. Uppsala, Sweden: Institute of an Entrepreneur: Experimental Evidence from Togo.” Draft for Evaluation of Labour Market and Education Policy. paper for the 2017 Center for the Study of African Economies 51. Elgin, Ceyhun, M. Ayhan Kose, Franziska Ohnsorge, and Shu Yu. (CSAE). Retrieved at: https://editorialexpress.com/cgi-bin/con- 2021. “Understanding Informality.” C.E.P.R. Discussion Paper Se- ference/download.cgi?db_name=CSAE2017&paper_id=493. ries, No. 16497. London: Centre for Economic Policy Research. 36. Carlsson, Magnus, and Stefan Eriksson. 2019. “Age Discrimina- 52. European Agency for Safety and Health at Work (EU-OSHA). tion in Hiring Decisions: Evidence from a Field Experiment in 2016. “Healthy Workplaces for All Ages: Promoting a Sustain- the Labor Market.” Labour Economics, 59: 173-183. able Working Life.” Luxembourg: European Agency for Safety 37. Chernina, Yevgenia, and Vladimir Gimpelson. 2022. “Do Wages and Health at Work. Grow with Experience? Deciphering the Russian Puzzle.” IZA 53. European Commission. 2012. “Special Eurobarometer 378: Ac- Discussion Paper Series, No. 15068. Bonn, Germany: Institute tive Aging.” January 2012. https://europa.eu/eurobarometer/ of Labor Economics. surveys/detail/1002. 38. Costa-Font, Joan, and Cristina Vilaplana-Prieto. 2022. “Care- 54. _____. 2015. “Discrimination in the EU in 2015.” Brussels: Di- giving Subsidies and Spousal Early Retirement Intentions.” IZA rectorate-General for Justice and Consumers. http://dx.doi. Discussion Paper Series, No. 15339. Bonn, Germany: Institute of org/10.2838/325154. Labor Economics. 55. European Network of Public Employment Services. 2019. “PES 39. Crawford, Joanne, Richard Graveling, Hilary Cowie, and Ken Strategies in Support of an Ageing Workforce: Study Report.” Dixon. 2010. “The Health Safety and Health Promotion Needs Luxembourg: EU Publishing Office. of Older Workers. Occupational Medicine, 60(3):184-92. 56. European Union Council. 2022. “Council Recommendation on 40. Crépon, Bruno, and Patrick Premand. 2018. “Creating New Po- Individual Learning Accounts.” Official Journal of the European COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAP T ER 13 413 Union. 2022/C 243/03. June 16, 2022. https://eur-lex.europa. of Labor 2014: 107. doi: 10.15185/izawol.107. eu/legal-content/EN/TXT/PDF/?uri=CELEX:32022H0627(03)& 72. Gruber, Jonathan, Kevin Milligan, and David Wise. 2009. “Social qid=1664444877706&from=EN. Security Programs and Retirement Around the World: The Re- 57. Fairlie, Robert, Sameeksha Desai, and AJ Herrmann. 2019. lationship to Youth Employment.” NBER Working Paper Series, “2018 National Report on Early-Stage Entrepreneurship.” Kauff- No. 14647. Boston, MA: National Bureau of Economic Research. man Indicators of Entrepreneurship. Kansas City: Ewing Marion 73. Guha, Pradyut, Bodrul Islam, and Md Aktar Hussain. 2020. “CO- Kauffman Foundation. VID‐19 Lockdown and Penalty of Joblessness on Income and 58. Ferranna, Maddalena, JP Sevilla, Leo Zucker and David Bloom. Remittances: A Study of Inter‐State Migrant Labourers from 2022. “Patterns of Time Use among Older People.” NBER Work- Assam, India.” Journal of Public Affairs, 21 (4): e2470. https://doi. ing Paper Series, No. 30030. Boston, MA: National Bureau of org/10.1002/pa.2470. Economic Research. 74. Gutierrez, Italo and Pierre-Carl Michaud. 2019. “Job Insecurity 59. Fouejieu, Armand, Alvar Kangur, Samuel Romero Martinez, and Older Workers’ Mental Health in the United States.” Health and Mauricio Soto. 2021. “Pension Reforms in Europe: How and Labor Markets, 47(July 11): 71-98. Far Have We Come and Gone?” IMF Fiscal Affairs Department 75. Göbel, Christian and Thomas Zwick. 2013. “Are Personnel Mea- DP/2021/016. Washington, D.C.: International Monetary Fund. sures Effective in Increasing Productivity of Old Workers?” La- 60. French, Eric, and John B. Jones. 2012. “Public Pensions and bour Economics, 22(June): 80-93. Labor Supply over the Life Cycle.” International Tax and Public 76. Hannon, Kerry. 2016. “5 Workplaces That Embrace Older Finance,19(2): 268–287. Workers.” Forbes. November 18, 2016. https://www.forbes. 61. Gassier, Marine; Lacina Trore, and Léa Rouanet. 2022. “Address- com/sites/nextavenue/2016/11/18/5-workplaces-that-em- ing Gender-Based Segregation through Information Evidence brace-older-workers/?sh=28306dc35914. from a Randomized Experiment in the Republic of Congo.” 77. Hardy, Morgan, Isaac Mbiti, Jamia McCasland and Isabelle World Bank Policy Research Working Paper Series, No. 9934. Salcher. 2019. “The Apprenticeship-to-Work Transition: Exper- Washington, D.C.: World Bank. imental Evidence from Ghana.” World Bank Policy Research 62. Generation. 2021. “Meeting the World’s Mid-career Moment.” Working Paper Series, No 8851. Washington, D.C.: World Bank. https://www.generation.org/wp-content/uploads/2021/07/ 78. Henkens, Kène, Hendrik van Dalen, and Hanna van Solinge. Meeting-the-Worlds-Midcareer-Moment-July-2021.pdf. 2021. The Rhetoric and Reality of Phased Retirement Policies. 63. Geppert, Christian, Yvan Guillemette, Hermes Morgavi, and Public Policy & Aging Report, 31(3): 78-82. David Turner. 2019. “Labour Supply of Older People in Ad- 79. Homrighausen, Pia, and Julia Lang. 2019. “Do Informational vanced Economies: The Impact of Changes to Statutory Re- Nudges Alter Firms’ Hiring Behavior of Older Workers?” [Confer- tirement Ages.” OECD Economics Department Working Paper ence Paper] Beiträge zur Jahrestagung des Vereins für Socialpo- Series, No. 1554. Paris: OECD Publishing. litik 2019: 30 Jahre Mauerfall - Demokratie und Marktwirtschaft 64. Ghosheh, Naj. 2008. “Age Discrimination and Older Workers: - Session: Labor Economics VI, No. D08-V3, ZBW. Kiel, Hamburg: Theory and Legislation in Comparative Context.” Conditions Leibniz-Informationszentrum Wirtschaft. of Work and Employment Series, No. 20. Geneva, Switzerland: 80. International Labour Organization (ILO). 2017. “World employ- International Labour Organization. ment social outlook: Trends for women.” Geneva, Switzerland: 65. Gibson, Donald E., and Lisa A. Barron. 2003. “Exploring the Im- International Labour Organization. pact of Role Models on Older Employees.” Career Development 81. _____. 2021. “World Social Protection Report 2020–22: Social International, 8(4): 198-209. protection at the crossroads ‒ in pursuit of a better future.” Ge- 66. Giesecke, Matthias. and Philipp Jäger. 2021. “Pension Incen- neva, Switzerland: International Labour Organization. tives and Labor Supply: Evidence from the Introduction of 82. Isele, Elizabeth, and Edward G. Rogoff. 2014. “Senior Entrepre- Universal Old-Age Assistance in the UK.” Journal of Public Eco- neurship: The New Normal.” Public Policy & Aging Report, 24 nomics, 203(C). (4): 141–147. https://doi.org/10.1093/ppar/pru043 67. Giles, John, Dewen Wang, and Wei Cai. 2012. “The Labor Supply 83. International Telecommunication Union (ITU). 2021. “Ageing in and Retirement Behavior of China’s Older Workers and Elderly a Digital World – From Vulnerable to Valuable.” Geneva, Swit- in Comparative Perspective.” In Aging in Asia: Findings from New zerland: ITU. http://handle.itu.int/11.1002/pub/818a937f-en. and Emerging Data Initiatives, edited by James P. Smith, and 84. Jayachandran, Seema. 2020. “Microentrepreneurship in Devel- Malay Majmundar. Washinton D.C.: National Academies Press. oping Countries.” NBER Working Paper Series, No. 26661. Bos- 68. Glaub, Matthias E., Michael Frese, Sebastian Fischer, and Maria ton, MA: National Bureau of Economic Research. Hoppe. 2014. “Increasing Personal Initiative in Small Business 85. Jedwab, Remi, Paul Romer, Asif Islam, and Roberto Samaniego. Managers or Owners Leads to Entrepreneurial Success: A The- 2021. “Human Capital Accumulation at Work: Estimates for the ory-Based Controlled Randomized Field Intervention for Evi- World and Implications for Development.”World Bank Policy Re- dence-Based Management.” Academy of Management Learning search Working Paper, No. 9786. Washington, D.C.: World Bank. & Education 13(3): 354-379. 86. Jensen, Robert. 2010. “The (Perceived) Returns to Education 69. Glick, Peter, Crystal Huang, and Nelly Mejia. 2015. “The Private and the Demand for Schooling.” Quarterly Journal of Economics, Sector and Youth Skills and Employment Programs in Low and 125(2): 515–48. Middle-Income Countries.” Washington, D.C.: World Bank. 87. Jensen, Robert. 2012. “Do Labor Market Opportunities Affect 70. Gragnolati, Michele, Ole Hagen Jorgensen, Romera Rocha, and Young Women’s Work and Family Decisions? Experimental Anna Fruttero. 2011. “Growing Old in an Older Brazil: Implica- Evidence from India.” Quarterly Journal of Economics, 127 (2): tions of Population Aging on Growth, Poverty, Public Finance, 753–792. and Service Delivery.” Directions in Development. Washington, 88. Jin, Bora and Lisa Baumgartner. 2019. “Ageism in the Work- D.C.: World Bank. place: A Review of the Literature.” Adult Education Research 71. Graham, Carol. 2014. “Late-life Work and Well-being.” IZA World Conference. https://newprairiepress.org/aerc/2019/papers/14. 414 CHAP T ER 13 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E 89. Johnson, Sheena J., Sabine Machowski, Lynn Holdsworth, ket Policies in Developing Countries? A Critical Review of Recent Marcel Kern, and Dieter Zapf. 2017. “Age, Emotion Regulation Evidence.” World Bank Policy Research Working Paper Series, No. Strategies, Burnout, and Engagement in the Service Sec- 8011. Impact Evaluation Series. Washington, D.C.: World Bank. tor: Advantages of Older Workers.” Revista de Psicología Del 106. _____. 2017b. “Identifying and Spurring High-Growth En- Trabajo Y de Las Organizaciones 33 (3): 205–16. https://doi. trepreneurship: Experimental Evidence from a Business Plan org/10.1016/j.rpto.2017.09.001. Competition.” American Economic Review, 107(8): 2278–2307. 90. J-PAL Policy Insight. 2023. “Vocational and Skill Straining Pro- 107. McKenzie, David, and Susana Puerto. 2017. “Growing Markets grams to Improve Labor Market Outcomes.” Cambridge: Abdul through Business Training for Female Entrepreneurs: A Mar- Latif Jameel Poverty Action Lab. ket-Level Randomized Experiment in Kenya.” IZA Discussion 91. Kalwij, Adriaan, Arie Kapteyn and Klaas Vos. 2010. “Retirement Paper Series, no. 10615, March 2017. Bonn, Germany: Institute of Older Workers and Employment of the Young.” De Econo- for Labor Economics. mist, 158: 341-359. 108. McKinsey Global Institute. 2021. “The future of work after 92. Knowland, Victoria, and Michael Thomas. 2014. “Educating the COVID-19.” McKinsey & Company. https://www.mckinsey. Adult Brain: How the Neuroscience of Learning Can Inform Edu- com/featured-insights/future-of-work/the-future-of-work- cational Policy.” International Review of Education, 60(1): 99-122. after-covid-19. 93. Kuddo, Arvo. 2018. “Labor Regulations throughout the World: 109. Meghir, Costas, and David Phillips. 2010. “Labour supply and An Overview.” Jobs Working Paper, No. 16. Washington, D.C.: taxes.” In Dimensions of Tax Design. Edited by Adam Stuart, World Bank. Timothy Besley, Richard Blundell et al. New York: Oxford Uni- 94. Kyyrä, Tomi and Virve Ollikainen. 2008. “To search or not to versity Press, Inc. search? The effects of UI benefit extension for the older un- 110. Munnell, Alicia and April Wu. 2013. “Will Delayed Retirement employed.” Journal of Public Economics, 92(10-11): 2048-2070. by the Baby Boomers Lead to Higher Unemployment Among 95. Lafortune, Jeanne, Julio Riutort, and José Tessada. 2018. “Role Younger Workers?” CRR working paper 2012-22. Boston: Center Models or Individual Consulting: The Impact of Personalizing for Retirement Research at Boston College. Microentrepreneurship Training.” American Economic Journal: 111. Munnell, Alicia H. and Abigail N. Walters. 2019. “Proposals to Applied Economics 10(4): 222–245. Keep Older People in the Labor Force.” Economic Studies at 96. Lam, David and Ahmed Elsayed. 2021. “The Interrelationship Brookings, January 2019. Washington, D.C.: Brookings Institute. of Growth, Formality, Informality, and Regulation.” In Labour 112. National Center on Senior Transportation (NCST). 2011. “Se- Markets in Low-Income Countries: Challenges and Opportunities. niority Community Service Employment Program: A Resource Oxford University Press. for SCSEP Providers.” Washington, D.C.: National Center on Se- 97. Laun, Lisa. 2017. ”The effect of Age-Targeted Tax Credits on nior Transportation. Labor Force Participation of Older Workers.” Journal of Public 113. Neumark, David, Ian Burn and Patrick Button. 2019. “Is It Harder Economics, 152: 102-118. for Older Workers to Find Jobs? New and Improved Evidence 98. Levin, Victoria. 2015. “Promoting Active Aging in Russia: Working from a Field Experiment.” Journal of Political Economy, 127 (2). Longer and More Productively.”World Bank: Washington, D.C. 114. Ng, Thomas W., and Daniel C. Feldman. 2008. “The Relationship 99. Liebman, Jeffrey. and Erzo Luttmer. 2015. Would People Be- of Age to Ten Dimensions of Job Performance.” Journal of Ap- have Differently If They Better Understood Social Security? plied Psychology, 93(2):392-423. Evidence from a Field Experiment. American Economic Journal: 115. Organization for Economic Cooperation and Development Economic Policy, 7 (1): 275-99. (OECD). 2008. “Improving Financial Education and Awareness 100. Lock, Christoph, Fabian Sting, Nikolaus Bauer, and Helmut of Insurance and Private Pensions.” Paris: OECD Publishing. Mauermann. 2010. “The Globe: How BMW is defusing the 116. _____. 2013. “OECD Employment Outlook 2013.” Paris: OECD demographic time bomb.” Harvard Business Review: Human Publishing. Resource Management. March 2010. 117. _____. 2017: “Pensions at a glance.” Paris: OECD Publishing. 101. Lorenz, Svenja, M. Pfister, Thomas Zwick. 2020. “Beware of the 118. _____. 2018. “Working Better with Age.” Belgium Country employer: Financial Incentives for Employees May Fail to Prolong Note. Paris: OECD Publishing. Old Age Employment.” ZEW Discussion Papers, No. 20-007, Mann- 119. _____. 2019a. “Pensions at a glance. Country profiles: China.” heim: Leibniz-Zentrum für Europäische Wirtschaftsforschung. Paris: OECD Publishing. 102. Martinez, Daniel, Oscar Mitnik, Edgar Salgado, Lynn Scholl, 120. _____. 2019b. “Working Better with Age.” Paris: OECD Publishing. Patricia Yanez-Pagans. 2019. “Connecting to Economic Oppor- 121. _____ 2019c. “Adapting to Demographic Change.” Paper pre- tunity? The Role of Public Transport in Promoting Women’s Em- pared for the first meeting of the G20 Employment Working ployment in Lima.” Journal of Economics, Race, and Policy, First Group under the Japanese G20 Presidency, 25-27 February Online: 26 November 2019. 2019, Tokyo. Retrieved at: https://www.oecd.org/g20/summits/ 103. Mateo Díaz, Mercedes, and Lourdes Rodriguez-Chamussy. osaka/OECD-Ageing-and-Demographic-change-G20-JPN.pdf. 2016. “Cashing in on Education: Women, Childcare, and Pros- 122. _____ 2019d. “Individual Learning Accounts: Design is key for perity in Latin America and the Caribbean.” Latin American success.” Policy Brief on the Future of Work. Paris: OECD Publishing. Development Forum. Washington, D.C.: World Bank and In- 123. _____. 2021. “OECD Employment Outlook 2021: Navigating ter-American Development Bank. the COVID-19 Crisis and Recovery.” Paris: OECD Publishing. 104. Mayhew, Ken and Bob Rijkers. 2004. “How to Improve the 124. Omrani, Nessrine and Ludivine Martin. 2019. “Understanding Human Capital of Older Workers, or the Sad Tale of the Magic Senior Entrepreneur Behavior.” Journal of Enterprising Culture, Bullet.” Paper prepared for the joint EC-OECD Seminar on Hu- 27:259-282. man Capital and Labour Market Performance, held in Brussels 125. Packard, Truman, Ugo Gentilini, Margaret Grosh, Philip O’Keefe, on Dec 8, 2004. https://www.oecd.org/els/emp/34932028.pdf. Robert Palacios, David Robalino, and Indhira Santos. 2019. “Pro- 105. McKenzie, David. J. 2017a. “How Effective are Active Labor Mar- tecting All: Risk Sharing for a Diverse and Diversifying World of COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAP T ER 13 415 Work.” Washington, D.C.: World Bank. 143. Staudinger, Ursula M., Ruth Finkelstein, Esteban Calvo, and 126. Pagés, Carmen. “Do payroll tax cuts boost formal jobs in devel- Kavita Sivaramakrishnan. 2016. “A Global View on the Effects oping countries?” IZA World of Labor 2017, 345. of Work on Health in Later Life.” Gerontologist, 56(S2): 281–292 127. Paweenawat, Sasiwimon W. and Lusi Liao. 2021. “Labor Supply 144. Thanapop, Sasithorn and Chamnong Thanapop. 2021. “Work of Older Workers in Thailand: The Role of Co-residence, Health, Ability of Thai Older Workers in Southern Thailand: a Compari- and Pensions.” ADBI Working Paper 1224. Tokyo: Asian Devel- son of Formal and Informal Sectors.” BMC Public Health 21: 1218. opment Bank Institute. 145. Thomas, Michael S.C., Victoria C.P. Knowland, and Cathy Rogers. 128. Picchio, Matteo. 2021. “Is training effective for older workers?” 2020. “The Science of Adult Literacy.” Social Protection and Jobs IZA World of Labor 2021, 121(2). Discussion Paper Series, No. 2001. Washington, D.C.: World Bank 129. Pit, Sabrina, Malcolm Fisk, Winona Freihaut, Fashola Akin- 146. United Nations Educational, Scientific, and Cultural Organiza- tunde, Bamidele Aloko, Britta Berge, Anne Burmeister, et al. tion (UNESCO). 2022. “Global Review of Training Funds: Spot- 2021. “COVID-19 and the Ageing Workforce: Global Perspec- light on levy-schemes in 75 countries.” Paris: United Nations tives on Needs and Solutions across 15 Countries.” Internation- Educational, Scientific and Cultural Organization. al Journal for Equity in Health 20 (1). https://doi.org/10.1186/ 147. United Nations Population Fund (UNFPA). 2012. “Ageing in the s12939-021-01552-w. Twenty-First Century: A Celebration and A Challenge.” New 130. Pizzinelli, Carlo, and Ippie Shibata. 2022. “Has COVID-19 In- York, NY: United Nations Population Fund. duced Labor Market Mismatch? Evidence from the US and the 148. United Nations (UN). 2007. “World Economic and Social Survey UK.” IMF Working Paper Series, No. 2022/005. Washington, D.C.: 2007: Development in an Ageing World. Department of Social International Monetary Fund. Affairs.” New York, NY: United Nations. 131. Poscia, Andrea, Umberto Moscato, Daniele Ignazio La Milia, 149. _____. 2019. “World Population Aging 2019: Highlights.” New Sonja Milovanovic, Jovana Stojanovic, Alice Borghini, Agnese York, NY: United Nations. Collamati, Walter Ricciardi, and Nicola Magnavita. 2016. “Work- 150. Van den Berg, Gerard J., Christine Dauth, Pia Homrighausen, place Health Promotion for Older Workers: A Systematic Litera- and Gesine Stephan. 2018. “Informing employees in small and ture Review.” BMC Health Services Research 16 (S5): 329. https:// medium sized firms about training: results of a randomized doi.org/10.1186/s12913-016-1518-z. field experiment.” IZA Discussion Paper Series, no. 11963. Bonn, 132. Prinz, Daniel, Michael Chernew, David Cutler, and Austin Frakt. Germany: Institute of Labor Economics. 2018. “Health and Economic Activity Over the Lifecycle: Lite- 151. van der Horst, Mariska. 2019. “Internalised Ageism and Self-Ex- rature Review.” NBER Working Paper Series, No. 24865. Boston, clusion: Does Feeling Old and Health Pessimism Make Individ- MA: National Bureau of Economic Research. uals Want to Retire Early?” Social Inclusion, 7 (3): 27–43. 133. Pricewaterhouse Coopers International Limited (PwC). 2018. 152. Veríssimo, João, Paul Verhaeghen, Noreen Goldman, Maxine “Golden Aging Index.” Available online: https://www.pwc.com/ Weinstein, and Michael T. Ullman. 2022. “Evidence That Age- hu/hu/csr/assets/golden-age-index-2018.pdf. ing Yields Improvements as Well as Declines across Attention 134. Riphahn, Regina T., and Rebecca Schrader. 2021. “Reforms of an and Executive Functions.” Nature Human Behaviour, 6: 97–110. Early Retirement Pathway in Germany and Their Labor Market https://doi.org/10.1038/s41562-021-01169-7. Effects.” IZA Discussion Paper Series, No. 14908. Bonn, Germa- 153. Vickerstaff , Sarah and Mariska Van der Horst. 2021. “The Im- ny: Institute of Labor Economics. pact of Age Stereotypes and Age Norms on Employees’ Retire- 135. Rocha, Romero. 2017. “Aging, Productivity and Wages: Is an ment Choices: A Neglected Aspect of Research on Extended Aging Workforce a Burden to Firms?” Revista Espacios, 38 (39). Working Lives.” Frontiers in Sociology, 6:686645. doi: 10.3389/ 136. Ronconi, Lucas. 2019. “Enforcement of Labor Regulations in fsoc.2021.686645. Developing Countries.” IZA World of Labor 2019: 457. doi: 154. Vodopivec, Milan, Dan Finn, Suzana Laporšek, Matija Vodopiv- 10.15185/izawol.457. ec, and Nejc Cvörnjek. “Increasing employment of older work- 137. Rutkowski, Michal. 2018. “Reimagining Social Protection.” Fi- ers: Addressing labour market obstacles.” Journal of Population nance and Development: December 2018. Washington, D.C.: Ageing, 12: 273–298. International Monetary Fund. 155. World Health Organization (WHO) 2021. “Global Report on 138. Schmieder, Julia. 2021. “Fertility as a Driver of Maternal Em- Ageism.” Geneva: World Health Organization. ployment.” Labour Economics, 72 (October): 102048. https:// 156. World Bank. 2016. “Live Long and Prosper: Aging in East Asia doi.org/10.1016/j.labeco.2021.102048. and Pacific.” Washington, D.C.: World Bank. 139. Singer, Christine, and Ott-Siim Toomet. 2013. “On Govern- 157. _____. 2017. Why Should We Care About Care? The Role of Child- ment-Subsidized Training Programs for Older Workers.” IAB care and Eldercare in Armenia. Washington, D.C.: World Bank. Discussion Paper, No. 21/2013, Nürnberg: Institut für Arbeit� - 158. _____. 2019. “World Development Report 2019: The Changing smarkt- und Berufsforschung. Nature of Work.” Washington, D.C.: World Bank. 140. Saraf, Priyam. 2019. “A Study of Fragility, Entrepreneurship and 159. _____. 2020. “A Silver Lining: Productive and Inclusive Aging Mental Health: Investing in Better Cognitive and Behavioral Skills for Malaysia.” Washington, D.C.: World Bank. for Small Medium Enterprise Entrepreneurs to Thrive in Con- 160. _____. 2021a. “Vietnam - Adapting to an Aging Society.” Wash- flict-Affected Areas of Pakistan.” Washington, D.C.: World Bank. ington, D.C.: World Bank. 141. Söderbacka, Tina, Linda Nyholm, and Lisbeth Fagerström. 161. _____. 2021b. “Aging and the Labor Market in Thailand: Labor 2020. ”Workplace Interventions that Support Older Employees’ Markets and Social Policy in a Rapidly Transforming and Aging Health and Work Ability - A Scoping Review.” BMC Health Ser- Thailand.” Washington, D.C.: World Bank. vices Research, 20 (472). 162. _____. 2022a. “Georgia Human Capital Review.” Washington, 142. Staubli, Stefan and Josef Zweimüller. 2013. “Does raising the D.C.: World Bank. early retirement age increase employment of older workers?” 163. _____. 2022b. “Charting a Course Towards Universal Social Journal of Public Economics, 108(C): 17-32. Protection: Resilience, Equity, and Opportunity for All.” Wash- 416 CHAP T ER 13 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E ington, D.C.: World Bank. 165. _____. 2011. “Why Training Older Employees is Less Effective.” 164. Zwick, Thomas. 2009. “Why Pay Seniority Wages?” ZEW - Centre ZEW - Centre for European Economic Research Discussion Pa- for European Economic Research Discussion Paper No. 09-005. per No. 11-046. http://dx.doi.org/10.2139/ssrn.1886428. http://dx.doi.org/10.2139/ssrn.1394351. Annex 13 TABLE 13A.1 Labor market data for older adults COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E Brazil Chile Colombia Costa Rica Mexico Russia South Africa EU27 OECD Demographic situation Old-age dependency ratio 14% 18% 13% 13% 11% 23% 9% 31% 64% Effective labour force exit age (years) Men 65 68 70 70 70 63 62 64 66 Women 62 63 63 64 66 61 57 63 64 Gender gap in exit (years) 3 5 7 6 4 2 5 1 2 Employment Employment rate, 50-74 (% of the age group) 36 49 51 45 50 38 31 48 51 of which 50-54 62 66 66 66 66 85 57 81 79 55-64 44 56 54 50 52 51 37 60 61 65-69 22 30 35 22 35 7 10 15 23 70-74 10 18 23 15 25 .. .. 7 11 Gender gap in employment, 55-64 (% point gap) 26 32 34 36 36 18 13 12 13 Job characteristics Incidence of part-time work, 55-64 (% of total employment) - 19 17 35 28 13 17 15 19 of which involuntary 55-64 (% of part-time work) - 45 36 - - 27 - 23 20 Incidence of temporary work, 55-64 (% of employees) - 19 18 6 - 6 - 5 8 Incidence of self-employment, 55-64 (% of total employment) 46 29 63 44 47 8 24 18 22 Full-time earnings, 55-64 relative to 25-54 (ratio) - 0.96 1.19 1.26 1.0 - - - 1.06 Dynamics Retention rate for workers currently aged 60-64 (% of employees t-5) 28 43 - 37 46 - - 53 52 Hiring rate 55-64 (% of employees) 34 16 24 14 13 - - 6 8 Joblessness Unemployment rate, 55-64 (% of the labour force) 7.9 6.9 10.5 12.0 2.6 3.6 11.2 5.2 5.2 Incidence of long-term unemployment, 55-64 (% of total unemployment) - - 7.9 7.1 2.6 26.0 68.6 47.9 38.7 Employability Share of 55-64 with tertiary education (% of the age group) 15 16 16 21 15 50 15 26 29 Absolute (% of all employed in the age group) - 33 - - 17 - - 40 41 Relative to employed persons aged 25-54 (ratio) - 0.67 - - 0.50 - - 0.81 0.79 Source: OECD older worker scoreboard, https://www.oecd.org/employment/ageingandemploymentpolicies.htm CHAP T ER 13 417 14 418 CHAPTER 14 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E Adequacy Pensions and Access to Healthcare Maintaining human capital during old age Gustavo Demarco a, Johannes Koettl b Miglena Abels c, and Andrea Petrelli b a Pensions Global Solution Group, World Bank b Social Protection and Labor, Middle East and North Africa, World Bank c Social Protection and Labor, Global, World Bank COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 14 419 INTRODUCTION Maintaining human capital during old age requires both an adequate pension (cash component) as well as ac- cess to affordable quality healthcare (healthcare component). Lack of universal/affordable healthcare and long- term care services for the elderly means governments may need to provide sufficiently higher old age financial assistance to poor elderly to compensate for high out-of-pocket health expenses. While countries with universal1 quality healthcare may be able to ensure a minimum standard of living for the elderly at a lower benefit level, in countries where the elderly face high out of pocket health expenses – pensions may need to be higher to achieve that same minimum living standard.2 Governments should aim to provide a “minimum lev- with benefits defined at poverty line level (Rofman, el of protection” for the disabled and elderly poor that Apella, and Vezza 2015). This chapter adopts the fall outside the reach of contributory social insurance. conclusion from these studies, that expanding non- Given the prevalence of informal jobs in developing contributory pensions to provide all informal work- countries, Bismarkian formal sector social insurance ers with retirement and coverage of other long-term programs (pensions, health insurance, and unem- social insurance risks would require enormous fiscal ployment benefits) have failed to provide sufficient efforts and is not recommended. Although noncon- coverage and protection. As a result, many countries, tributory pensions play a necessary role in providing including low-income countries, have introduced income security in old age for the poorer informal non-contributory mechanisms to cover the informal workers, fiscal constraints will limit both the size of sector and the self-employed. Some countries have this kind of pensions and the income and age to be chosen to cover all elderly, regardless of their income eligible to receive the old age benefit. The study also level at a substantially higher cost, while others have highlights that even if noncontributory pensions are opted for means-tested pensions. Constrained fiscal devised to be affordable now, a major challenge is en- space, multiple competing development objectives, suring that the program remains affordable over time, and a growing number of older people due to demo- given the population aging and the potential for dis- graphic ageing generally means that developing coun- cretionary increases in the benefit amounts. tries could either afford to provide a lower benefit to a In designing non-contributory pensions, pol- larger share of the elderly population, or a higher ben- icy makers should consider questions such as the efit to a more limited share of the elderly population. following: (i) what percentage of the elderly are Determining what level of pensions is “adequate”, poor?; (ii) are elderly poorer (or more vulnerable) what constitutes a “minimum level of protection” and than other age groups?; (iii) what are the living ar- who should be eligible are the biggest design challeng- rangements of the elderly – do they live with other es today, for several well-documented reasons (high family members or alone?; (iv) what is their per cap- costs, incentives to formalize, etc.). A recent study for ita consumption relative to the per capita consump- the South Asia Region (SAR) region (Demarco et al. tion of non-elderly populations; (v) do they spend 2022) presents the challenges to expand social pen- more on healthcare than other age groups?; (vi) do sions in a region where most population in the infor- they spend more on healthcare out of pocket (and if mal sector is poor and needs to rely on non-contrib- not, could it be because there are other barriers in utory programs to provide financial support to the the way to seeking and obtaining healthcare alto- elderly. A similar framework was used in studies for gether); (vii) what benefit level could be considered the Latin America and the Caribbean (LAC) region “adequate”; (viii) what would be the available fiscal discussing the trade-off between the expansion of space to expand benefits – both now and into the non-contributory pensions and the fiscal costs, even future – vis-à-vis competing development needs. 1   It is worthwhile to note that universal contributory health insurance programs, even when they exist, do not always provide sufficient elderly coverage, for instance, health insurance does not always cover long-term care. 2   While the chapter aims to underscore the need for both adequate pension as well as quality healthcare in order to support human capital in old age, it does not explicitly advocate for any single one approach to achieving this goal; it merely posits that in cases where the elderly lack healthcare, non-contributory old age pensions may need to be correspondingly adjusted to ensure the same level of wellbeing as in cases where sufficient/quality/high coverage healthcare is available to the elderly population. The chapter does not argue against broader initiatives that would provide increased affordable quality healthcare to the elderly through funds and risk pooling mechanisms. 420 CHAPTER 14 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E Including the costs of access to quality health solution, but it may take longer to implement. care in the definition of “adequate” pensions, does This chapter is organized as follows: (i) a review not necessarily imply that increased pensions are of current global patterns of non-contributory pen- preferrable to addressing the deficiencies of the sion provision (design, coverage, eligibility criteria, health system; it may just reflect the additional finan- benefit levels, costs) complimented by a discussion cial needs pensioners will face if access to universal around adequacy, equity, and fiscal sustainability (ii) health of good quality is not granted. Improving this a review of healthcare provisions for the elderly with- access is, no doubt, a more efficient and permanent out access to social insurance, (iii) recommendations. FINANCIAL PROTECTION The difficulties to expand contributory Social Insurance in developing countries has driven a growing number of governments to conclude that non-contributory programs are necessary to ensure minimum income protection at older ages. Several countries have moved in that direction in the past decade, notably in response to the im- pact of the COVID-19 crisis on vulnerable groups such as elderly and persons with disabilities. In the last two decades, around 30 countries have introduced some form of a non-contributory benefit targeted to the elderly. Analytical work and policy dialogue supported by development agencies has been consistent with this approach and provided analytical framework to support the policy shift and respond to the fiscal challenge of expanding tax-based pensions (Palacios and Robalino 2019). This chapter sets out to present a general framework to un- derstand the rationale and objectives of the different models. The design and scope of non-contributory programs example, Egypt’s Karama program, provides cash targeted to the elderly varies considerably across the transfers to poorer elderly and persons with disabil- developing world, both in terms of eligibility criteria ities in a targeted manner, while Mexico3 launched a and the level of the benefits (Table 14.1). Some coun- universal program for all above age 65, regardless of tries have chosen to provide a universal pension to their income. Argentina has adopted more complex all individuals above a certain age, while others lim- modalities combining untargeted and targeted ben- it access to those below a certain income level. For efits on a contributory and non-contributory basis. TABLE 14.1 Design and scope of non-contributory programs targeted to the elderly Program Name Targeting Age Albania: Social Pensions pension tested 70 Argentina: Universal Pension for the Elderly pension tested 70 Armenia: State Social Pension pension tested 65 Bangladesh: Old Age allowance means tested 65m/62f Belarus: Social pensions (disability and old age) pension tested 60m/55f Bolivia: Dignity Pension universal 60 Botswana: Old-Age Pension universal 65 Brazil: Beneficio de Prestacao Continuada (BPC / Continuous Cash Benefit) means tested 65 Bulgaria: Old age pension - not related to labor activities means tested 70 Cabo Verde: National Center for Social Pensions means tested 60 Canada: Old Age Security Pension universal, residency 65 Chile: Basic Solidarity Old Age Pension (PBS) pension tested 65 Colombia: Social protection program for the elderly means tested 59m/54f 3   México universal pension: Programa Pensión para el Bienestar de las Personas Adultas Mayores COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 14 421 Program Name Targeting Age Costa Rica: Régimen no contributivo de pensiones por monto básico means tested 65 Ecuador: Pension para Adultos Mayores means tested 65 Egypt: Karama elderly assistance means tested 65 Eswatini: Old Age Grant universal 60 Fiji: Social Pension Scheme pension tested 68 Georgia: Universal old age pension universal 65m/60f Greece: Social solidarity allowance for uninsured elders means tested 67 India: Indira Gandhi National Old Age Pension Scheme means tested 60 Indonesia: Elderly Social Assistance Programme means tested 70/60* Kazakhstan: Basic state pension pension tested 63m/58f Kenya: Older Persons Cash Transfer OPCT means tested 65 Kiribati: Elderly Fund Pension universal 65 Kosovo: Basic Pension (social pension, non-contributory) universal 65 Lesotho: Old age pension pension tested 70 Malaysia: Financial Assistance for Older Person means tested 60 Maldives: Old Age Basic Pensions Scheme pension tested 65 Mauritius: BRP - zero pillar universal-residency test 60 Mexico: Programa Pensión para el Bienestar de las Personas Adultas Mayores universal 65 Mongolia: Social Welfare Pension means tested 60m/55f Namibia: Old Age Pension universal 60 Nepal: Senior Citizens Allowance pension tested 68 New Zealand: Supperannuation universal, residency 65 Panama: 120 to the 65 pension tested 65 Papua New Guinea: New Ireland Social Pension universal regional 60 Paraguay: Pensión Alimentaria para Adultos Mayores means tested 65 Peru: Pension 65 means tested 65 Philippines: Social Pension for the elderly means tested 60 Samoa: Senior Citizens Benefit universal 65 Seychelles: Retirement Pension universal 63 South Africa: Old-age grant means tested 60 Sri Lanka: New Elders Assisstance Program means tested 70 Tajikistan: Social pensions pension tested 65m/60f Tanzania: Zanzibar’s Universal Social Pension universal 70 Thailand: Old Age Allowance pension tested 60 Timor-Leste: Elderly support allowance universal 60 Tonga: Social Welfare Scheme for the Elderly means tested 70 Uruguay: Non-contributory pensions for old age and disability means tested 70 Viet Nam: Monthly subsidy for elderly means tested 60 Source: World Bank Atlas of Social Protection Indicators of Resilience and Equity (ASPIRE). *Indonesia: Elderly Social Assistance Programme 70/60 if chronically ill 422 CHAPTER 14 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E Lower ages of eligibility and/or a lower income age of 60 is eligible (percentages in the graph ex- threshold for means-testing unsurprisingly results ceed 100 percent because the denominator is pop- in a higher count of beneficiaries. Figure 14.1 be- ulation ages 65+ whereas the retirement age is 60). low shows non-contributory pensioners as a share Conversely, coverage in Uruguay, where the age for of the population above age 65. Coverage is highest eligibility is 70 and benefits are means tested, stands in Bolivia, Namibia, and Mauritius where the age of at 15 percent of the population above age 65. Figure eligibility is 60 and the benefit is universal, mean- 14.2 presents the data on beneficiaries of social pro- ing there isn’t a means test and everyone above the grams in LAC countries over time. FIGURE 14.1 Coverage of non-contributory pensions, % population 65+ Source: World Bank ASPIRE dataset FIGURE 14.2 Old age social pension beneficiaries, % population 65+ Source: World Bank ASPIRE dataset COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 14 423 All these models have two main objectives in out of pocket healthcare expenditures suggest that common – to ensure a minimum level of income poor and vulnerable elderly in lower income coun- support to the largest possible number of (poor or tries may be inadequately positioned to maintain a vulnerable) elderly. What that “minimum level” of decent4 and healthy life at older ages. Out-of-pocket income support needed is, and who is “poor”, can expenditure accounts for 48 and 43 percent of to- vary from country to country. Some countries have tal health expenditure in lower middle income and assigned greater importance on expanding cover- low-income countries respectively, whereas it ac- age at the expense of a lower level of pensions, while counts for 32 percent in upper middle-income coun- others have set more strict eligibility conditions (in- tries and only 14 percent in high income countries. cluding higher eligibility ages or income threshold However, increasing the benefit level without levels for means testing) – opting to provide high- carefully defining the pool of potential beneficiaries er benefits to fewer beneficiaries. Governments are may lead to programs quickly becoming fiscally un- faced with the need to balance the potential disin- affordable. A recent study for the SAR region (De- centives to formalization and higher costs that a marco et al. 2022) presents the challenges to expand high level of non-contributory pensions spell with social pensions in a region where most population the potential regressivity and limited impact on in the informal sector is poor and needs to rely on poverty of a low level of benefits (Figure 14.3). non-contributory programs to provide financial While the overall objective of non-contributory support to the elderly. The study concludes that ex- pensions is to ensure a minimum standard of liv- panding noncontributory pensions to provide all ing in old age, some countries provide benefits that informal workers with retirement and coverage of exceed the poverty line, while others only provide other long-term social insurance risks would require a fraction of the poverty line. Across the Organiza- enormous fiscal efforts and is not recommended. Al- tion for Economic Cooperation and Development though noncontributory pensions play a necessary (OECD) countries, means-tested non-contributory role in providing income security in old age for the pensions average around 19 percent of gross aver- poorer informal, fiscal constraints will limit both the age earnings. It is interesting to note however that size of this kind of pensions and the income and age some countries, including Australia, New Zealand, to be eligible to receive the old age benefit. The study and the US have pension amounts for elderly el- also highlights that even if noncontributory pensions igible couples that are less than twice the benefit are devised to be affordable now, a major challenge is amount a single elderly person receives. In Austra- ensuring that the program remains affordable over lia, an eligible couple receives 1.5 times the benefit time, given the population aging and the potential an elderly person living alone receives. In the devel- for discretionary increases in the benefit amounts. oping countries where data is available, universal, Figure 14.5 shows the current and projected share and residency-based pensions average 18 percent of 60+ individuals in total population in South Asia of GDP per capita, whereas means-tested pensions region – most countries are projected to see their average around 10 percent of GDP per capita. The elderly populations double or more as a share of current poverty line is approximately at 20 percent total population. The figure shows expenditure on of GDP per capita in lower-income countries, and a non-contributory pensions – expenditure is high- little lower in middle- and higher-income countries. est in countries with universal coverage and high However, in only 8 of the countries where data is benefits as a share of GDP per capita. Fastest ageing available (n=34), non-contributory old age pensions countries with universal pensions will likely need to equal or exceed 20 percent of GDP per capita (Figure budget a growing, potentially unsustainable, fiscal al- 14.4). Benefits below the poverty line alongside high location to their income support for the elderly. 4   The ILO Convention No. 102 recommends that non-contributory pensions, including means-tested old-age pensions, should guarantee that the provision offered is at least sufficient to maintain the family of the beneficiary in health and decency (Convention No. 102, Art. 67(c)). 424 CHAPTER 14 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E FIGURE 14.3 Benefit amount, % GDP per capita Source: Original calculations for this publication based on World Development Indicators and World Bank ASPIRE dataset COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 14 425 FIGURE 14.4 Expenditure on non-contributory pensions, % GDP (latest year) Source: World Bank ASPIRE dataset FIGURE 14.5 Projected share of 60+ in total population in SAR Source: Original calculations for this publication based on data from World Bank ASPIRE 426 CHAPTER 14 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E ACCESS TO HEALTH AND LONG-TERM CARE SERVICES While rapid progress in medical science and treatments explain the important extension in life expectancy and ag- ing worldwide, health and long-term care costs can quickly become overwhelming. These costs are higher in the last years of life, so older people face a much higher risk of encountering such catastrophic costs that would not only eat up their pension and other social benefits, but also their savings and other assets. So, to fulfill the financial protec- tion mandate, countries not only need to provide pensions but also access to affordable health and long-term care. In many countries, pension systems link to health has a universal national health system. care, and in a few countries even long-term care. There is large consensus on the motivation and Again, this seems to work well in the Bismarckian need to provide healthcare for the elderly, but the systems when the economy is highly formalized concept of adequacy pensions usually misses the and coverage—both in terms of beneficiaries and implications of out-of-pocket healthcare expenditures, contributors—is broad. In these countries, typical- which is in turn dependent on the health care financ- ly social insurance type pension systems are linked ing model and the quality of health service delivery. with social health insurance so that beneficiaries The table below, using the categories pro- of pensions are also covered by—and contributing posed by Scheil-Adlung (2013), includes some to—health insurance. There are some exceptions, additional dimensions which may help us identify notably the US, which has a separate health pro- additional factors to incorporate in the concept of gram for the elderly (Medicare) and the UK, which adequacy pensions. TABLE 14.2 Additional dimensions to identify factors to incorporate in the concept of adequacy pensions Social Health Insurance National Health Traditional National Health Community-based Private Health Services (Bismarckian) Insurance Health Insurance Insurance Coverage All Formal sector All Selective (commu- Selective (voluntary nity members) members) Financing Public budget Contributions Public budget and Local governments Individual contri- (group or national contributions and NGOs butions (low or no risk pooling) risk-pooling) Sufficiency Yes, if all health ser- Yes Yes, if all health ser- Yes, if all health ser- Yes vices are provided vices are provided vices are provided efficiently efficiently efficiently No, if some services No, if some services No, if some services are excluded or the are excluded or the are excluded or the quality is poor quality is poor quality is poor Out of pocket costs Low Low Low Low High Source: Adapted from Scheil-Adlung (2013) The table shows five basic models to provide expense when quality of services is low. health coverage for the elderly, with different finan- National Health Services provided in many cial implications: (i) National Health Services (also countries are the main mechanism of health cover- known as “Public Health”); (ii) Health Insurance (un- age for the poorer (elderly and non-elderly). Coun- der three possible variants: contributory, national or tries with strong public health systems will provide community based), and (iii) private health insurance. health care at null or low cost. This would be the The green shadow means that the probability of only case in which adequacy pensions can be de- the members to incur in costs is null or low, while red fined at a level equivalent to the poverty line (p): means they must incur in costs, or they may need to do so. This may happen because the costs are on the AP = p  [1] members (by design), because the original sources of funding may result insufficient, or because the popu- However, if public health does not cover some lation covered may need to seek alternatives at their treatments, or the services are of poor quality, some COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 14 427 health services, treatments, and medicines may Health programs and overcome some of their limita- need to be paid out of pocket, and the level of ade- tions and challenges. In several countries, NHI and quate pension would change to: CBHI coexist with NHS, but they usually provide additional services and more expedited treatment, AP = p + α nc + β q  [2] especially for targeted groups such as children, el- derly, and persons with disabilities. One of such ex- Where α and β are coefficients to increase the amples is Egypt, where a National Health Insurance level of adequacy pension in the cases of non-covered program was launched to provided poorer groups services (nc) or to compensate for quality deficits (q, with better access to the services provided by the with β≤0). Both parameters are dependent on the pre-existing Public Health hospitals and facilities. probability of occurring and the costs incurred. The case of traditional Health Insurance is dif- The source of financing appears in yellow in the ta- ferent because the health services provided are usu- ble, to reflect the risk of a potential limitation of financ- ally wider and, in general, of higher quality than in ing sources in the future (unsustainable social expendi- the case of public hospitals. However, these services ture), resulting in the need for further adjustments in are usually available to workers in the formal sector, the level of pensions to be considered “adequate”. which in developing countries represent a small per- Similar situations can be observed in the cases centage of the total labor force and usually excludes of National Health Insurance and Community Based the growing group of self-employed, migrant work- Insurance, except that these programs were devel- ers, and other forms of work which are becoming oped in some countries to complement the Public prominent in the world, such as platform workers. MAP 14.1 High shares of informality in developing countries Source: ILOSTAT However, formal sector workers’ health insurance is at (Maps 14.1 - 14.3) and highlighted in multiple publi- risk of becoming unsustainable due to the increasing cations, aging population is happening at a high speed, costs of health and pensions associated with demo- even in young countries, many of which will have large graphic change. As shown in the following figures number of the elderly in the coming decades. 428 CHAPTER 14 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E MAP 14.2 Global Aging in 2020 Source: Adapted from Harasty and Ostemeier (2020) MAP 14.3 Global Aging in 2050 Source: Adapted from Harasty and Ostemeier (2020) COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 14 429 The case of the voluntary private health insur- treatments. A possible solution would be to use the ance is beyond the scope of this chapter. It is lim- OOP variable corresponding to a higher income ited to a small number of higher income groups of group (for example, the lowest income quintile in- the population, so they have no implications on the stead of the lowest decile households). It may be analysis of “adequacy pensions”. useful to incorporate this indicator (or estimate its In summary, minimum, or social pensions level) in the definition of adequacy pensions. should be defined as “adequate” if they are able to Impoverishment5 due to out-of-pocket spend- provide income protection to prevent poverty and ing occurs even in countries where the entire pop- some minimum level of income to provide for health ulation is officially covered by a health insurance expenses at older ages, as shown in equation (2). Ide- scheme or by national or subnational health services ally, access to good quality health services would be (Wagstaff et al. 2018). Out-of-pocket spending on a better solution, but we will assume that this would health can add to the poverty head count and the require more structural reforms of the health system, depth of poverty by diverting household spending which could not be achieved in the shorter term. from non-health budget items. Data from WHO The main problem with that formula is that and World Bank shows that the size of out-of-pocket most of the variables (notably nc, q and the param- expenditures is highest among low and lower mid- eters α and β) are difficult to observe or measure. dle-income countries, followed by middle and upper The Elder Economic Security Standard Index middle-income countries and is lowest among high (Elder Index), developed by the Gerontology Insti- income countries (Figure 14.6). OOP spending on tute at the University of Massachusetts, Boston (El- health care tends to be higher across countries where der Index 2022), provides a measure of the income the share of total public health spending channeled that older adults in the United States need to meet through social security funds is lower. While high their basic needs and age in place with dignity. This OOP spending can add to poverty, low OOP spend- includes not only housing and food but also the cost ing doesn’t necessarily mean a good thing, especially of healthcare. The Elder Index is one example of a if public health spending is lower and the country methodology that facilitates tailoring the adequacy is low-income. Countries with low levels of public measure to the elderly while still providing a relative- spending on healthcare that also report lower OOP ly simple way to evaluate retirement security. Even so, costs suggests that the lower OOP expenditure levels the extension of such methodologies faces important may, to a degree, be resulting from people forgoing data challenges, particularly in developing countries. healthcare altogether at higher rates – lower OOP In the absence of detailed data required for an costs among low-income countries with lower levels accurate measurement of “adequate pensions”, we of health spending channeled through social securi- propose the use of a proxy variable. Out-of-pock- ty funds may signal inadequate financial resources et health expenditures (OOP) have frequently been resulting in healthcare done without. suggested to increase the needs of poor, above the Analysis of household survey data also shows that poverty level. There is abundant information on having any OOP health costs is widespread among OOP, although the data not always discriminates low and lower-middle income countries with around the expenditures by income level and age. If that in- 8 out of 10 households reporting OOP health costs formation was available, then the proper definition across all income quintiles. A deeper look at the house- of adequacy pensions could simply be reduced to: hold composition by age shows that a higher percent- age of households with elderly vs households without AP = p + OOP  [3] elderly report having OOP health expenditures. Fur- ther analysis using the income distribution shows that The caveat of this approach is that “OOP” may un- the largest percentage of households reporting OOP derestimate the needs of health coverage, especially costs fall in the richest quintile (Q5), while the lowest for poorer population, who may not be able to incur percentage of households reporting OOP costs are in so much OOP as their health needs would require, the lowest income quintile (Q1): only 75 percent of precisely because of their limited income, and they households with elderly in the Q1 report having OOP may just not attend medical consultations or stop costs while 87 percent of households with elderly in 5   Impoverishing out-of-pocket metrics include indicators to identify both people impoverished and further impoverished by out-of-pocket health spending, using various poverty lines (e.g., the global extreme poverty line, a relative poverty line). 430 CHAPTER 14 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E FIGURE 14.6 Out-of-pocket expenditure (% of current health expenditure) Source: WHO Data Portal. Q5 report having OOP costs. These findings align percentage points more than households without with both the thinking that healthcare needs increase elderly in the same income quintile. In the richest with age and that lower income individuals may forgo income quintile, households with elderly spend 1.5 care at higher rates for financial reasons. percentage points more than households without Household survey analysis shows that house- elderly in the same income quintile. holds with elderly have 1.3 percentage points high- On average, a household with elderly that be- er OOP health expenditure relative to households longs to the bottom 20 percent spends 5.17 per- without elderly (Table 14.3). Households with cent of welfare (based on household expenditure), elderly in the lowest income quintile spend 1.29 whereas households in Q5 spend only 4.4 percent TABLE 14.3 Comparison of average shares of OOP health expenditure between households with elderly in the poorest quintile versus households with elderly in the richer quintiles (1) (2) (3) (4) (5) VARIABLES sh_hexp_1 sh_hexp_1 sh_hexp_1 sh_hexp_1 sh_hexp_1 hh_older_65 0.0129*** 0.0121*** 0.0149*** 0.0119*** 0.0149*** (0.0029) (0.0017) (0.0019) (0.0011) (0.0010) hh_size -0.0007*** -0.0001 -0.0005*** -0.0004** -0.0008*** (0.0002) (0.0002) (0.0001) (0.0001) (0.0002) Constant 0.0379*** 0.0274*** 0.0288*** 0.0277*** 0.0267*** (0.0015) (0.0014) (0.0009) (0.0009) (0.0007) Observations 50,191 58,894 64,584 71,737 88,508 R-squared 0.0048 0.0072 0.0127 0.0082 0.0143 Note: Robust standard errors in parentheses; *** p<0.01, ** p<0.05, * p<0.1 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 14 431 FIGURE 14.7 Average share of out of pocket health expenditure by quintile and presence of persons age 65+ Source: Original calculations for this publication based on household survey data from World Bank ASPIRE. of welfare – representing a 16 percent higher bur- each family member receives an equal share of the den on poor households with elderly. Comparing overall household welfare. However, elderly mem- average shares of OOP health expenditure between bers may not always receive an equal share, expos- households with elderly in the poorest quintile ver- ing elderly individuals to an even greater vulnera- sus households with elderly in the richer quintiles bility than household survey results suggest. also shows that elderly poor devote a larger share of In summary, households with elderly face higher their income to OOP health costs relative to richer healthcare costs in general across all income groups households with elderly (Figure 14.7). relative households without elderly. Poor households Households with elderly in the poorest income with elderly spend a larger share of their welfare on quintile are more likely by 0.40 percentage points OOP health expenditures. As such, elderly in the to be pushed below the poverty line ($3.2 per day) bottom 40 percent of the income distribution face whereas richer households with elderly report no increased vulnerability as their healthcare needs – impact on their welfare due to OOP health spending. along with higher OOP health costs increase as they In addition, households with elderly and households become older. Findings from this household survey without elderly have a 4 percentage points and 1.5 analysis of low and lower middle-income countries percentage points higher likelihood of incurring at suggest that elderly poor among these countries may least 10 or 25 percent of OOP expenditure on health choose to forgo healthcare at lower rates if they re- respectively as a share of their welfare. ceive an adequate non-contributory pension that Welfare is calculated on a per capita basis at not only covers their basic living expenses but also the household level, which means analysis assumes provides for access to adequate healthcare. 432 CHAPTER 14 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CONCLUSIONS To define a level of pensions as “adequate”, we need to include considerations on access to health care. Well-de- signed programs can either offer a package of pensions plus health coverage or invest in improved public health and long-term care for the elderly. Not doing so exposes many of the programs launched in recent years to the risk of failure. This may be the case both of programs to provide universal pensions (without provisions for healthcare), and of the universal national health care system that is de-linked from any social protection benefits. In the case of social insurance, the link between liance on the general safety net. A separate program pensions and social health insurance for the elderly to provide healthcare for elderly has been adopted in exists in many cases but is concerningly missing in several countries. The US Medicare program is one others. In the cases where health insurance exists, it such example of a program is delinked from pen- sometimes only covers the workers during the active sions, while extended programs for elderly such as life, but not after retiring. The link between pensions the “Programa de Atención Médica Integral” (PAMI) and health programs may prevent the lower income – the Integrated Medical Assistance Program – ad- pensioners from falling in poverty. But, as explained opted in Argentina in the early 1970s only apply to in the main text, this can only apply to formal sector pension beneficiaries (in a fairly laxed manner). workers, and these are a minority of the workforce in Programs designed to expand social pensions most low- and middle-income countries. or non-contributory programs should explicitly The Bismarckian social insurance model has provide enough resources to cover the minimum many advantages and is very effective in highly for- expenses at poverty level, plus the projected out of malized economies, although even among those, so- pocket health expenses. However, the fiscal implica- cietal ageing and the changing nature of work pose tions of providing adequacy pensions are stark and considerable challenges for the pay-as-you go fi- require careful analysis and projections to ensure nancing model going forward. In recent years, an al- fiscal sustainability. The approach proposed in this ternative to the Bismarckian social insurance model chapter may provide further evidence to support has been gaining popularity in the form of basic pen- targeting fiscal support to non-contributory pen- sions, old age allowances, or simply an increased re- sion programs, as opposed to universal benefits. REFERENCES 1. Center for Social and Demographic Research on Aging, Universi- International Labor Organization (ILO). ty of Massachusetts Boston and Gerontology Institute, Universi- 6. Neelsen, Sven, Patrick Hoang-Vu Eozenou, Marc-Francois ty of Massachusetts Boston. 2017. “The National Elder Economic Smitz, and Ruobing Wang. 2022. “The 2022 Update of the Security Standard™ Index: Methodology Overview.” Center for Health Equity and Financial Protection Indicators Database: An Social and Demographic Research on Aging Publications. 16. Overview.” Health, Nutrition, and Population (HNP) Discussion 2. Demarco, Gustavo C., Ernesto Brodersohn, Miglena Abels, Clé- Paper. Washington, D.C.: World Bank. ment Joubert, and Emilio Basavilbaso. 2023. “Social Insurance 7. Packard, Truman, Ugo Gentilini, Margaret Grosh, Philip O’Keefe, among Informal Workers in South Asia.” In Hidden Potential: Re- Robert Palacios, David Robalino, and Indhira Santos. 2019. “Pro- thinking Informality in South Asia, edited by Maurizio Bussolo tecting All: Risk Sharing for a Diverse and Diversifying World of and Sidharath Sharma, 249–318. Washington, D.C.: World Bank. Work.” Washington, D.C.: World Bank. 3. Elder Index. 2022. The Elder Index™ [Public Dataset]. Boston, 8. Palacios, Robert J., and David A. Robalino. 2020. “Integrating MA: Gerontology Institute, University of Massachusetts Boston. Social Insurance and Social Assistance Programs for the Future Retrieved from ElderIndex.org. World of Labor.” IZA Discussion Paper No. 13258. Bonn, Germa- 4. Ghannam, Alaa, and Ayman Sebae. 2021. “Healthcare Protection ny: IZA Institute of Labor Economics. Policies during the COVID-19 Pandemic: Lessons towards the 9. Rofman, Rafael, and Ignacio Apella. 2020. When We’re Sixty-Four: Implementation of the New Egyptian Universal Health Insurance Opportunities and Challenges for Public Policies in a Population Law.” Social Protection in Egypt: Mitigating the Socio-Economic Aging Context in Latin America. Washington, D.C.: World Bank. Effects of the COVID-19 Pandemic on Vulnerable Employment. 10. Rofman, Rafael, Ignacio Apella, and Evelyn Vezza. 2015. “Beyond https://fount.aucegypt.edu/faculty_journal_articles/876/. Contributory Pensions : Fourteen Experiences with Coverage 5. Harasty, Claire, and Martin Ostermeier. 2020. “Population Age- Expansion in Latin America.” Washington, D.C.: World Bank. ing: Alternative Measures of Dependency and Implications for 11. Scheil‐Adlung, Xenia. 2013. “Health Protection: More than Finan- the Future of Work.” ILO Working Paper 5. Geneva, Switzerland: cial Protection.” In Scaling up Affordable Health Insurance, edited COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 14 433 by Alexander S. Preker, Marianne E. Lindner, Dov Chernichovsky, Retrospective Observational Study.” The Lancet Global Health 6 and Onno P. Schellekens, 13–48. Washington, D.C.: World Bank. (2): e180–92. https://doi.org/10.1016/s2214-109x(17)30486-2. 12. Skirbekk, Vegard, Joseph L Dieleman, Marcin Stonawski, Krys- 14. World Bank. 2022. The Atlas of Social Protection Indicators of tian Fejkiel, Stefanos Tyrovolas, and Angela Y Chang. 2022. “The Resilience and Equity  (ASPIRE) dataset. https://www.world- Health-Adjusted Dependency Ratio as a New Global Measure bank.org/aspire. of the Burden of Ageing: A Population-Based Study.” The Lancet 15. World Health Organization (WHO). 2022. Global Health Obser- Healthy Longevity 3 (5): e332–38. vatory Data Portal. https://www.who.int/data/maternal-new- 13. Wagstaff, Adam, Gabriela Flores, Marc-François Smitz, Justine born-child-adolescent-ageing/indicator-explorer-new/mca/ Hsu, Kateryna Chepynoga, and Patrick Eozenou. 2018. “Prog- out-of-pocket-health-expenses-as-percentage-of-total- ress on Impoverishing Health Spending in 122 Countries: A health-expenditure. 434 CHAP T ER 15 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E Prioritizing Action Chapter 15 Priority Setting for NCD Control and Health System Investments Chapter 16 Control of Non-Communicable Diseases for Enhanced Human Capital: The Case for Whole-of-Society Action 15 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAP T ER 15 435 Priority Setting for NCD Control and Health System Investments David Watkins a,b, Sali Ahmed a,b, and Sarah Pickersgill b a Department of Medicine, University of Washington b Department of Global Health, University of Washington 436 CHAP T ER 15 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E INTRODUCTION As underscored in several background papers developed for the HLI Commission, noncommunicable diseases (NCDs) represent a major and growing threat to the health of the world’s aging populations. Unfavorable trends in risk factors like smoking, alcohol use, unhealthy diet, and low physical activity, when combined with accelerations in population growth and aging, paint a picture of a future world dominated by a greater number of NCD cases and deaths in nearly all countries (Foreman et al. 2018; Watkins et al. 2018). Yet many NCD deaths are amenable to simple and relatively inexpensive interventions, such as provision of basic medical care for acute heart attacks, or they are largely preventable, for example through tobacco control policies (Prabhakaran et al. 2018). Ever since the first UN High-Level Meeting in 2011, ficiency in health systems. This chapter seeks to the global NCD agenda has faltered and has suf- provide guidance on the sort of NCD interventions fered from insufficient national and international that could generate high returns for the money resources (Nishtar et al. 2018). Only a handful of spent, thereby improving spending efficiency over countries are on track to achieve the Sustainable the long run. These interventions could be consid- Development Goal (SDG) target 3.4, which calls ered top priorities for health benefits packages and for a one-third reduction in premature mortality universal health coverage systems. We argue that from NCDs between 2015 and 2030 (NCD Count- any new health sector resources available for NCDs down Collaborators 2020). The COVID-19 pan- should be spent on these interventions rather than demic made matters much worse by causing major on ones that are politically popular (e.g., chronic disruptions to health systems. In the aftermath of hemodialysis) or establish new markets for industry the pandemic, the medium-term fiscal and mac- (e.g., novel on-patent chemotherapy agents) but do roeconomic outlook remains uncertain for many not generate much health per dollar spent. low-income countries (LICs) and middle-income We acknowledge that each health system is countries (MICs) because of tepid growth rates and unique and that countries with different resource large government debts made worse by inflationary levels and disease contexts will have different priori- pressures (IMF 2022). ties. The primary objective of this chapter is to pres- The challenge for ministries of health that want ent a framework for setting priorities around NCD to act on NCDs will be to find ways of doing more interventions and demonstrate how it could be used with the same (or potentially fewer) resources. This to pare down a list of candidate interventions to a is fundamentally a challenge to achieve greater ef- manageable set that fits within a given health budget. IDENTIFYING CANDIDATE NCD INTERVENTIONS This chapter builds heavily on previous projects the authors have led. The list of interventions presented here (see Table 15.1 below) is modified from recommendations that came from the Disease Control Priorities Project, specifically, the 3rd edition (DCP3), published between 2015 and 2018 (Jamison et al. 2018). The methods we use for modeling intervention cost and impact are adapted from a report we prepared for the NCD Countdown 2030 Collaborators that provided guidance to countries on how to get back on track to achieve the SDG 3.4 target (NCD Countdown Collaborators 2022). However, this chapter takes a broader view of NCDs than the Count- down report and considers any DCP3 intervention that can reduce NCD mortality, broadly defined, including surgical care and mental healthcare. The annex to this chapter provides an overview of methods. From a cost and volume-of-care standpoint, these those heavily affected by HIV/AIDS). Still, most interventions represent a significant proportion of countries are nowhere near full implementation of what health systems do for NCDs. For example, these interventions. For example, fewer than half cardiovascular disease (CVD) primary prevention, of persons with hypertension worldwide are on diabetes management, and mental healthcare to- treatment, and in LICs like Mozambique the per- gether address the plurality of primary care need centage is closer to 15 percent (NCD Risk Factor for chronic disease in most countries (except for Collaboration 2021). COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAP T ER 15 437 TABLE 15.1  Interventions considered in this chapter Intervention cluster Specific interventions Interventions outside the health sector Tobacco excise taxes (risk factor reduction) Alcohol excise taxes Smoking regulations and information/education/communication Alcohol regulations Sodium regulations and information/education/communication Trans fat bans Outpatient cardiometabolic and Diabetes screening/treatment respiratory disease care Cardiovascular disease primary prevention Aspirin for suspected acute coronary syndromes Cardiovascular disease secondary prevention Heart failure chronic treatment Chronic pulmonary disease treatment Outpatient mental, neurological, and Injection drug use harm reduction measures substance use disorder care Alcohol use screening/brief intervention Depression chronic treatment Bipolar disorder chronic treatment Schizophrenia chronic treatment Epilepsy acute and chronic treatment First-level hospital cardiometabolic and Medical management of acute coronary syndromes respiratory disease care Heart failure acute treatment Treatment of acute exacerbations of chronic pulmonary disease First-level hospital surgical care Screening and treatment of early-stage cervical cancer Management of bowel obstruction Management of appendicitis Repair of hernias Repair of gastrointestinal perforations Referral hospital services Percutaneous coronary intervention for acute coronary syndromes Advanced care for severe acute-on-chronic pulmonary disease Treatment of early-stage breast cancer Treatment of early-stage colorectal cancer We first looked at the cost and impact of scaling plementing the package of interventions could avert up all the interventions in Table 15.1 to achieve 80 up to 180 million deaths (or 2.7 billion disability-ad- percent population coverage in all countries by 2050 justed life-years [DALYs]) by 2050, at an incremen- or sooner. The gap in coverage and thus the “incre- tal cost of US$ 5.4 trillion, translating to US$ 20,000 mental cost” as a share of the current health budget per death averted or US$ 2000 per DALY averted, would be modest for high-performing upper-MICs the latter of which is just over one-third of GDP per but very large for LICs with weak health systems. capita per DALY averted for these countries. Table 15.2 presents our findings. Overall, fully im- 438 CHAP T ER 15 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E TABLE 15.2  Cost and impact of all NCD interventions by 2050, by world region Deaths averted DALYs averted Incre-mental cost Projected public Total cost per Total cost vs. projected through 2050 through 2050 through 2050 Total cost in 2050 spending on health capita 2050 public spending on Country income group (millions) (millions) (USD billions) (USD billions) in 2050 (USD billions) (USD) health in 2050 LICs 12 240 140 17 28 14 59% Lower-MICs 100 1600 1700 200 720 45 27% Upper-MICs 70 860 3600 490 2500 190 20% All LICS and MICs 180 2700 5400 710 3300 87 22% Notes: DALYs = disability-adjusted life-years. Estimates are for a linear scale-up by 2 percent per year to a target of 80 percent coverage by the year 2050. (Not all countries would achieve 80 percent coverage if current coverage is very low.) Incremental costs and health gains are the difference between the status quo, in which coverage is assumed to remain constant, and this scale-up scenar- io. “Public spending on health” is from National Health Accounts data (general government expenditure on health from domestic sources). The total cost is the sum of the incremental cost and the cost of maintaining the status quo. Numbers may not add up due to rounding. While the package overall would be very cost-ef- yond the spending increases that would result purely fective, the budgetary consequences would be con- from economic growth, these countries would strug- siderable. The final column in Table 15.2 shows the gle to afford the entire package of interventions. (We total cost for the year 2050 as a share of projected stress that our cost estimates assume the interventions public-sector spending on health in 2050. This share are all provided to everyone and at no cost at the point would range from 20 percent in upper-MICs to 59 of care, which is in keeping with the progressive uni- percent in LICs. The reason for this discrepancy is versalist approach to health benefits package design that most LICs and lower-MICs are under-spending but is far from reality in many countries that provide on health, so without political action to allocate addi- larger benefits packages to smaller segments of the tional resources to the health sector, i.e., above and be- population or with considerable out-of-pocket costs.) A FRAMEWORK FOR PRIORITIZING NCD INTERVENTIONS In our view, priority setting starts with an assessment of cost-effectiveness. We first looked at the cost-effective- ness of all the clinical interventions in Table 15.1, grouping LICs and MICs according to WHO’s health system typology (Stenberg et al. 2017) to acknowledge differences in implementation capacity. Figure 15.1 summarizes our findings; the analytical methods are detailed elsewhere (NCD Countdown Collaborators 2022) and summa- rized in Annex 15.1 on methods. Figure 15.1 shows that there is considerable vari- ment, chronic depression treatment) would be very ation in intervention cost-effectiveness by health cost-effective. Diabetes care would be the only in- system type—which is correlated with epidemi- tervention that would not be cost-effective in any ological, demographic, and economic differenc- setting. However, diabetes is a special case, because es—underscoring the importance of local analysis most of the health benefits of caring for people with rather than reliance on international literature for diabetes are from CVD prevention (which is cap- cost-effectiveness assessments. About half of the tured in the “CVD primary prevention” and “CVD interventions would be very cost-effective (blue secondary prevention” interventions) rather than tones) in nearly all settings. Some of the surgical glycemic control, which is the focus of this inter- and mental health interventions (e.g., emergent vention. Implementing comprehensive diabetes hernia repair, chronic schizophrenia treatment) care that includes glycemic control and CVD pre- would be on the borderline of cost-effective in most vention would be cost-effective, though more costly settings, whereas others (e.g., appendicitis manage- than doing CVD prevention alone. COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAP T ER 15 439 FIGURE 15.1 Cost per DALY averted for 24 clinical interventions, by health system type Notes: ACS = acute coronary syndrome; IDU = injection drug use; CVD = cardiovascular disease; COPD = chronic obstructive pulmonary disease; PCI = percutaneous coronary intervention; ICER = incremental cost-effectiveness ratio; GDP = gross domestic product. The five health system types (x axis) are based on a typology used by WHO,10 where “conflict” and fragile” include nations in active conflict or who are post-conflict or otherwise at risk of collapse (respectively); “HS1” includes stable countries with low resources, whereas “HS2” and “HS3” include stable countries with medium and high resources, respectively. The thresholds used for cost-effectiveness in the key are based on our review of different proposed cost-effectiveness thresholds in the literature, summarized elsewhere.8 Use of these thresholds here does not imply our endorsement of them. 440 CHAP T ER 15 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E Table 15.1 also includes several interventions able up-front investments or specialist training. Ta- “outside the health sector” that produce health. In ble 15.4 below summarizes our semi-quantitative our view, cost-effectiveness analysis is an incorrect assessment of each intervention’s relative perfor- approach to assessing the efficiency of these inter- mance on these three criteria. The construction of ventions. In a related analysis we did for the Copen- the equity and financial risk protection indicators is hagen Consensus Center, we did benefit-cost anal- summarized in DCP3 volume 9, chapter 3 (Watkins yses of these interventions using standard methods et al. 2017). The construction of the implementation for valuing health gains in economic terms and in- feasibility indicator is summarized in the aforemen- cluding the costs and benefits of the interventions tioned NCD Countdown report (appendix, pp. 13- outside the health sector (Watkins, Ahmed, and 15, NCD Countdown Collaborators 2022). Pickersgill 2023). Table 15.3 below presents bene- The assessments above are intended to serve as fit-cost ratios for these policies, again aggregated by inputs to country-level (or subnational-level) pri- health system type for comparison purposes. ority setting processes. Each health ministry will As the Table demonstrates, all these interven- inevitably place different weight on each of these tions are incredibly cost-beneficial, with tobacco criteria, and most ministries will also consider oth- control providing the most impressive returns. In er criteria alongside these (Eregata et al. 2020). Be- fact, the benefit-cost ratios for tobacco are much low we present a stylized approach that synthesizes higher than most cross-sectional analyses would these criteria in a systematic way and illustrates how suggest (Jha et al. 2012), because the benefits of they can be used to generate a final, high-priority cessation (and prevention of initiation) accumulate package for each type of health system. over decades (Jha et al. 2013). Because all these in- We start by grouping interventions into three terventions are very cost-beneficial and inexpensive categories based on cost-effectiveness: provisional- for governments to implement, we include them by ly high-priority (<0.25x GDP per capita per DALY default in our stylized, locally-tailored, high-priori- averted), provisionally medium-priority (0.25-1.0x), ty packages (below). and provisionally low priority (>1.0x). The provi- Beyond cost-effectiveness, we consider three sionally high-priority category is based on the most other criteria for identifying high-priority health stringent cost-effectiveness threshold proposed in sector interventions for NCDs: equity, financial risk the literature (Ochalek, Lomas, and Claxton 2015). protection, and implementation feasibility. Equity Interventions in this group would, along with the six and financial risk protection have been endorsed by intersectoral interventions (Table 15.3), be the start- WHO (along with cost-effectiveness) as core criteria ing point for most countries’ high-priority packages. for designing health benefits packages (Ottersen and Yet other interventions might merit inclusion Norheim 2014). We view implementation feasibility in a high-priority package on other grounds. For as a critical criterion for NCDs, because many NCD example, interventions in the provisionally medi- interventions are complex and place high demands um-priority group that are attractive in terms of on health systems, and they may require consider- their equity or financial protection characteristics TABLE 15.3  Value for money among interventions outside the health sector Intervention Conflict Fragile HS1 HS2 HS3 Tobacco excise taxes 290 310 170 310 320 Tobacco regulations and IEC 250 270 97 270 300 Alcohol regulations 150 250 110 200 270 Alcohol excise taxes 110 190 52 130 210 Sodium regulations and IEC 64 42 NA 53 130 Trans fat bans 32 50 2.6 71 79 Notes: IEC = information, education, and communication. See footnote to Figure 1 for an explanation of the five health system types. The quantities presented in the cells are benefit-cost ratios from a societal perspective, with costs and benefits discounted at 5 percent and a time horizon of 2023-2050 for the analysis. Sodium policies are “N/A” in HS1 because average sodium consumption in these countries is already at WHO-recommended levels (Pickersgill et al. 2022). COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAP T ER 15 441 (i.e., ++ or +++, per Table 15.4) could be moved to lustration purposes, we moved these interventions the provisionally high-priority group. On the other into the provisionally medium-priority group for hand, interventions in the provisionally high-prior- countries with weaker health systems (i.e., Conflict, ity group that are challenging to implement (i.e., +, Fragile, and HS1 types) but retained them in the per Table 15.4) could be moved to the provision- provisionally high-priority group for countries with ally medium-priority or low-priority group. For il- stronger systems (i.e., HS2 and HS3 types). TABLE 15.4  Non-cost-effectiveness criteria for prioritizing clinical interventions Financial Implementation Intervention Equity impact risk protection feasibility Aspirin for suspected ACS + + +++ Heart failure chronic treatment + +++ ++ Treatment of early-stage breast cancer + ++ + Epilepsy acute and chronic treatment +++ + +++ IDU harm reduction measures ++ ++ +++ Depression chronic treatment + + +++ CVD primary prevention + + +++ Pulmonary rehabilitation + + ++ Heart failure acute treatment + ++ ++ Medical management of ACS + ++ ++ Management of appendicitis + + +++ Asthma/COPD acute treatment + ++ ++ CVD secondary prevention + + ++ Cervical cancer screening and treatment + + ++ Repair of gastrointestinal perforations + + +++ Treatment of early-stage colorectal cancer + ++ + Alcohol use screening/brief intervention + + +++ Management of acute ventilatory failure + ++ + Management of bowel obstruction +++ ++ +++ Repair of hernias + + +++ PCI for ACS + ++ + Asthma/COPD chronic treatment + + ++ Schizophrenia chronic treatment + + +++ Bipolar disorder chronic treatment + + +++ Diabetes screening/treatment + + ++ Notes: ACS = acute coronary syndrome; CVD = cardiovascular disease; IDU = injection drug use; COPD = chronic obstructive pulmonary disease; PCI = percutaneous coronary intervention. Figure 15.2 shows the results of this stylized We stress that categorizing interventions as prioritization exercise. Again, these categorizations medium- or low-priority does not automatically are meant to be illustrative rather than prescriptive. imply they should be de-implemented in countries. As the Figure indicates, such a high-priority pack- Oftentimes, interventions with these characteristics age would contain between 15 and 18 interventions are already being provided, even in low-income for given country and would, include all six inter- countries, although often at low coverage and con- sectoral interventions. siderable out-of-pocket cost. What the high-priori- 442 CHAP T ER 15 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E ty categorization implies is that any new resources hand, some low-priority interventions should be for NCDs (i.e., over and above current spending) de-implemented, or not allowed to enter the system should be devoted preferentially to scaling up these until sufficient cost controls are put in place. Many of interventions to maximal coverage (80 percent or the novel targeted cancer drugs fall into this category. higher) and minimal out-of-pocket cost (20 percent Prioritization of interventions also provides a or lower (McIntyre, Meheus, and Rottingen 2017)). framework for sequencing them over time. In gen- On the other hand, medium-priority interven- eral, intersectoral policies are so effective at preven- tions could be maintained at their current coverage tion and are so inexpensive that they should almost and out-of-pocket cost levels until sufficient resourc- always be rolled out during the earliest possible es become available to scale them up; they might policy window. In the spirit of Primary Health Care also be allowed to persist in the system if they help (Watkins et al. 2018), we argue that interventions achieve longer-term health system objectives (see that are very cost-effective and can be delivered below). Low-priority interventions are a more het- through primary care platforms by non-physician erogeneous category, however. Some might be al- health workers should also be scaled up as early as lowed to persist and financed through cost recovery possible. These two sets of interventions could be (Eregata et al. 2020). One example from our analysis the targets of the bulk of national and internation- would be bipolar disorder treatment. On the other al resources during the Sustainable Development FIGURE 15.2 Final prioritization of NCD interventions, by health system type Note: see footnotes in Figure 1 for acronyms and explanation of health system types. The colors above reflect the authors’ summative judgment about the priority level of each intervention based on value for money and, in the case of clinical interventions, equity, financial protection, and implementation feasibility. COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAP T ER 15 443 Goal period. Less cost-effective outpatient and hos- even if treatment of breast or colorectal cancer are pital-based interventions could be phased in more not high priorities for a country’s benefits package slowly, perhaps with greater emphasis beyond 2030 now, they eventually will be. Delaying the necessary once the first wave of NCD interventions has been upfront investments in health workforce could set sufficiently implemented. this country even further back in the future. The exception to this general approach is spe- The good news is that, compared to the entire cialized NCD care. Building specialist capacity takes list of interventions in Table 15.1, the high-priority many years; if a national government wants to have packages outlined in Figure 15.2 would still gener- a well-functioning cancer system in 2035, it needs ate meaningful health gains but at lower cost. Table to start building physician subspecialty programs 15.5 presents the same estimates of cost and impact and raising capital for large items like new hospi- by country income group as in Table 15.2, but only tals, advanced imaging, and pathology services in for the high-priority packages. Overall, fully imple- 2025. Physicians are often the bottleneck to rolling menting these interventions could avert up to 150 out these programs, as the most specialized ones million deaths (or 2.2 billion DALYs) by 2050, at an (e.g., cardiac surgeons) require a decade or more of incremental cost of US$ 1.3 trillion, translating to postgraduate training. What this means for benefit US$ 5600 per death averted or US$ 620 per DALY packages and national NCD strategies is that a mod- averted, the latter of which is just over 10 percent est amount of new financial resources for NCDs of GDP per capita per DALY averted for this group during a given policy cycle should be set aside for of countries. The budgetary implications of this capacity-building and long-term investments, with package would be more modest, ranging from 6.2 a focus on building the necessary capacity to deliver percent of total public-sector health spending in up- medium-priority interventions. Put more simply: per-MICs to 20 percent in LICs. TABLE 15.5  Cost and impact of locally tailored high-priority NCD packages, by world region Deaths averted DALYs averted Incre-mental cost Total cost Projected public Total cost per Total cost vs. projected through 2050 through 2050 through 2050 in 2050 spending on health in capita 2050 public spending on Country income group (millions) (millions) (USD billions) (USD billions) 2050 (USD billions) (USD) health in 2050 LICs 8.8 180 51 5.5 28 4.6 20% Lower-MICs 82 1300 470 57 720 13 7.9% Upper-MICs 55 650 830 160 2500 61 6.2% All LICS and MICs 150 2200 1300 220 3300 27 6.7% Notes: see footnotes from Table 15.2, which are also relevant to this Table. POLICY IMPLICATIONS In this chapter, we present a framework for selecting interventions to prevent and treat NCDs in LICs and MICs. We look at 30 recommended interventions, which include 24 personal health services—most of which can be deliv- ered through primary healthcare systems—and six “primordial prevention” interventions outside the health sector. We summarize the evidence around the properties of each intervention in the domains of value for money, equity impact, financial risk protection afforded, and feasibility of implementation. We demonstrate how a balanced use of these criteria could allow countries to accelerate process on NCDs, practically doubling or tripling the rate of intervention scale-up compared to the status quo (see annex) and averting 240-280 million deaths by 2050. To achieve these health gains, MICs would need to be mark” spending levels on NCD interventions for the spending at least 6 to 8 percent of their health budgets, years 2030, 2040, and 2050, assuming full implemen- and LICs 20 percent of their budgets, on high-priori- tation of all high-priority and medium-priority inter- ty interventions by 2050. Table 15.6 provides “bench- ventions (see Figure 15.2) by 2050. 444 CHAP T ER 15 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E TABLE 15.6  Total cost per capita for priority NCD interventions by year and country income group 2030 2040 2050 Country income group High High + medium High High + medium High High + medium LICs 1.7 2.7 3.0 4.8 4.6 7.3 Lower-MICs 6.2 10 9.8 16 13 22 Upper-MICs 38 61 52 87 61 100 All LICS and MICs 17 28 23 39 27 45 Notes: Costs are total costs, i.e., sum of current spending and incremental costs required to scale up to target coverage. Costs per capita increase over time because coverage increases at a modest 2 percent per year and because of population aging and a continued epidemiological shift towards NCDs. Numbers may not add up due to rounding. While drugs, consumables, and equipment sis in the global nursing workforce (Buchan and would comprise the plurality of costs, greater in- Catton 2023). An additional 5.8 million physicians vestment in health systems would also be needed. and 79,000 health facilities would also be required We looked at the additional health workforce and for full implementation of the package. These find- health facilities that would be required to fully im- ings suggest that, alongside greater required public plement the high-priority NCD packages (Table contributions to universal health coverage schemes 15.7). The most striking finding from this analysis to cover the cost of human and physical capital, gov- is that an additional 48 million nurses would need ernments also need to invest in expanded pre-ser- to be in the LIC and MIC workforce by 2050. This vice training for the health workforce, especially in finding is consistent with the well-documented cri- building more schools of nursing. TABLE 15.7  Workforce and facility requirements for implementing the high-priority NCD package by 2050 Current density Required increase in Target density in 2050 for Input (most recent year) density by 2050 full implementation Nurses 2.6 per 1000 7.3 per 1000 10 per 1000 Physicians 1.3 per 1000 0.88 per 1000 2.2 per 1000 Health facilities 4.9 per 100,000 1.2 per 100,000 6.1 per 100,000 Notes: Most recent density estimates are for the year 2018 for health workforce and 2013 for health facilities, as per the WHO Global Health Observatory. Results are average values across all LICs and MICs. The latter two columns only reflect the increases needed to implement the NCD package; additional increases would be needed to scale up interventions for communicable diseases, injuries, and other health conditions. It is also possible that COVID-19-related burnout and accelerated retirements from nursing and medicine have reduced the density of health workers and larger increases will be required. Still, these interventions could substantially ad- for individuals in their early 60s, and 57 percent to vance the healthy aging agenda. Figure 15.3 shows 71 percent for individuals in their early 80s. Rough- how far the intervention packages could get LICs ly speaking, across all age groups, the HLI package and MICs towards the avoidable NCD mortality would get the world about halfway to the mortality frontier (Chang et al. 2024) by the year 2040. (In rates observed in the best-performing countries. this analysis, the frontier is defined as the 20th per- Still, there is a significant gap between achievable centile cause-specific mortality rate rather than the and avoidable mortality. Part of the reason the pack- lowest observed rate.) age has a greater impact at older ages (Figure 15.3) is As the figure shows, the full NCD package and that a greater share of NCD deaths at older ages are the high-priority subset would accelerate movement from causes addressed by the HLI package, where- towards the NCD mortality levels projected for the as at younger ages there are other drivers of NCDs best-performing countries. Compared to the base- (e.g., liver disease from chronic infections, rheumatic line/status quo scenario, the progress towards the heart disease, cancers other than those featured in frontier would be 22 percent to 28 percent for indi- the HLI package, etc.). Additionally, it is possible that viduals in their early 40s, 38 percent to 47 percent there are other risk factors that are not addressed by COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAP T ER 15 445 FIGURE 15.3 Contribution of NCD interventions towards avoidable all-cause mortality the HLI package interventions that are driving NCD cially inexpensive to implement, require spending deaths in younger ages; for example, diets high in of political capital and anticipatory management of saturated fats and refined carbohydrates and low in relationships with food, beverage, and tobacco in- fiber have been documented to account for early-on- dustries that will probably lobby against them. On set colorectal cancer (Carroll et al. 2022), which is a the other hand, bringing industry to the table can in rapidly increasing problem in most countries (GBD some cases result in more ambitious policy targets 2019 Diseases and Injuries Collaborators 2020). and better implementation. For example, the South In addition to reducing adult mortality (the fo- African government engaged with industry from the cus of this chapter), these interventions could im- very beginning of its salt reformulation regulations, prove nonfatal outcomes by reducing the incidence resulting in mandatory (as compared to voluntary) of disease overall and by reducing the severity of targets and encouraging preliminary findings on chronic illness, especially for mental and substance changes in dietary sodium consumption (Charlton use disorders. Still, there are other intervention op- et al. 2021). Additionally, doing more on tobacco tions for countries seeking to improve nonfatal out- and alcohol taxes can pave the way for develop- comes, such as post-stroke rehabilitation and cus- ing a more comprehensive fiscal policy approach todial care for individuals with dementia. Assessing to NCDs that includes, for example, introduction a broader set of interventions around healthy aging of sugar-sweetened beverage taxes and removal of was also outside the scope of this work but will be subsidies on high-energy density agricultural com- an important area of emphasis in the upcoming 4th modities. We note that sugar-sweetened beverage edition of Disease Control Priorities. taxes appear to be cost-beneficial, but their impact Identifying and prioritizing interventions is not on population health is modest compared to alcohol enough: countries also need to implement them. and tobacco taxes (Watkins et al. 2019). This chapter does not cover implementation con- We acknowledge several important limitations siderations except tangentially (see Table 15.4), but of this work, most of which were reviewed in the these issues are discussed in the aforementioned NCD Countdown report (NCD Countdown Col- NCD Countdown report (NCD Countdown Collab- laborators). We did not have sufficient data to im- orators). That report also contains a detailed assess- plement nonlinear cost functions or model econ- ment of the financing landscape for NCDs and pro- omies of scope or scale, so our cost estimates will vides some general guidance for ministries of health differ from the real world. We are still uncertain as well as the development assistance community. about the effect of the COVID-19 pandemic on The intersectoral interventions, while finan- NCD incidence and mortality as well as health sys- 446 CHAP T ER 15 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E tem resources in the medium term. We do not know tions in NCD mortality, helping them to achieve much about the current coverage level of most of the SDG 3.4 target and post-SDG targets for NCDs. these interventions in many countries, requiring We anticipate that this guidance could help bring us to employ generic assumptions about the status a necessary focus and fiscal discipline to national quo scenario as well as the rate of scale-up required NCD strategies and plans. Additionally, our list of to achieve 80 percent population coverage by 2050. interventions and their associated health system in- These data gaps hinder our ability to produce pre- vestments could be used as benchmarks for future cise estimates of incremental costs and health gains. World Bank lending efforts to support NCD pro- In summary, we provide a set of recommen- gram development in lower-resource countries. dations to help countries achieve significant reduc- REFERENCES 1. Brouwer, Elizabeth D., David Watkins, Zachary Olson, Jane Administration. https://fda.report/media/116833/Food-Label- Goett, Rachel Nugent, and Carol Levin. 2015. “Provider Costs ing--Nutrition-Labeling-of-Standard-Menu-Items-in-Restau- for Prevention and Treatment of Cardiovascular and Related rants-and-Similar-Retail-Food-Establishments---Final-Regula- Conditions in Low- and Middle-Income Countries: A System- tory-Impact-Analysis.pdf (accessed 9 December 2022). atic Review.” BMC Public Health 15 (1). https://doi.org/10.1186/ 8. Foreman, Kyle J, Neal Marquez, Andrew Dolgert, Kai Fukutaki, s12889-015-2538-z. Nancy Fullman, Madeline McGaughey, Martin A Pletcher, et al. 2. Buchan, James, and Howard Catton. 2023. “Recover to Rebuild: 2018. “Forecasting Life Expectancy, Years of Life Lost, and All- Investing in the Nursing Workforce for Health System Effec- Cause and Cause-Specific Mortality for 250 Causes of Death: tiveness.” Geneva, Switzerland: International Council of Nurses. Reference and Alternative Scenarios for 2016–40 for 195 Coun- https://www.icn.ch/sites/default/files/2023-07/ICN_Recov- tries and Territories.” The Lancet 392 (10159): 2052–90. https:// er-to-Rebuild_report_EN.pdf. doi.org/10.1016/s0140-6736(18)31694-5. 3. Carroll, Kaitlin L., Andrew D. Frugé, Martin J. Heslin, Elizabeth 9. GBD 2019 Diseases and Injuries Collaborators, Cristiana Ab- A. Lipke, and Michael W. Greene. 2022. “Diet as a Risk Factor for bafati, Kaja M. Abbas, Mohammad Abbasi, Mitra Abbasifard, Early-Onset Colorectal Adenoma and Carcinoma: A Systematic Mohsen Abbasi-Kangevari, Hedayat Abbastabar, et al. 2020. Review.” Frontiers in Nutrition 9 (June). https://doi.org/10.3389/ “Global Burden of 369 Diseases and Injuries in 204 Countries fnut.2022.896330. and Territories, 1990–2019: A Systematic Analysis for the 4. Chang, Angela Y., Gretchen A. Stevens, Deigo S. Cardoso, Bo- Global Burden of Disease Study 2019.” The Lancet 396 (10258): chen Cao, and Dean T. Jamison. 2024. “The Economic Value of 1204–22. https://doi.org/10.1016/S0140-6736(20)30925-9. Avoidable Mortality.” In Unlocking the Power of Healthy Longev- 10. International Monetary Fund (IMF). 2022. World Economic Out- ity: Compendium of Research for the Healthy Longevity Initiative. look, October 2022: Countering the Cost-of-Living Crisis. Vol. 2022 Washington, D.C.: World Bank. (002). Washington, D.C.: International Monetary Fund. 5. Charlton, Karen E., Barbara Corso, Lisa Ware, Aletta E. Schutte, 11. Jamison, Dean T, Ala Alwan, Charles N Mock, Rachel Nugent, Leanda Wepener, Nadia Minicuci, Nirmala Naidoo, and Paul David Watkins, Olusoji Adeyi, Shuchi Anand, et al. 2018. “Uni- Kowal. 2021. “Effect of South Africa’s Interim Mandatory Salt versal Health Coverage and Intersectoral Action for Health: Reduction Programme on Urinary Sodium Excretion and Blood Key Messages from Disease Control Priorities , 3rd Edition.” The Pressure.” Preventive Medicine Reports 23 (September): 101469. Lancet 391 (10125): 1108–20. https://doi.org/10.1016/s0140- https://doi.org/10.1016/j.pmedr.2021.101469. 6736(17)32906-9. 6. Eregata, Getachew Teshome, Alemayehu Hailu, Zelalem Adug- 12. Jha, Prabhat, Rachel Nugent, Stéphane Verguet, David Bloom, na Geletu, Solomon Tessema Memirie, Kjell Arne Johansson, and Ryan Hum. 2012. “Chronic Disease Prevention and Control.” Karin Stenberg, Melanie Y. Bertram, Amir Aman, and Ole Frith- Copenhagen Consensus Challenge 2012 Paper. Tewksbury, MA: jof Norheim. 2020. “Revision of the Ethiopian Essential Health Copenhagen Consensus Center. https://copenhagenconsen- Service Package: An Explication of the Process and Methods sus.com/sites/default/files/chronicdisease.pdf. Used.” Health Systems & Reform 6 (1): e1829313. https://doi.org 13. Jha, Prabhat, Chinthanie Ramasundarahettige, Victoria Lands- /10.1080/23288604.2020.1829313. man, Brian Rostron, Michael Thun, Robert N Anderson, Tim 7. Food and DrugAdministration (FDA). 2014. “Food Labeling: Nu- McAfee, and Richard Peto. 2013. “21st-Century Hazards of trition Labeling of Standard Menu Items inRestaurants and Simi- Smoking and Benefits of Cessation in the United States.” The lar Retail FoodEstablishments.”Washington, D.C.: Food and Drug New England Journal of Medicine 368 (4): 341–50. https://doi. COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAP T ER 15 447 org/10.1056/NEJMsa1211128. Bertram, Callum Brindley, Andreia Meshreky, James E Rosen, 14. Mcintyre, Di, Filip Meheus, and John-Arne Røttingen. 2017. et al. 2017. “Financing Transformative Health Systems towards “What Level of Domestic Government Health Expenditure Achievement of the Health Sustainable Development Goals: Should We Aspire to for Universal Health Coverage?” Health A Model for Projected Resource Needs in 67 Low-Income Economics, Policy and Law 12 (02): 125–37. https://doi. and Middle-Income Countries.” The Lancet Global Health 5 (9): org/10.1017/s1744133116000414. e875–87. https://doi.org/10.1016/s2214-109x(17)30263-2. 15. NCD Countdown 2030 Collaborators. 2020. “NCD Count- 24. United Nations (UN). 2022. WorldPopulation Prospects 2022. down 2030: Pathways to Achieving Sustainable Development https://population.un.org/wpp/ (accessed 1 September 2022). Goal Target 3.4.” The Lancet 396 (10255): 918–34. https://doi. 25. Vassall, Anna, Sedona Sweeney, Jim Kahn, Gabriela B. Gomez, org/10.1016/s0140-6736(20)31761-x. Lori Bollinger, Elliot Marseille, Ben Herzel, et al. 2017. “Reference 16. ———. 2022. “NCD Countdown 2030: Efficient Pathways Case for Estimating the Costs of Global Health Services and In- and Strategic Investments to Accelerate Progress towards the terventions.” Seattle, WA: Global Health Cost Consortium, Uni- Sustainable Development Goal Target 3.4 in Low-Income and versity of Washington. Middle-Income Countries.” The Lancet 399 (10331): 1266–78. 26. Watkins, David A, Jinyuan Qi, Yoshito Kawakatsu, Sarah J Pick- https://doi.org/10.1016/s0140-6736(21)02347-3. ersgill, Susan E Horton, and Dean T Jamison. 2020. “Resource 17. NCD Risk Factor Collaboration, Rodrigo M Carrillo-Larco, Requirements for Essential Universal Health Coverage: A Mod- Goodarz Danaei, Leanne M Riley, Christopher J Paciorek, elling Study Based on Findings from Disease Control Priorities, Gretchen A Stevens, Edward W Gregg, et al. 2021. “Worldwide 3rd Edition.” The Lancet Global Health 8 (6): e829–39. https:// Trends in Hypertension Prevalence and Progress in Treatment doi.org/10.1016/s2214-109x(20)30121-2. and Control from 1990 to 2019: A Pooled Analysis of 1201 27. Watkins, David A., Dean T. Jamison, Anne Mills, Rifat Atun, Kris- Population-Representative Studies with 104 Million Partici- ten Danforth, Amanda Glassman, Susan Horton, et al. 2017. pants.” The Lancet 398 (10304). https://doi.org/10.1016/s0140- “Universal Health Coverage and Essential Packages of Care.” 6736(21)01330-1. Edited by Dean T. Jamison, Hellen Gelband, Susan Horton, Pra- 18. Nishtar, Sania, Sauli Niinistö, Maithripala Sirisena, Tabaré bhat Jha, Ramanan Laxminarayan, Charles N. Mock, and Rachel Vázquez, Veronika Skvortsova, Adolfo Rubinstein, Festus Gon- Nugent. PubMed. Washington (DC): The International Bank tebanye Mogae, et al. 2018. “Time to Deliver: Report of the for Reconstruction and Development / The World Bank. 2017. WHO Independent High-Level Commission on NCDs.” The https://www.ncbi.nlm.nih.gov/books/NBK525285/. Lancet 392 (10143): 245–52. https://doi.org/10.1016/s0140- 28. Watkins, David A., Gavin Yamey, Marco Schäferhoff, Olusoji 6736(18)31258-3. Adeyi, George Alleyne, Ala Alwan, Seth Berkley, et al. 2018. 19. Ochalek, Jessica, James Lomas, and Karl Claxton. 2015. “Cost “Alma-Ata at 40 Years: Reflections from the Lancet Commis- per DALY Averted Thresholds for Low- and Middle-Income sion on Investing in Health.” The Lancet 392 (10156): 1434–60. Countries : Evidence from Cross Country Data.” Research Paper https://doi.org/10.1016/S0140-6736(18)32389-4. 122. York, UK: Centre for Health Economics (CHE), University of 29. Watkins, David, Sali Ahmed, Sarah Pickersgill, and Saleema York. https://doi.org/10.7490/f1000research.1113912.1. Razvi. 2023. “Best Investments in Chronic, Noncommunicable 20. Ottersen, Trygve, and Ole F Norheim. 2014. “Making Fair Disease Prevention and Control in Low- and Lower–Middle-In- Choices on the Path to Universal Health Coverage.” Bulletin come Countries.” Journal of Benefit-Cost Analysis, July, 1–17. of the World Health Organization 92 (6): 389–89. https://doi. https://doi.org/10.1017/bca.2023.25. org/10.2471/blt.14.139139. 30. Watkins, David, Jessica Hale, Brian Hutchinson, Ishu Ka- 21. Pickersgill, Sarah J., William T. Msemburi, Laura Cobb, Nicole taria, Vasilis Kontis, and Rachel Nugent. 2019. “Investing in Ide, Andrew E. Moran, Yanfang Su, Xinpeng Xu, and David A. Non-Communicable Disease Risk Factor Control among Ado- Watkins. 2022. “Modeling Global 80-80-80 Blood Pressure Tar- lescents Worldwide: A Modelling Study.” BMJ Global Health 4 gets and Cardiovascular Outcomes.” Nature Medicine 28 (8): (2): e001335. https://doi.org/10.1136/bmjgh-2018-001335. 1693–99. https://doi.org/10.1038/s41591-022-01890-4. 31. World Bank. 2022. Antiretroviral therapy coverage (% of people 22. Prabhakaran, Dorairaj, Shuchi Anand, David Watkins, Thomas living with HIV). Health Statistical Database/ https://data.world- Gaziano, Yangfeng Wu, Jean Claude Mbanya, Rachel Nugent, bank.org/indicator/SH.HIV.ARTC.ZS (accessed 22 October 2022). et al. 2018. “Cardiovascular, Respiratory, and Related Disorders: 32. World Health Organization (WHO). 2020. Global Health Es- Key Messages from Disease Control Priorities, 3rd Edition.” The timates: Life Expectancy and Leading Causes of Death and Lancet 391 (10126): 1224–36. https://doi.org/10.1016/s0140- Disability. Global Health Observatory. Geneva: World Health 6736(17)32471-6. Organization. https://www.who.int/data/gho/data/themes/ 23. Stenberg, Karin, Odd Hanssen, Tessa Tan-Torres Edejer, Melanie mortality-and-global-health-estimates. 448 CHAP T ER 15 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E ANNEX 15.1 METHODS Intervention selection and aggregation ed in our analysis, which was focused on the next 25 years. Our analysis looks at both clinical and in- The starting point of this analysis is a set of interven- tersectoral interventions through a benefit-cost lens. tions recommended in the third edition of the Dis- ease Control Priorities (DCP3) series (Jamison et al. Modeling intervention costs 2018). DCP3 was a set of nine volumes containing 172 chapters that covered all major areas of global health Our cost estimates build on those done for a previ- interest. The volumes were published over 2015- ous publication from the Disease Control Priorities 2018, with the final volume being a synthesis of the Project (Watkins et al. 2020) and the aforemen- previous eight volumes. DCP3 chapters underwent tioned NCD Countdown report (NCD Countdown peer review in a process overseen by the US National Collaborators 2022). Costs borne by governments Academy of Medicine. Each chapter covered a partic- in implementing the intersectoral policies were ular health topic (e.g., tuberculosis, cancer screening, estimated on a per-capita basis, using published neurological disorders) and synthesized the evidence costing studies or grey literature (e.g., government in a series of recommended interventions that (i) pro- budget reports). For the clinical interventions, the vide good value for money, (ii) are feasible to imple- focus was on unit costs (e.g., cost per patient-year of ment in low- and middle-income countries, and (iii) chronic treatment, cost per episode for acute care, address a significant cause of death or disability. etc.) to healthcare sector. All interventions were as- These criteria were applied to systematic reviews sumed to be publicly financed (i.e., through univer- of economic evaluations of health interventions done sal health coverage systems), so out-of-pocket costs in low- and middle-income country settings, supple- currently paid by households would be shifted to mented by other information such as clinical and im- governments and accounted for in our estimates. plementation studies and expert judgment. The latter We primarily sourced unit cost data for the clin- was especially important, because robust cost-effec- ical interventions from DCP3’s systematic reviews of tiveness information is lacking for many services that cost and cost-effectiveness studies (see, e.g., a review are being provided, and topical expertise is needed to of cardiovascular disease treatment costs by Brouw- be able to make sense of the literature. For example, er et al. 2015). Since NCD costing studies are few, we DCP3 did not identify any published studies of the selected the highest-quality study that we identified cost-effectiveness of cardiopulmonary resuscitation, that most closely reflected the medical components but this intervention was recommended because it is of the intervention in question. All costs were up- an essential part of the emergency healthcare system dated to 2020 US dollars using procedures recom- and is nearly impossible to study using convention- mended by the Global Health Costing Consortium al economic evaluation techniques. In other words, (Vassall et al. 2017). They were then extrapolated to DCP3 strove to avoid both “false negatives” and other countries in two stages. First, we decomposed “false positives” (Jamison et al. 2018). costs into traded and nontraded components. Trad- DCP3’s final list of recommended interventions ed components were assumed to be constant across was separated into 218 health sector interventions countries. Nontraded components were adjusted and 71 intersectoral interventions. The entire list of based on ratios of gross national income (GNI) per interventions is available in a series of appendices ac- capita across countries. As described in the DCP3 companying the DCP3 capstone paper in the Lancet costing paper and its appendix (Watkins et al. (Jamison et al. 2018). For this analysis, we selected 2020),this approach has several limitations but rep- 30 interventions that are proven to reduce mortali- resents the most feasible approach for a global-level ty from NCDs and can achieve meaningful impacts modeling study. Our costs were meant to be illus- by 2050. HPV vaccination for adolescent girls, for trative of the magnitude and range of costs across example, is highly cost-effective but has at least a 30- LICs and MICs rather than precise estimates used year lag between administration and attainment of for budgeting country-level NCD programs. Coun- any meaningful health benefits, so it was not includ- try-level empirical (micro)costing of the 30 inter- COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAP T ER 15 449 ventions was outside the scope of this chapter. inform our approach, which used an offset parame- Unit costs were then multiplied by the popula- ter that was applied to the estimated economic ben- tion requiring each intervention and further by the efits from improved health (see below). For tobacco target coverage level of the intervention each year. For and alcohol policies, the offset value was 0.9, and for example, the cost of an intervention costing $20 per sodium and trans-fat policies, it was 0.5. patient-year that addressed a chronic disease with a prevalence of 1 million cases and a current coverage of Modeling intervention health 30 percent was calculated as $20 * 1,000,000 * 30 per- and economic outcomes cent = $6,000,000. The “incremental” cost of increas- ing coverage of that intervention by a certain amount We quantify improvements in health as a reduction would be calculated as the difference in coverage year in mortality and disability rates, and therefore total over year. We defined full coverage of each interven- deaths and disability-adjusted life-years (DALYs), fol- tion as 80 percent of the population covered by the lowing the scale-up of an intervention. To do this, we year 2050, consistent with DCP3 and WHO assump- used a population model we developed for the NCD tions (Stenberg et al. 2017; Jamison et al. 2018). Epi- Countdown report (NCD Countdown Collaborators demiological and demographic data used to estimate 2022). In brief, this model combined demographic population in need were taken from the WHO (WHO projections (UN 2022) – (including population counts 2020), UN Population Division (UN 2022), and Glob- and all-cause mortality rates) with cause-of-death al Burden of Disease 2019 Study (GBD 2019 Diseases data (WHO 2020)28 and disease incidence and preva- and Injuries Collaborators 2020). Coverage data were lence rates (GBD 2019 Diseases and Injuries Collabo- taken from the literature, WHO, or expert opinion. rators 2020). The baseline projection that we used as a As Table 15.2 shows, different countries currently reference for calculating intervention-specific health have different levels of coverage for the various NCD gains was calibrated to the UN Population Division interventions, so the incremental cost and impact (be- medium projections, representing a business-as-usual low) of achieving 80 percent target coverage will be scenario for intervention implementation. different. We imposed a standard 2 percent per year Changes in disease-specific mortality and dis- increase in coverage for all interventions and coun- ability rates were a function of (i) the effectiveness tries in our model, meaning that they all increase of the intervention on these outcomes, usually ex- coverage at the same rate but reach 80 percent at dif- pressed as a rate ratio or hazard ratio, and (ii) the ferent time points (and some do not reach 80 percent change in intervention coverage. Effectiveness data by 2050, especially LICs). The assumption of a 2 per- were usually taken from clinical trials, favoring me- cent annual increase is based on recent trends in hy- ta-analytic estimates when available. We did our own pertension coverage (1.0-1.5 percent annual increases searches for effectiveness data for the NCD Count- in well-performing countries (NCD Risk Factor Col- down report (see appendix). Intervention-specific laboration 2021)) and in antiretroviral drug therapy effectiveness parameters are detailed in the online coverage (4-5 percent annual increases (World Bank appendix (see Github URL below). We multiplied 2022)). Hence a 2 percent increase reflects an accelera- each literature-based effect size by 0.70 to account tion of progress on access to NCD care but within rea- for imperfect implementation in real-world settings son, considering that NCD programs do not have the (NCD Countdown Collaborators 2022). same resources or political momentum as HIV/AIDS. To calculate the economic value of reduced mor- For the non-health sector interventions, there tality and disability, we multiplied projected DALYs are two major types of costs that are borne outside by the standardized time series estimates for the the government/healthcare sector. The first type is value of a DALY that were used throughout the Co- the cost to firms of implementing government regu- penhagen Consensus project. One potential benefit lations. Again, we used literature-based estimates of of tobacco and alcohol taxes is a gain in revenue to these costs and extrapolated them across countries, governments. We took a societal perspective on costs like we did for the clinical interventions (above). and benefits, so these revenue gains are fully offset by The second type of cost is the forgone consumer additional costs to consumers—i.e., they are, func- surplus due to taxes and regulations on unhealthy tionally, transfer payments. products. As outlined in our paper for the Copen- All input data, including citations of the litera- hagen Consensus Center (Watkins, Ahmed, and ture used to estimate the cost of each intervention, Pickersgill 2023), we used recommendations from are available at https://github.com/Disease-Con- US-based regulatory impact analyses (FDA 2014) to trol-Priorities/CCC. 16 450 CHAPTER 16 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E Control of Non-Communicable Diseases for Enhanced Human Capital The case for whole-of-society action Ramesh Govindaraj a and Sundararajan Srinivasa Gopalan b a Health, Nutrition, and Population, Africa Region, World Bank b Independent Researcher COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 16 451 INTRODUCTION This chapter has been prepared as background to the Healthy Longevity Initiative (HLI), under the ambit of the Health, Nutrition, and Population (HNP) Global Practice of the World Bank Group (WBG). Under this initiative, a series of papers were commissioned on various aspects of healthy longevity, most of them focusing on different facets of non-communicable diseases (NCDs), such as the current status of the problem, its economic burden, the financing of interventions, fiscal instruments for controlling NCDs, behavioral aspects, relationship with the labor market, long-term care needs, and the gender dimension. Although of direct relevance, these aspects are dealt with only briefly in this chapter as these are covered in other chapters of this compendium. This chapter focuses specifically on a “whole-of-society” (WoS) approach to preventing and mitigating NCDs. Its main objectives are: through infection,” including other endocrine disor- ders (beyond diabetes), malnutrition (both under- i. Briefly review the current status of NCDs, and over-nutrition), auto-immune diseases, chronic their economic costs, and their impact on hu- kidney disease, various occupational hazards, and man capital; injuries – both accidental and intentional. The ra- tionale for broadening the scope is that it brings the ii. Argue for WoS efforts to address NCDs, in relationship between NCDs and non-health sectors view of the two-way relationship of NCDs with into sharper focus and further strengthens the case relevant economic sectors; and for WoS action against NCDs. As this chapter was prepared against the back- iii. Make recommendations on the way forward on drop of the COVID-19 pandemic, it includes a WoS efforts. section on the two-way links between NCDs and COVID-19, in order to glean lessons on the value of Scope and limitations a WoS approach to NCDs in the post-COVID-19 era. The chapter is primarily based on desk reviews, The influential 4x4 framework on NCDs proposed online research, and informal interviews with tech- by the World Health Organization (WHO) focused nical experts at the WBG and other development on four sets of diseases, i.e., cardiovascular diseas- entities, with some inputs from internal and external es (CVD), diabetes mellitus, chronic respiratory reviewers. The main sources include WHO docu- diseases, and cancers, along with four behavioral ments, WBG documents, technical journals, and risk factors (tobacco use, harmful use of alcohol, the World Bank’s operational database. Examples unhealthy diets, and physical inactivity). WHO and country case studies presented in the paper are subsequently expanded this to a 5x5 framework, meant only to be illustrative and not comprehensive. including mental health issues among the set of dis- Since this work is commissioned by the WBG, eases and air pollution as a key risk factor. This lim- the review of global experience on a WoS approach ited focus helped in advocacy but was not intended to NCDs includes a special focus on WBG opera- to neglect or downplay the importance of a host of tions. A self-examination of how far the approach other NCDs. In this chapter, a broader, more com- has been adopted by the institution is germane to plete definition of the term “NCDs” is considered, the paper’s scope and message. i.e., “any disease/disability that is not transmissible WHY FOCUS ON NCDs? NCDs are making an increasingly significant contribution to the burden of disease globally, including in the developing world. Key NCD risk factors are widely prevalent across the world and are responsible for a large share of the current total of disability-adjusted life years (DALYs) lost. NCDs affect all age groups and both sexes (encompassing the entire range of gender identities) but have a disproportionate impact on older populations. By virtue of their effects on human capital, NCDs are increasingly among the important constraints to economic 452 CHAPTER 16 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E growth. NCDs have a differential impact on men and women. Assessing the economic impact/financial burden on societies or individuals of NCDs—either individually or collectively—is a challenging task, as is the computa- tion of NCD expenditures. But it is clear that the economic impact of NCDs—both due to lost productivity and lost lives—is large and growing. NCDs also compromise future economic and human development because of the inter-generational impacts of poverty and ill-health. NCDs are making an increasingly significant con- 2019 or COVID-19). However, 78 percent of NCD tribution to the burden of disease globally, includ- deaths occur in low- and middle-income coun- ing in the developing world (Figure 16.1). Accord- tries (LMICs) – (WHO 2021). While low-income ing to the latest available Global Burden of Disease countries continue to wrestle with older and new- (GBD) report, 42 million deaths (74 percent of er pathogens, they are also seeing a rise in NCDs, global mortality) and the loss of 1.62 billion dis- thus having to manage a double burden of NCDs ability-adjusted life years (DALYs), are attributable and CDs. Notwithstanding the severe impact to NCDs, 63.82 percent of total DALYs (IHME of the ongoing COVID-19 pandemic on human 2019; The Lancet 2020). This is not counting 4.2 lives and economies, the disease burden caused by million deaths and 249 million DALYs lost to inju- NCDs continues to be equally—if not more—con- ries – which should be included in the definition of sequential for human capital. In fact, the effects of NCDs. NCDs constitute the bulk of the disease bur- COVID-19 are often compounded and worsened by den in high-income countries (HICs), which have co-morbidities, most of which are NCDs. Therefore, largely controlled communicable diseases (CDs), even as health systems are still grappling with the with the notable exception of emerging infections pandemic, it is timely to consider the multifaceted (the most recent of which is the coronavirus disease challenges and solutions resulting from NCDs. FIGURE 16.1 Trend in global deaths and disease burden by major cause (annual rate, 1990 - 2019) Source: IHME Global Burden of Disease (2019). [Accessed: July 15, 2022] Note: The rate of mortality is defined as the number of deaths due to cause divided by the mid-year population. COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 16 453 Figure 16.2 shows the increase in mortality rates million deaths are attributed to tobacco products and lost DALYs due to NCDs in LMICs from 1990 to and smoking, and roughly 7 million deaths are 2019. Conversely, during the same period, the shares linked to road traffic accidents. 22 percent of adults of mortality and lost DALYs attributable to communi- are estimated to be hypertensive; 650 million peo- cable diseases and maternal causes declined steadily. ple are estimated to be obese; and the prevalence of Key risk factors for NCDs are highly prevalent diabetes has nearly quadrupled in the past 40 years. globally (WHO 2021). Alcohol abuse contributes Air pollution ranks fourth among the risk factors to 11 percent of the global death toll; 28 percent of contributing to deaths – both among males and fe- adults have a sedentary work/lifestyle; salt intake is males. Globally, 6.5 million annual deaths are esti- extremely high, at an average of 9-12 grams per day; mated to be due to air pollution. Figure 16.3 shows each year, more than 180,000 deaths are attributed that most of the top 10 risk factors causing deaths to high sugar consumption globally; more than 7 are related to NCDs. FIGURE 16.2 Trend in deaths and disease burden by major cause in LMICs (annual rate, 1990 - 2019) Total Deaths - LMICs Total DALYs - LMICs Source: IHME Global Burden of Disease (2019). [Accessed: July 15, 2022] Note: The rate of mortality is defined as the number of deaths due to cause divided by the mid-year population. 454 CHAPTER 16 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E FIGURE 16.3 Global Deaths Attributed to Major Risk Factors, by sex, 2019 Source: Adapted from Murray et al. 2020 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 16 455 FIGURE 16.4 Global disease burden from NCDs by age, 1990 to 2019 Source: Adapted from IHME/Our World in Data 2020 [Accessed: July 15, 2022] Note: Disease burden is measured in disability-adjusted life years (DALYs) measured as the sum of years of life lost and years lived with disability. NCDs affect all age groups and both sexes but Gender impacts have a disproportionate impact on older popula- tions (Figure 16.4). NCDs have a differential impact on men and wom- By virtue of their effects on human capital, en. In addition to certain cancers (e.g., breast, cer- NCDs are an important constraint to economic vix, ovarian, and prostate) that are gender-specific, growth. As the world’s population has aged at dra- the exposure to risk factors for many other NCDs matic speed, governments have been increasingly has been found to differ by gender. According to concerned about how to maximize the contribution research by the Pan-American Health Organiza- of the older population to all aspects of human capi- tion (PAHO), “Women are significantly more like- tal. This is particularly important, as the proportion ly to be obese than men. PAHO’s 18-country study of older adults has risen in virtually all countries, found that 40 percent of women in Canada and over and especially in developing countries. According- 70 percent of women in Nicaragua and Belize are ly, governments are now engaging in policy inter- overweight or obese. Women’s higher rates of obe- ventions to make the additional years healthy and sity lead to their increased vulnerability to NCDs, optimally productive, in order “to develop and particularly diabetes” (PAHO 2011). Worldwide, maintain the functional ability that enables well-be- 48 percent of adult men smoke compared with 12 ing through middle and older ages” (WHO 2020a). percent of women (WHO 2010), thus increasing One significant aspect of longevity is that the pro- men’s risk of lung cancer. Women experience less ductivity of older, knowledgeable, healthy adults apparent and often different symptoms of CVD enhances and increases the pool of human capital, than do men and, consequently, are less likely to be thereby contributing to economic growth. Though diagnosed and treated appropriately (DeVon et al. NCDs pose a powerful threat to countries’ human 2008). Apart from the differential risks and prev- capital, productivity, and future economic growth, alence rates of NCDs, gender disparities in terms they have received relatively little attention in a con- of financial security, employment, household de- certed intersectoral manner. cision-making power, and socio-cultural factors which determine gender roles and health care ac- cess are likely to be critical factors that could impact 456 CHAPTER 16 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E women and men differently. PAHO (2011) recom- (curative as well as palliative). The cost of treatment mends three approaches to address NCDs in a gen- and the continuum of care is more difficult to es- der-sensitive manner: (i) pro-poor and gender-sen- timate, due to challenges in attribution of medi- sitive health policies; (ii) use of primary health care cal care costs to diseases, and in tracking life-long for NCD prevention and control; and (iii) invest- costs. It is also not clear whether the study includ- ment in girls, women, and their health. Chapters 8 ed recurrent expenditures or only the capital costs.  - 11 of this compendium delve deeper into gender If recurrent costs are not included (as “programs” issues relevant to NCDs. are usually funded from the investment budget and typically don’t count routine recurrent costs such as Economic aspects salaries and consumables), that would be another source of underestimation. Notwithstanding these Assessing the economic impact and financial bur- limitations, the effort invested in the Chilean study den of NCDs on societies or individuals is a chal- is laudable, and it is worth replicating/expanding in lenging task, whether the diseases are considered other countries around the world. singly or collectively. The computation of NCD It is clear that the economic impact of NCDs— expenditures is also difficult. This is due to issues both due to lost productivity and lost lives—is large with definition (e.g., which diseases and risk factors and growing. Childhood malnutrition can adverse- to include), estimating the prevalence of specific ly impact the learning and earning potential of af- NCDs/risk factors, and attributing expenditures fected children and would have a negative effect on specifically to prevention and management. The human capital in the medium- to long-term. Labor problem is compounded when expenditures are in- units lost to NCD deaths and the direct medical curred, and the impact is dispersed, across econom- costs of treating NCDs can be expected to reduce the ic sectors. Most costs/expenditures and financial/ quality and quantity of the labor force and human economic NCD burden calculations are, therefore, capital. Economists increasingly express concern estimates derived from various modeling exercises. that NCDs will result in long-term macroeconomic Chapters 3 and 4 discuss the economic aspects of impacts on labor supply, capital accumulation, and NCDs, and therefore this paper does not present a GDP worldwide with the consequences being most comprehensive review of this subject. severe in developing countries. A study on the mac- The Chile Expenditure Study. One exception roeconomic burden of NCDs in the United States to the typical estimations that is worth highlighting, estimated that the economic burden of all NCDs in however, is a unique report on Chile by the Ministry 2015-2020 is US$265,000 per capita, and that the of Health, PAHO, and the Economic Commission total NCD burden roughly corresponds to an an- for Latin America and the Caribbean, which pro- nual tax rate of 10.8 percent on aggregate income vides some estimates of expenditures on NCDs by (Chen et al. 2018). A recent study on the association selected ministries, highlighting clearly the multi- between NCD morbidity and employment status in sectoral nature of such expenses. It shows that Chile’s the Republic of Korea found that chronic disease Ministry of Social Development reported spending increased the risk of unemployment markedly, with 0.7 percent, the Ministry of Sports 22.8 percent, the CVD and cancer having the most profound nega- Ministry of Interior and Public Safety 0.1 percent, tive effect (Kwon et al. 2020). the Ministry of Environment 11.8 percent, the Min- NCD risk factors such as tobacco smoking, istry of National Assets 0.1 percent, and the Minis- obesity, and heavy drinking have negative econom- try of Health 3.5 percent of their respective budgets ic impacts, including through effects on adult earn- in 2013 on NCDs. In absolute terms, the Ministry of ings. The capacity of health care systems to manage Health spent the most on NCDs, as could be expect- the consequences of increased risk factors, and the ed. All told, the government spent 1.4 percent of its diseases associated with them, ultimately deter- budget on NCDs (Cuadrado et al. 2015). mines changes in morbidity and mortality. Ac- The study has limitations, but the approach counting for the effects of health care systems would holds considerable promise. It focuses only on pri- require including health outcome measures–such as mary prevention activities, including health promo- the child stunting and adult survival rates used in tion. This grossly underestimates total expenditures the human capital index (HCI)–as additional prox- on NCDs, as the largest share of expenditures would ies for latent health status (World Bank 2020b). As be on care (including long-term care) and treatment an example of the economic impact of NCDs, avail- COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 16 457 able evidence indicates that the estimated cost of tional impacts of poverty and ill-health. For exam- new cancer cases in 2009 was US$286 billion glob- ple, poor nutrition may not only lead to NCD-relat- ally (Bloom et al. 2011). This estimate is based on ed morbidity in the affected individuals, but it could the cost-of-illness approach and includes treatment also impair in-utero growth and compromise fetal and care costs, research and development costs as- development, thus harming the next generation.  sociated with cancer control, and foregone income Taking a life-course approach to the issues of pov- due to inability to work. erty and NCDs that acknowledges the connections NCDs also compromise future economic and between social and health conditions over the lifes- human development because of the inter-genera- pan is, therefore, useful. IMPACT OF THE COVID-19 PANDEMIC ON NCDs The COVID-19 pandemic has thrown into sharp relief the intricate relationship between COVID-19 and NCDs. COVID-19 has significant bi-directional links to NCDs. Many of the interventions required for fighting the pandemic are also those required to fight NCDs. During the pandemic, many countries have seen the value of stronger tobacco and alcohol controls. The COVID-19 pandemic, often compounded by NCDs, has clearly had a significant negative impact on human capital around the world through disruptions to education and the loss of jobs and lives. The pandemic also caused major disruption to NCD control programs and services, highlighting the critical intersection of health systems resilience and health emergency preparedness and response. COVID-19 has significant bi-directional links to general immunity, while at the same time contrib- NCDs. On the one hand, co-morbidities, most of uting significantly to NCD prevention and control. which are NCDs, and risk factors such as obesity Many of the interventions and conditions required and hypertension increase the severity of illness for fighting the pandemic are also those required and mortality risks of COVID-19; on the other, pro- to fight NCDs (The Lancet 2020): disease surveil- longed economy-wide restrictions imposed by most lance, a strong civil society, robust public health, countries to contain the spread of COVID-19 have clear communication, and equitable access to resil- resulted in reduced access to critical care for NCD ient universal health care systems. COVID-19 could patients, thus accentuating the associated morbidity provide new insights into interactions between the and mortality. Lockdowns due to COVID-19 out- immune system and NCDs, and potentially change breaks might worsen the problem of diabetes mel- the way we understand and treat these diseases. On litus by curtailing access to medication and other the other hand, it might also generate new long- care, and in some cases reducing physical activity. term disabilities that will add to the NCD burden. Patients with diabetes have had at least twice the Based on a recent WHO survey of 107 coun- risk of severe COVID-19 (and death) compared to tries, three-quarters of countries reported a con- non-diabetic individuals (Chan et al. 2021). Most siderable degree of disruption of NCD services due importantly, COVID-19, by the sheer magnitude to COVID-19 (WHO 2020a). Two-thirds of coun- of its societal impact, touching almost every sector tries reported that ensuring the continuity of NCD of the economy, has compelled humanity to mount services was included in the list of essential health whole-of-society responses (albeit with varying de- services in their national COVID-19 response plan. grees of effectiveness), much as the Human Immu- Twenty-eight countries reported that there was ad- nodeficiency Virus/Acquired Immune Deficiency ditional funding allocated for NCDs in the govern- Syndrome (HIV/AIDS) did in most of Sub-Saharan ment budget for COVID-19 response. The pandemic Africa in 1990s. This experience could be very valu- has exposed the consequences of failing to address able, as it offers a platform to extend the same WoS several of the risk factors predicting COVID-19 in- approach beyond the pandemic. cidence and outcomes. In England and Wales, the During the COVID-19 pandemic, many coun- excess risk of dying from COVID-19 was 2.4 times tries have seen the value of stronger tobacco and al- higher for females and 1.9 times higher for males cohol controls, an important step towards reducing with longstanding health problems (ONS 2020). NCDs. There have also been national efforts to safe- Obesity alone posed a 1.9 times higher risk of dying guard and improve nutrition as a means to bolster from COVID-19, after adjusting for all other comor- 458 CHAPTER 16 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E bidities (Williamson et al. 2020). A large prospective individuals from their loved ones have exposed and, South African study showed that higher physical ac- in some cases, worsened trends in mental health and tivity levels predicted substantially lower severity and substance abuse (Ettman et al. 2022). For two years, mortality from COVID-19 (Steenkamp et al. 2022). screening programs for cancers, primary prevention The pandemic has amplified the importance programs for tobacco and alcohol, and health pro- of recognizing gaps in how we care for those living motion efforts supporting active lifestyles ceased or with multiple chronic diseases (Hajat and Kishore were deferred in many settings. 2018), with studies showing the presence of multiple The COVID-19 pandemic has clearly had a sig- co-morbidities in the majority of COVID-19 deaths nificant negative impact on human capital around (Agarwal et al. 2022). The neglect of the health ser- the world, through disruptions to education and vices for older adults has manifested in declining life the loss of jobs and lives. These disruptions have expectancy in developed countries such as the Unit- been compounded by the adverse effect of the pan- ed Kingdom (UK) and the United States of America demic on NCD management. The double impact of (USA), together with a greater number of years spent COVID-19 and NCDs on human capital around the in poor health (Marmot et al. 2022). The impact of world is therefore a matter of serious concern. the pandemic on families and the related isolation of RATIONALE FOR A WHOLE-OF-SOCIETY APPROACH TO NCDs Most health problems have a multifactorial causation and consequently warrant a multidimensional approach, but this is even truer for NCDs. This section clarifies, conceptually, what is meant by “society-wide action” or a “whole-of-society (WoS) approach,” as compared to other terms such as “multisector,” “intersectoral,” and “whole-of-government (WoG).” WoS is a comprehensive concept going beyond the WoG approach to include non-state actors (for-profit and non-profit), communities, households, and individuals. The need to involve all branches of government (the judiciary, the legislative, and the executive) and political leaders from opposition parties is emphasized, as the latter play a critical role in supporting or opposing any government initiatives. A framework for the bidirectional relationship of cause and effect linking NCDs with various sectors is presented. In practice, the goal of broad inclusiveness implied in WoS approaches needs to be balanced with the require- ment of shaping an effective alliance for change. This may involve some selectivity, especially in the early stages of a change process, in order to achieve results. Most health problems have a multifactorial causation co-morbidities are at a much greater risk of hospi- and consequently warrant a multidimensional ap- talization and death, warranting an immediate focus proach. This is perhaps even truer for NCDs, as so- on MCC, which have not received sufficient atten- cial, economic, environmental, gender, demograph- tion to date (Yach 2022). However, while arguing ic, and other dimensions contribute significantly to for non-health sectors’ actions, the health sector’s the risk of suffering from NCDs and to the eventual particular responsibility to address NCDs cannot be outcomes faced by those afflicted. Therefore, medi- ignored. For instance, while non-health sectors are cal/health interventions, by themselves, are far from critically important in the prevention of NCDs, the sufficient to address NCDs and their negative im- treatment and long-term care of these diseases and pacts on human capital and on the quality of life; to ensuring continuity of care are squarely within the be effective, they need to be complemented by social, purview of the health sector. legal, educational, environmental, and other actions. Before addressing the main substantive top- In addition to the causation of NCDs, their impacts, ics of this paper, it is helpful to clarify conceptu- too, straddle several sectors and, in fact, the econo- ally what is meant by “society-wide action” or a my as a whole. The multidimensional nature of the “whole-of-society approach.” This term has been problem and its impact, and the resulting need for consciously chosen, in preference to others such as WoS solutions, are at the core of this paper. In par- “multisectoral,” “intersectoral,” “interdisciplinary,” ticular, multiple chronic conditions (MCC) are often and “whole-of-government,” which have been used prevalent in the same individual, and the COVID-19 with varying connotations—often without sufficient pandemic showed that persons with two or more clarity and often interchangeably—by proponents COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 16 459 of the idea that many health problems cannot be at least two problems with this limited approach: (a) effectively addressed by the health sector or minis- it ignores the important roles to be played by the tries of health alone. Every one of these terms has a legislative branch and the judiciary in NCD control; legitimate place in our lexicon, but with significant and (b) it ignores the opposition political parties, differences in meaning. which may currently be out of power but would be important stakeholders if they take the reins of pow- • Multisectorality entails several sectors of the er and may even be influential while sitting in the economy—usually represented by various min- opposition benches of the country’s legislature, par- istries of government—working together to- liament, or congress. Especially in countries where wards a common goal. Such cooperation across the legislative branch is controlled by one political sectoral ministries is not easy to accomplish; party, while the executive branch is held by anoth- most governments are organized into bureau- er, it is easy to see that the involvement of just one cratic silos, making working across sectors par- branch would be inadequate. The judiciary, ignored ticularly challenging. As a result, special efforts in most initiatives, could play a vital role as well; are needed to overcome these hurdles and ob- in many instances, “judicial activism” could be the tain multisector involvement. main force behind important legal measures – for instance, when political imperatives have stymied • Intersectoral is a term that goes beyond multisec- action by the legislative branch. For example, in torality, in that not only are multiple sectors in- some countries the courts have taken it upon them- volved, but they are also actively interacting with selves to pull up the executive branch for neglecting each other, coordinating effectively, and cooper- the environment. Taking suo moto cognizance of ating towards a common goal (NCD prevention issues of public interest, courts can and have tak- and control, in our case). If multisector involve- en such actions and will likely continue to do so. ment is challenging, intersectoral coordination In other instances, groups seeking to weaken public is even more difficult. But without such coop- health measures have also used the courts (World eration and coordination, interventions against Bank 2022). The most recent example of public NCDs are unlikely to be optimally effective. health interventions being undermined if political forces and opinion-makers fail to align with public • Whole-of-government approaches (sometimes policy manifested as political hurdles placed in the referred to as “Health in all”) connote broad- way of measures to control the SARS-CoV-2 virus. er engagement than only the various sectoral Public health strategies including lockdowns and ministries. They include an aspect that is often mask mandates in public places, and even vacci- overlooked, i.e., the political economy of any nation, have been targeted by lawsuits and political serious initiative. Since many preventive mea- prohibitions in some jurisdictions. sures against NCDs are behavioral, they need WoS concepts go beyond whole-of-govern- broad-based acceptance from the population, ment efforts, as they must involve non-state actors and therefore warrant a concerted effort across as well. Many of the causal determinants of NCDs the political spectrum, i.e., non-partisan buy-in. require the active participation of the private sec- “Whole-of-government” would also mean the in- tor, industry, and civil society organizations such clusion of all branches of government, not just the as non-governmental and community-based orga- executive branch, to which “multisectoral” or “in- nizations. Strong efforts to develop public-private tersectoral” approaches tend to limit themselves. partnerships and community mobilization are cru- cial to the success of NCD control programs. It is worth elaborating on the concept of involv- The scope of a WoS approach is depicted in ing all branches of government. Typically, in every Figure 16.5, which shows a Venn diagram high- government, economic sectors are represented by lighting the all-encompassing nature of the term different line ministries, and efforts to involve the “whole-of-society.” This term comprises all parts of “whole government” seek to involve all these minis- the government with all its branches, including po- tries; but this approach considers only the executive litical leaders who might not currently hold power; branch of the government and can hardly be called different types of non-state actors, both commercial “whole-of-government” (let alone “whole-of-soci- and non-profit; civil society organizations, opinion ety,” as explained in the next paragraph). There are leaders, and community organizations; and, last but 460 CHAPTER 16 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E not least, the individuals and households that con- ance for change. The purpose of such an alliance stitute the society. In short, what distinguishes the of key stakeholders is to work together to achieve term “whole-of-society” from other designations shared objectives. Other stakeholders, including like multisector, intersectoral, and whole-of-gov- those opposed to the objectives, ultimately need to ernment is its comprehensive scope and inclusive be consulted and their views taken into account. nature. In practice, the goal of broad inclusiveness However, some selectivity in participation, particu- implied in WoS approaches needs to be balanced larly at early stages of the change process, is likely to with the requirement of shaping an effective alli- be required for a WoS strategy to succeed. FIGURE 16.5 Components of a “whole-of-society” approach Source: Original figure for this publication It is important to stress that policy action on icies. Indeed, as elaborated below, so-called “judicial NCDs is the culmination of a constant negotia- activism” has played a pivotal role in promoting ac- tion amongst the executive, legislative, and judicial tion on NCDs and their risk factors in many coun- branches of government and non-state actors. In tries, when the legislative or the executive branch has other words, the process is profoundly political, and been slow to act or has faced resistance from political institutional structures and processes are key to how vested interests. By the same token, broad agreement policies are elaborated and implemented. As noted, across the political spectrum is important to build a it would be a mistake to ignore the important roles consensus on appropriate NCD-related policies and of legislative and judicial processes in policy making, ensure continuity in the implementation of policies as well as in the implementation of policies and pro- and programs across successive administrations. grams. These roles are at least as important as that of Similarly, the role of sub-national actors, particular- the executive in determining the course of NCD pol- ly in federal systems, has not been given sufficient COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 16 461 importance in the literature on WoS approaches to ples include court rulings on the mask mandate for NCD prevention and mitigation. public transportation advocated by the United States There are many reasons for the expanding role Centers for Disease Control and Prevention and the of the judiciary in national health policies. These “vaccine-or-test” mandate from the United States include the inability of the legislative and executive Occupational Safety and Health Administration. branches of government to develop explicit policies This chapter alludes to the many benefits of es- in some settings, the growing share of national re- tablishing a national political consensus on social sources devoted to health care, and the increasingly and health policies. However, the public response litigious nature of society globally (Anderson 1992). to the COVID-19 pandemic in the United States Proponents of judicial intervention have suggest- and Brazil, for example, has been highly politicized. ed a number of advantages of the courts’ becom- More generally, although it is well known that party ing involved in social policy issues: the promotion identification can shape basic political perception, of minority rights, the promotion of more humane there have been increasing concerns about the po- conditions in social institutions, restrictions on bu- liticization of science globally. This includes parti- reaucratic arbitrariness, and, more generally, the san divisions around the world over public health promotion of positive social change (Kagan 1991). issues such as vaccination, genetically modified or- On the other hand, it has also been suggested that ganisms, and climate change (Kerr, Panagopoulos, judges may not always have the technical back- and van der Linden 2021). Responses to NCDs and ground or experience necessary to evaluate complex their risk factors are no exception. Cases in point social policy issues; that the decentralized nature of include the polarization of views in various coun- the judicial system can lead to inconsistent treat- tries on the use of alcohol and sugar-sweetened bev- ment of similar cases; that the focus of the courts’ erages, gun control, and other issues. review is often narrow (Anderson 1992); and that, The polarization of views observed at the na- more generally, “excessive reliance on the courts, tional level also plays out at the global level. De- instead of self-government through democratic pro- veloped countries that export alcoholic beverages cess, may deaden a people’s sense of moral and polit- often complain about the perceived “array of new ical responsibility for their own future” (Cox 1976). and existing tariff and non-tariff barriers in export These arguments notwithstanding, there are markets” in the developing world. Health protec- several instances of courts in developed and de- tion and labelling, particularly for food and drink, veloping countries intervening in lieu of govern- are emerging as a dominant theme in many of ments—based on human rights and interpretation the “specific trade concerns” that countries raise in of the constitution—to promote or impede public international fora, highlighting the balance govern- health interventions. For example, responding to ments must strike between trade and health in areas public interest litigation on air pollution, the judi- such as reducing obesity, discouraging unhealthy ciary in India has asked governments in many cases eating and alcohol abuse, and protecting children, to constitute national and state regulatory boards for example, by regulation or by helping consumers or environmental courts, or issued directions to re- to make better informed choices for themselves; we mind statutory authorities of their responsibility to would contend that this choice between trade and protect the environment. Similarly, expressing se- health is a false one, and health must not be traded rious concern over the large number of deaths due for commerce in unhealthy goods. to road traffic accidents in the country, the Supreme On the importance of sub-national actors in Court of India in 2001 made wearing seat belts man- how policies and laws relevant to NCDs are inter- datory for front seat occupants in cars and directed preted and implemented in countries with federal the chief secretaries of the states and union territo- systems, the United States provides a striking ex- ries in India to implement the order. However, it was ample. The United States Constitution gives states only in 2005 that the Union government passed a broad authority on a variety of issues, including law that made front seat belt use in cars mandatory, health regulations. For example, Section 2 of the while the draft rules for the mandatory use of rear Twenty-first Amendment to the Constitution has seat belts were developed as recently as September been interpreted by the courts and others as giv- 2022. In contrast, courts have sometimes limited the ing broad authority over the regulation of alcoholic government’s ability to use emergency powers to beverages to the states and limiting the power of the promote public health and protect people. Exam� - national government to intrude upon state alcohol- 462 CHAPTER 16 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E ic beverage control policies. States, in turn, can and private sector is also critical, particularly in LMICs in many cases have delegated authority to counties where the private sector is often the primary pro- and localities. As a result, the availability of alcohol- vider of health and health services. The section on ic beverages, their prices, and the terms and con- engaging non-state actors in this chapter deals with ditions under which they can be obtained (for ex- the engagement with non-state actors, including ample, whether a county is “dry,” or whether a state for-profit and non-profit entities, in some detail. itself exercises a monopoly on the sale of wines and Overall, the case for a WoS approach has been spirits) have varied substantially across the country. made by many influential authors, a sample of As part of the WoS approach, the role of the which is provided in Box 16.1 below. BOX 16.1 Rationale for WoS approach – a sample of arguments 1. The argument for the involvement of non-state actors is strongly advanced by Galambos and Sturchio in their book on “Non-Communicable Diseases in the Developing World,” (Galambos, Sturchio, and Whitehead (2013); and several relevant studies may be found in “The Road to Universal Health Coverage” by Sturchio, Kickbusch, and Galambos (2019). 2. The United States National Academy of Medicine, in its “Global Roadmap for Healthy Longevity,” defines healthy longevity as the state in which years in good health approach the biological life span, with good physical, cognitive, and social functioning. The report argues that healthy longevity enables well-being across populations and urges that all-of-society transformation is needed to ensure that ageing societies worldwide thrive. Among actions critical for healthy longevity, it highlights sup- porting productive engagement in work, volunteering, and lifelong education; reducing ageism; promoting social inclusion and financial security; improving housing, public spaces, infrastructure, and transportation; increasing the coordination and affordability of equitable healthcare; and investing in public health (National Academies of Medicine 2022). 3. Mahat and Thapa, in an article in the Journal of Nepal Medical Association, state that countries that are trying to take a multi- sectoral approach in addressing NCDs often encounter challenges in operationalizing it. The authors recommend nuancing the multisectoral approach to NCDs in order to better inform its application. They note that often it is limited to the forma- tion of high-level steering committees, which are rarely functional, thus warranting more robust institutional mechanisms. With limited technical capacity, and even less control over financial resources, ministries of health seem to play a weak role in driving a multisectoral approach. The authors identify political commitment and leadership as the most important fac- tors for a successful implementation of whole-of-society and whole-of-government approaches, with the development of common knowledge and understanding along with evidence generation as other critical factors (Mahat and Thapa 2019). 4. George Alleyne and Sania Nishtar, in their work titled “Sectoral Cooperation for the Prevention and Control of NCDs,” attempt to separate multisector from intersectoral action and postulate that the former refers to action involving the various sectors of government, while ‘intersectoral’ refers to the interactions among the government, civil society, and the private sector (Alleyne and Nishtar 2013). 5. Leppo and colleagues, in their book Health in All Policies – Seizing Opportunities, describe how various sectors could incorpo- rate health in all of their policies – the so-called ‘health-in-all policies’ approach. They focus on relevant issues such as health equity, workplace health, mental health, nutrition, alcohol, and tobacco (Leppo et al. 2013). Table 16.1 summarizes the bi-directional rela- sectors are included in a proposed framework that tionship between selected non-health sectors and shows schematically the bi-directional relationship NCDs. These are just a selected set of sectors that between various sectors of the economy and NCDs clearly exemplify such relationships. Additional (Figure 16.6). COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 16 463 TABLE 16.1 Bi-directional relationship between selected non-health sectors and NCDs – some illustrations Sector (Examples) How are NCDs affected by? How do NCDs affect? Education Since NCDs are lifestyle diseases, education NCDs adversely affect educational outcomes both influences knowledge, attitudes, and practices, indirectly (impact on teachers/parents) and directly including positive health seeking behaviors. (malnutrition, type 1 diabetes, accidents/injuries, visual impairments). Social Protection & Labor Inadequate labor laws, occupational health NCDs impact jobs, livelihoods, and labor and industrial safety measures, and social productivity. assurance/insurance programs can cause NCDs or exacerbate their negative impacts. Agriculture Increased availability and affordability of NCDs can affect agricultural production, profits, and diverse foods can decrease NCD prevalence. financial sustainability. Use of insecticides/pesticides can cause NCDs. A literature review did not yield a comprehen- been developed for other purposes (e.g., to monitor sive framework capturing the two-way cause-effect NCD interventions). Therefore, an original frame- relationship between NCDs and relevant sectors of work for multisectoral causes and impacts of NCDs the economy, in order to demonstrate the rationale is proposed in Figure 16.6 and further elaborated in for a WoS approach. The few frameworks relevant the paragraphs that follow. to NCDs available in the literature appear to have FIGURE 16.6 Framework for multisector causation and impact of non-communicable diseases Source: Original figure for this publication 464 CHAPTER 16 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E Education. The education sector is highly rele- in Finland (e.g., Puska and Jaini 2020) showed how vant to all health issues, especially those that require NCDs could be addressed through specific agricul- healthy behaviors for prevention and control, since tural changes (reduced support for dairy, more for inculcating good knowledge, attitudes, and practices fruits and vegetables); city planners’ pushing phys- relevant to health from an early age becomes essen- ical activity; and, critically, strong and consistent tial. NCDs are frequently seen as “lifestyle diseases” leadership by health leaders. Some of that work was – since healthy lifestyles often constitute the most globalized by WHO in the 1990s, with WHO’s Eu- effective preventive device against them. Therefore, ropean and Eastern Mediterranean Regions in par- the relevance of the education system to NCD con- ticular building on these initiatives. trol and prevention is self-evident. There is also a Transportation. As we consider injuries among relationship in the other direction, i.e., NCDs can NCDs, and as road traffic accidents are a significant adversely affect education outcomes both through cause of injuries around the world, the involvement of their debilitating impact on parents and teachers the transportation sector becomes critical to ensure and by directly afflicting children. Examples of road safety, both in terms of the design of trans- common childhood NCDs include but are not lim- portation infrastructure and in terms of improved ited to: malnutrition (including childhood obesity); behaviors – through education as well as regulations. type I diabetes among children; accidents and inju- Environment. Air pollution is a significant ries at home, school, or on playgrounds; pediatric NCD risk factor (added to WHO’s 4x4 framework asthma; and visual impairments – mostly refractive when it was expanded into a 5x5 framework). It errors or a deficiency of vitamin A. Early identi- stands to reason that the ministry of environment fication of NCDs among children and their timely and relevant private sector actors work together to management are often best achieved in the educa- address this and other environmental factors con- tion system, which clearly implies a critical role for tributing to NCDs. teachers as well as parents. Commerce/trade. Nearly every good or service Social protection and labor. Given the poten- that gets traded–internationally, as well as domesti- tially devastating impact of NCDs on livelihoods cally–has an impact on health, positive or negative. (Kankeu et al. 2013), their links to social safety nets Regulating the trade of specific goods with proven and the work of labor ministries become vitally im- deleterious health effects, especially through NCDs, portant. Lack of access to adequate social protection is clearly an area that requires the involvement of can exacerbate the negative effects of chronic debil- this sector/ministry. itating illness on the affected individuals and their Law/justice. Many interventions against NCDs families. Therefore, there is a clear need to develop require a legal framework to be in place, e.g., traffic social protection programs targeting those who suffer safety, access to alcohol, tobacco, narcotics, firearms, from or are affected by NCDs. Occupational health just to name a few. Therefore, the involvement of and industrial safety are areas where the labor min- this sector is self-evident. istry needs to ensure appropriate interventions that Finance. In addition to the critical function of could contribute to the prevention and mitigation of resource allocation to various ministries, ministries NCDs, especially those which are linked to occupa- of finance have a key role to play in taxation, which tional hazards such as industrial pollution, exposure is relevant to interventions that seek to raise the tax to toxic chemicals, and risks of accidents and injury. on unhealthy commodities such as alcohol, tobacco, Agriculture and urban policy. Work under- sugary drinks, etc., and possibly earmarking such taken by Lepp, Puska, and others in North Karelia additional revenues to the control of NCDs. COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 16 465 ENGAGING NON-STATE ACTORS (FOR-PROFIT AND NON-PROFIT) A strong case exists for coopting and engaging non-state actors in preventing, managing, and controlling NCDs. But public-private partnership on NCD prevention and control cannot be premised merely on good intentions; bringing the private sector on board requires a clear-eyed approach on the part of governments, bearing in mind the risks emerging from the profit motive of commercial entities that diverge from public health interests. Several examples of innovative and mutually beneficial public-private partnerships focused on the prevention and control of NCDs are presented, although they are limited in scale and scope. The important role that the non-profit private sector (NGOs and CSOs) can play is stressed, particularly given their ubiquitous presence in LMICs, particularly in the poorest regions and countries of the world, and their comparative advantage with regard to community mobilization. There are several reasons why it makes sense, and involving the private sector in public health glob- is in fact necessary and potentially mutually ben- ally, while acknowledging the existence of distrust eficial, for state actors to co-opt and engage with and skepticism in several quarters; the authors go private sector actors in preventing, managing, and beyond the issue of whether the private sector needs controlling NCDs. The private sector has at least to be engaged and attempt to address the question four motivations to invest in NCDs: (i) its com- of how such engagement could be put to best use mitment to its employees (keeping the workforce (Sturchio and Goel 2012). healthy), (ii) its profit motive (new products and Authors associated with the United States Na- new markets), (iii) its social responsibility (doing tional Institutes of Health have made a strong case the right thing) – (Hancock, Kingo, and Raymond for partnership with the private sector, distinguish- 2011), and (iv) its recognition that such investment ing between “interacting, engaging, and partner- in the well-being of consumers has long-terms ing” (Collins, Mikkelsen, and Axelrod 2019). Their benefits for economic development, which in turn analysis stresses the need for working actively with is good for business. The private sector (except of the private sector for the prevention and control of course those industries producing and selling goods NCDs, rather than merely promoting an interaction that harm human health) is thus uniquely posi- between the public and private sectors. The authors tioned to catalyze and enhance multi-stakeholder suggest several possible opportunities for such en- efforts to address NCDs at the global, national, and gagement/partnership. First, since the private sector local levels. The private sector also has several key already plays a dominant role as a provider of NCD attributes—motivation, geographical presence and services, especially in the curative domain, they reach, resources, convening power, and credibility recommend that governments mounting a com- with local communities—that make it a potential prehensive NCD control and management strategy partner in this effort. Aside from NCD service pro- leverage these already existing capacities through vision, the private sector also offers special capaci- effective public-private partnership (PPP) models. ties, including the creation and adoption of innova- Second, they suggest that a broader engagement tions (Laryea and Cueni 2019), greater production/ with industry could be forged to address CVD and cost efficiencies, new technologies, strong research other NCDs, going beyond the private sector’s role capabilities, and supply chain capacity, that can be in service provision and encompassing domains leveraged to optimize WoS NCD efforts. The role such as technological innovation, controlling input of private actors in NCDs is not limited to service costs so as to reduce the economic burden, conduct- provision and should include a significant emphasis ing high-quality research, improving supply chains, on financing, as well, i.e., the mobilization of private and more. Third, the authors note that moving funding for NCD interventions—either as a part of from private sector engagement to multi-stakehold- corporate social responsibility or even as a core ob- er partnerships calls for the development of new jective. Effective control of NCDs could yield tan- rules of the game, especially if the private sector gible benefits to the private sector, in terms of em- is to be involved through mechanisms other than ployees’ health and productivity, for example. In a contracting; such rules include requiring complete report of the Center for Strategic and International and regular reporting of epidemiological data, link- Studies (CSIS) project on US Leadership in Devel- ing of patient information, telemedicine, and other opment, Sturchio and Goel make a strong case for innovations. Finally, the authors propose the WHO 466 CHAPTER 16 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E Global Coordination Mechanism as an example of (a) single-company efforts addressing NCDs broad- an innovative platform for multi-stakeholder en- ly, e.g., by Novo Nordisk, Vitality, and Teva; these gagement at the global level. Similar coordination firms usually partner with NGOs and have some mechanisms are required at the national level. relationship to the countries or cities in which they Public-private partnership on NCD prevention work; (b) multi-company efforts addressing one or and control cannot be premised merely on good more risk factors: e.g., IFBA, supported by the 10 intentions; bringing the private sector on board re- largest multinational food companies; and (c) a coa- quires a clear-eyed approach on the part of govern- lition of public and private entities addressing CVD, ments (Allen and Bloomfield 2016). It is important to asthma, and diabetes. These initiatives are usually a recognize that the interests of the industries that pro- response to WHO or governments, and they often duce tobacco products, processed foods, or pharma- engage NGO partners. However, they rarely achieve ceuticals often diverge from those of public health. scale for two main reasons: companies need to be Thus, coopting these industries to work towards the encouraged to cooperate and find the non-compet- larger interests of the society can pose significant itive aspects of a relationship where they can all win challenges. Involving private sector partners with together; and they need government leaders who diverging interests comes with risks as well as ben- understand how markets work and know how to efits for public health objectives. Clear policy goals, leverage regulations that encourage change. WHO identification of value-added, creation of appropriate now actively encourages public-private approaches engagement platforms, establishment of statutory/ through its new unit focused on the “commercial regulatory frameworks, and documented agree- determinants of health.” Box 16.2 presents some il- ments on reporting and benefit-sharing are critical lustrative examples (IARD 2019; Laryea and Cueni for effective public-private collaboration on NCDs. 2019). One area where the private sector’s role appears vital It is worth underlining the important role that is fighting malnutrition. In their article titled “Can the non-profit private sector can play, particularly the food industry help tackle the growing global bur- given its ubiquitous presence in LMICs, especially den of undernutrition?”, Yach et al. (2010) argue: “If in the poorest regions and countries of the world. we are to combat global undernutrition successfully, In an article in the British Medical Journal, Katie efforts must be sustained by multiple stakeholders Dain, a Commissioner on the WHO Independent from various sectors. Only through new and innova- High-Level Commission on NCDs and co-chair tive public–private sector partnerships can we truly of the WHO Civil Society Working Group for the make a difference”. Large corporations in the food Third United Nations High Level Meeting (HLM) and beverage industries wield substantial market on NCDs, argues that HIV/AIDS, Ebola, and the cli- power which can translate into political influence. mate crisis provide good examples for civil society Worldwide, the top 10 packaged food companies ac- engagement (Dain 2019). Dain’s article is a part of count for 15.2 percent of sales, while the top 10 soft the “Solutions for Prevention and Control of NCDs” drink companies account for 52.3 percent of sales. In collection, a collaborative output of the WHO Glob- recent years, 10 major multinational food and bev- al Coordination Mechanism on NCDs and the BMJ. erage companies have worked together within the Defining civil society in the broadest terms, Dain as- International Food and Beverage Alliance (IFBA) to serts that the international community has been slow increase their commitments to public health. While to recognize and meaningfully engage civil society engaging industry “heavyweights” is important, for organizations in the response to NCDs. She identi- best results, major multinational companies need to fies supportive legal, social, and policy environments be joined by the myriad of small- and medium-sized and the documentation of best practices as essential enterprises in developing and implementing pro- prerequisites and exhorts the United Nations (UN) grams to improve the health of the public (Alexan- and governments to “walk the talk” in this regard der, Yach, and Mensah 2011). and the NCD community to hold governments to Globally, there are several examples of inno- account. In this context, it is notable that a major vative and mutually beneficial public-private part- review of public-private partnerships seeking to de- nerships focused on the prevention and control of velop quality and affordable new technologies for NCDs, but many of these efforts have been limited health highlighted the general neglect of NCDs and in scale and scope. Such efforts can be categorized as: mental health (Bulc and Ramchandani 2021). COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 16 467 BOX 16.2 Examples of Public-Private Partnerships for Non-Communicable Diseases • Access Accelerated: A platform consisting of 20 biopharmaceuticals; civil society organizations, e.g. UICC, WHF, NCD Alliance, and City Cancer Challenge; and international agencies, e.g., PATH and the World Bank. • “Ask, Listen, Learn”: Responsibility.org, alongside a team of educators specializing in elementary and middle-level education, developed “Ask, Listen, Learn” with the goal of increasing conversations between parents and kids about alcohol and re- ducing underage drinking. The program encourages open, factual conversations between parents, educators, and children about alcohol, and informs kids about the dangers of underage drinking. • AstraZeneca launched Healthy Heart Africa in 2014 to reach 10 million hypertensive patients across Africa over the ensuing 10 years in an effort to address the World Health Organization’s target of reducing premature CVD deaths by 25% by 2025. • The British insurance company Bupa’s Global Chief Medical Officers Network is an alliance of clinicians and executives whose mission is to improve health and wellbeing globally by engaging private sector leaders not typically included in the health conversation in dialogue with governments, civil society, and multilateral agencies to share best practices. • The Global Smoke-free Worksite Challenge is a partnership between Johnson & Johnson, the Mayo Clinic, the American Cancer Society, the US Department of Health and Human Services, and others to reduce tobacco in the workplace. • IARD drunk driving project in the Dominican Republic: The International Alliance for Responsible Drinking (IARD) engaged a wide range of government and private stakeholders – along with academia, NGOs, civil society – and helped lead meetings and workshops about road safety issues. These events allowed many Dominican road-safety agencies and experts to become aware of others’ work and expertise. • International Food and Beverage Alliance (IFBA) is a global alliance of twelve food and beverage companies, dedicated to advancing health and nutrition. Their engagement with WHO led to changes in food and beverage marketing to children across many companies. The Healthy Weight Commitment Foundation was a partnership with the Robert Wood Johnson Foundation, the US President’s Office, and academia. • Eli Lilly launched the Lilly NCD Partnership in 2011, committing US$30 million over five years to leverage public-private part- nerships to combat NCDs in Brazil, India, Mexico, and South Africa. • Merck’s Capacity Advancement Program, established in 2012, is a public-private partnership with Ministries of Health, heath fa- cilities, local communities, universities, and the media to expand diabetes-related health care professional capacity in Africa. • In partnership with Medtronic, the NGO Partners In Health (PIH) launched the NCD Synergies Initiative to facilitate collabora- tion among low-income countries facing similar challenges in their national strategic planning to address NCDs. • Novo Nordisk and several state governments in India have collaborated to improve diagnosis, treatment, and care for diabetes patients. A major impact of Novo Nordisk has been in initiating the Oxford Health Alliance, which built many public-private partnerships, launched a multi-city effort toward ending diabetes, and supported the UN resolution on diabetes that led to the UN NCD resolution. These actions show how some private sector actors have for decades steadily supported a wide range of policy, advocacy, training, and specific projects related to NCDs, including creation of the World Diabetes Alliance. • The Pan American Forum for Action on Non-Communicable Diseases is a public-private partnership whose goal is to address the NCD epidemic in the Americas by implementing and promoting the Pan American Health Organization (PAHO) strategy for prevention and control of NCDs. • The Pfizer Global Health Challenge program has pledged more than US$47 million to support cancer- and tobacco-control organizations to combat NCDs. • The Workplace Wellness Alliance, launched in 2010 as a response to the World Economic Forum Meeting in 2008, aims to make wellness a priority in the workplace to improve employee health and productivity. 468 CHAPTER 16 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E WHAT HAS BEEN ACHIEVED SO FAR, AND WHAT HAVE WE LEARNED? Several examples of WoS approach to NCD control are presented and lessons are gleaned from them. Concrete empirical evidence of the effectiveness of a WoS approach for NCDs is hard to come by because most of the ex- amples are limited in scope, and/or did not include rigorous impact evaluations. However, experience shows that it is vitally important to mitigate three factors that usually impede inter-ministerial cooperation: (a) financial rules, which make it difficult for one ministry to share resources with another; (b) turf battles regarding roles and responsibilities and leadership; and (c) lack of domain knowledge in relevant ministries. Other lessons include: (i) the importance of genuine involvement and active participation of the relevant sectors from the beginning of the process; (ii) identifying clear roles and responsibilities for every sector; (iii) the need for an appropriate institutional home for coordination—physically as well as organizationally; and (iv) the need for a special effort to include non- state actors (i.e., the for-profit private sector, non-profit organizations, civil society, and community-based groups). This section presents examples of WoS approaches vent and control NCDs (WHO “best buys”) in low- that have been adopted in different contexts, based and lower-middle-income countries. While all these on a review of available knowledge, and derives global initiatives did call for WoS actions, that has lessons from these experiences. The approaches remained mostly on the level of rhetoric, with little described below encompass regional and national concrete action taken to adopt such an approach. The NCD programs, as well as programs targeting spe- NCD Alliance, citing a new NCD Countdown 2030 cific NCDs or individual risk factors for NCDs. It Report published by The Lancet, highlights the slow should be stressed that what is presented is just a progress towards SDG target 3.4, pointing out that sample of such efforts/studies around the world, and many countries are falling behind on global commit- not an exhaustive review. The selection is limited by ments to tackling premature deaths from chronic dis- the amount of available evidence about programs eases (NCD Alliance 2020b). that adopted a truly comprehensive approach in- Our overarching finding, based on a literature volving the whole society in actions against NCDs; search, is that concrete empirical evidence of the ef- among such programs, it is even more difficult to fectiveness of a WoS approach for NCDs is limited. find those with rigorous impact evaluations that Unlike nutrition and HIV/AIDS, there is a paucity could provide the required evidence of effectiveness. of policy evaluations and assessments of the impact The global commitment to combat NCDs and of WoS frameworks/approaches on tackling NCDs, calls for whole-of-society action are not new. In 2011, or even of the impacts of specific risk factor initia- the first HLM of the United Nations General Assem- tives on health outcomes. Few studies have exam- bly, in its political declaration, highlighted the impor- ined important aspects like the political economy of tance of prevention and control of NCDs. This was NCDs or undertaken rigorous institutional analy- followed by similar declarations in 2014, 2018, and ses. Nor have there been evaluations of the impli- 2019. The WHO, for its part, put out a Global Action cations of emerging health risks, such as pollution Plan for 2013-20 (NCD GAP), accompanied by nine or climate change, for NCD burdens or national voluntary, global NCD targets for 2025 and 25 indi- responses to these diseases. The evidence is particu- cators to track the progress made.1 In 2015, as part of larly scanty in LMICs. A partial exception concerns the Sustainable Development Goals (SDGs), SDG 3a tobacco taxation and action on sugar sweetened exhorted all countries to strengthen the implementa- beverages (SSBs), though these too have only been tion of the WHO Framework Convention on Tobacco studied in limited settings. Figure 16.7 shows that Control, while with SDG 3.4, countries committed to action against tobacco and SSBs had to involve le- reduce premature mortality from NCDs by one-third gal action, taxation, and engagement of the private and promote mental health and well-being. In 2018, sector. Of course, consumers had to be convinced of WHO also published a document titled “Saving Lives, the importance of these measures and accept them Spending Less: A strategic response to noncommu- for the measures to be successful. Many countries nicable diseases,” which laid out for the first time the lack the necessary regulatory framework to imple- health and economic benefits of implementing the ment such initiatives, and where regulations exist, most cost-effective and feasible interventions to pre- enforcement may be weak. 1   The decision WHA72(11) in 2019 extended the period of the action plan to 2030 to ensure its alignment with the 2030 Agenda for Sus- tainable Development. COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 16 469 FIGURE 16.7 Global policy actions relevant to sugar-sweetened beverages Source: Presentation on ”Evidence base for actions by national actors on regulation of sugar sweetened beverages” by Prof. Yik Ying Teo of the Saw Swee Hock School of Public Health, National University of Singapore at the Prince Mahidol Award Conference (PMAC) 2019. It is also noteworthy that global investments in get 3.a), although tobacco use is “steadily declining.” combating NCDs have been relatively low, despite Progress has been made across countries in reducing the potential for good returns on such investments global suicide rates, from 12.9 (2000) to 10.6 (2016) in terms of both health outcomes and economic per 100,000. Progress on reducing harmful alcohol benefits. A discussion paper by the NCD Alliance ti- consumption (target 3.5), however, has been stag- tled “Bridging the gap on NCDs: from global prom- nant. Moreover, road traffic accidents increased from ises to local progress” identifies significant resource 1.31 million in 2013 to 1.35 million in 2016 (target 3.6 gaps in the first decade of the SDGs and points out per 100,000 population). In the case of tobacco, suc- that global commitments in this area have not yet cess with increasing excise taxes demands complex been followed up (NCD Alliance 2020a). It is crit- health-finance-commerce agreements within govern- ically important that gaps between the resources ments. A WoS approach is needed to resolve this. An needed and those actually available to countries intervention that requires adequate priority is smok- be assessed and remedied if global commitments ing cessation through cheap and effective nicotine re- are to be translated into substantive actions on the placement therapy (NRT) – (Thakur et al. 2021). ground. Such an effort to fill the resource gap calls The World Health Statistics 2020 show that the for effective mobilization of financial and human world is not on track to achieve the objectives of the resources from private as well as public sectors. This NCD GAP and SDG target 3.4 on NCDs. Despite the is an area where a WoS approach would be essential. global attention paid to NCDs over the past two de- cades, progress toward reducing the burden of NCDs Global/regional initiatives has been slow. In 2020, only 31 UN Member States were on track to achieve a 33 percent reduction in Thakur and colleagues reviewed progress and chal- the risk of premature mortality from NCDs by 2030 lenges in achieving NCD targets for the SDGs. They against a 2015 baseline (SDG target 3.4.1). Progress found that only 14 of 181 countries had improved towards the nine voluntary global targets set for 2025 their national monitoring on tobacco use (SDG tar- against a 2010 baseline is also off track. Further- 470 CHAPTER 16 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E more, with the exception of tobacco use, there has various challenges in bringing sectors together to not been a significant change in the trends for NCD develop policies to address the increasing NCD risk factors across WHO regions over the past decade burden in the region (Juma et al. 2018). Stronger (WHO 2020b). On the positive side, Thakur e t a l. coordination mechanisms with clear guidelines, (2021) noted that India adopted the WHO Global including approaches for capacity-building and Action Plan for NCDs 2013-2020 and developed its resource generation, were identified as prerequisites. own national action plan with indicators, including a Such mechanisms should include approaches for target of reducing the number premature deaths by capacity building and resource generation to enable 25 percent by 2025. While the Global Action Plan multisectoral action in NCD policy formulation, lists nine targets and 25 indicators, India has set a implementation, and monitoring of outcomes. tenth target to address household air pollution. Other recommendations included increasing NCD WHO’s “best buy” strategies. The WHO’s “best awareness among various sectors, a health sector buy” interventions are a menu of policy options leadership role, and implementation strategies to and cost-effective interventions for the prevention counter “industry interference” that blocks NCD and control of major NCDs. The menu has six ob- prevention measures (Juma et al. 2018). jectives, the implementation of which, by Member Another review of NCD policies in Africa States, is expected to promote the achievement of synthesized the findings o f 21 studies and found the nine NCD targets by 2025. inadequate studies on NCD policy, unsatisfactory A geopolitical analysis of 151 countries with NCD-related policy development, poor policy regard to NCD policy implementation, however, implementation, lack of policy equity to combat observed that, though most countries have en- NCDs, and lack of data (Kassa and Grace 2022). It dorsed WHO’s NCD “best buy” policies, relatively also identified inadequacies in awareness, funding, little is known about global implementation pat- and inter-sectoral coordination and recommended terns and about the geopolitical factors affecting that policy makers strengthen the inclusion and implementation (Allen et al. 2020). The a nalysis implementation of “best buy” strategies identified found that the mean NCD policy implementation by WHO. score was 49.3 percent in 2017, with Costa Rica and Multisectoral Action Plans (MAPs) and Iran scoring highest (86.1 percent) and Haiti and National Strategic Plans. Perhaps inspired by the South Sudan scoring lowest (5.5 percent); between examples of HIV/AIDS and nutrition, multisectoral 2015 and 2017, the scores rose in 109 countries action plans and/or national strategic plans against but regressed in 32 countries. The most common- NCDs have been attempted in some countries. An ly implemented policies were clinical guidelines, example of concrete steps towards multisectoral graphic warnings on tobacco products, and NCD action against NCDs is the development and risk factor surveys. Nutrition-related policies saw promotion of the NCD MAP toolkit by WHO, to gains, while those related to alcohol and physical assist policymakers and program managers in activity were the most likely to have been less effec- developing, implementing, and monitoring national tively implemented or even dropped. multisectoral plans. It covers the main steps from Similarly, assessing progress under Health situation assessment, stakeholder engagement, and 2020 in WHO’s European Region, Zuidberg et al. setting national NCD targets to implementation and (2020) found a wide variation of the ‘Health 2020 monitoring and evaluation. “Best buys” form the Index’. Scores ranged from 82.8 in Sweden to 30.0 basis for NCD Multisectoral Action Programs in Turkmenistan, and the study noted a clear east- (MAPs) in many countries of the world. west gradient, with countries in the west perform- In Türkiye, for example, the MAP for NCDs ing better than those in the east. While all quintiles 2017-2025 has two phases. The underlying deter- showed improvement in terms of premature mor- minants of NCDs and the risk factors mean that tality, mortality from external causes, life expectan- multisectoral, whole-of-government, and WoS res- cy and infant mortality, there was a uniform lack of ponses are required to prevent and control them. The progress on overweight. influence of public policy in sectors such as trade, A review of multisectoral action in NCD preven- taxation, agriculture, urban development, and food tion policy development in five Sub-Saharan African production is frequently required for health gains to be countries (Cameroon, Kenya, Malawi, Nigeria, and made in the area of NCDs. Intersectoral action can be South Africa) concluded that such multi-sectoral complex and challenging but there is a sense in Türkiye action is possible in African countries but identified that adequate experience now exists about which institu- COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 16 471 tional processes promote intersectoral policy practice. vided the mandate to implement policies and pro- In the Caribbean countries, all the NCD MAPs grams relating to prevention and control of NCDs. include at least nine of the 14 WHO “best buys” pol- icy options, and all include the “best buys” related to • Threats: a) Lack of support from other agencies; smoke-free environments, unhealthy diet, physical b) Political change may affect the direction of the activity, and cervical cancer screening. Countries in overall agenda. this region strongly emphasize improving awareness and strengthening and enforcing policies, legisla- Illustrative national initiatives tion, and regulations to enable risk factor prevention and control, especially relating to unhealthy diets. Japan. Multisectoral cooperation in NCD control A study of Barbados’ Multi-sectoral National NCD in Japan has been well established and practiced. Commission reported that, while the country was “Health Japan 21,” initiated in 2000, was the third na- successful in engaging with leaders in civil society tional health promotion campaign aimed to establish organizations and the private sector, the engage- professional health care teams at primary care level. ments among different government ministries were In Japan, decision-making and implementation of limited (Unwin et al. 2016). It has been noted that the NCD control are usually based on evidence (Wu mechanisms required for cooperation between sec- et al. 2017). Survey data are used to predict trends tors within the government are different from those in NCDs and related behavioral risk factors, and to required for cooperation between different parts of evaluate the effects of prevention and control strat- the society, and between the government and non- egies and measures. Implementation of “Health Ja- state actors. Therefore, it is important to find appro- pan 21” across different areas has provided evidence priate mechanisms for both “whole-of-government” for the government to make further decisions. The and “whole-of-society” approaches. Japanese government also provides specific funds to In Malaysia, an analysis of the 2010-2014 Na- support applied research on NCD control. tional Strategic Plan for NCDs found that there has Japan has a horizontal management system of been a strong unification of “whole-of-government” NCDs under a decentralized model (Shibuya et al. and multisectoral involvement in the prevention of 2011) as highlighted in Table 16.2. The system has NCDs. However, the lack of measurable indicators contributed much to the success of NCD control and evaluation criteria has hampered assessment of in Japan. Various major actors also participated in the progress achieved. The new 2016 to 2025 plan creating and implementing NCD policies in Japan, has therefore highlighted overarching principles, including the central government, ministries, local including multisectoral action, evidence-based government, community health care professionals, strategies, use of a lifecycle/life-course approach, insurers, and private industries. One distinctive and empowerment of people and communities. The point of Japan’s health care system is the prominence Malaysian government also conducted a SWOT of community health workers (Kabayama et al. analysis of its National Strategic Plan for NCDs 2014). Another noteworthy aspect of Health Japan 2010-2014, and found the following: 21 (2013–2022) is the active engagement of the pri- vate sector. The plan includes a target to, “increase • Strengths: a) The unification of “whole-of gov- the number of corporations that deal with health ernment” and multisectoral involvement of other promotion and educational activities.” Between 2012 agencies in the prevention of NCDs. and 2022, the number of corporations that were ex- pected to achieve the proposed target rose from 420 • Weaknesses: a) Lack of indicators and evaluation to 3,010. As part of Health Japan 21, under the lead- criteria; b) No specific target for each strategy ership of the Ministry of Health, Labor, and Welfare outlined; c) No specific measurable achievements also launched the Smart Life Project to encourage in each strategy. three concrete strategies: (i) proper exercise, (ii) ap- propriate dietary habits, and (iii) smoking cessation. • Opportunities: a) Development of public-private The project encourages corporations to take the nec- partnerships; b) Document recognized worldwide essary measures for tackling NCDs and commends and highlighted as an example by WHO; c) Pro- corporations with outstanding outcomes.2 2   Government of Japan, Ministry of Health, Labor, and Welfare. Health Service Bureau. 472 CHAPTER 16 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E TABLE 16.2 NCD Control and Prevention System in Japan “Healthcare at age 40+ years, medical care at age 70+ years” Key features: Law-based MS governance + Decentralized NCD management + Focus on primary prevention (“Lifestyle Diseases”) Development Unhealthy lifestyle  Risk Status  NCD  Rehabilitation Focuses Overeating Overweight Cancers Smoking High blood pressure Diabetes Drinking High glucose CVD Not exercising High TC Stroke Living with Stress Actions Health Japan 21 Specific health check/ Surveillance Selfcare guidance Treatment Long-term care Cancer screening Systems A national health Mass screening program Cancer registry Long-term care system promotion and disease Health system prevention initiative • Health insurance • Healthcare delivery Acts/Laws Health Promotion Act Act on Assurance of Cancer Registration The Long-term Care Medical Care for the Elderly Promotion Act Insurance Act Health Promotion Act Health Insurance Act Act on Assurance of Medical Care for the Elderly Source: Wu et al. (2017) China. In China, NCD control measures main- China also recognizes the importance of mul- ly include three aspects: (i) surveys or surveillance tisectoral cooperation in NCD control. In 2012, 15 of NCDs and related risk factors, (ii) health edu- Chinese government ministries jointly issued the cation and health promotion in the general pop- “Work plan on NCD prevention and control in China ulation, and (iii) comprehensive interventions in (2012-2015)” to propose mechanisms of multisec- high-risk populations (Table 16.3). Of the three as- toral cooperation and define responsibilities of rel- pects, management of high-risk populations is the evant sectors. The publication of the work plan was most important in China. NCD control tasks were a landmark for NCD control in China. However, designed to be shared by multiple organizations in due to the lack of political incentives and effective the Chinese context, including national and local mechanisms, along with conflicts of interest, mul- Centers for Disease Prevention and Control (CDC), tisectoral cooperation in NCD control remains a hospitals, institutions of health education, and the partially fulfilled promise of the central and local Institute of Maternal and Children Healthcare. In governments. One of the major risk factors that re- addition, the National Cancer Center, National main inadequately addressed in China is tobacco Center for Cardiovascular Disease, Prevention Of- in contrast with the rapid decline in tobacco use in fice for Endemic Diseases, and other academic or- Japan (Wu et al. 2017). ganizations are also involved in NCD control. COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 16 473 TABLE 16.3 Major NCD Management Initiatives in China Key features: Surveys of NCDs/risk factors + High-risk population focus + Multiple agencies involved but not truly WoG Issuing Institutes Types of Policies Topics Number of Policy Documents Overall 1984-1999 2000-2008 2009 & after General Office of the Decisions, Opinions, Re- Medical reform; Population aging; Sports; 11 1 5 5 Communist Party of ports, Notices Rural healthcare service; Tobacco control. China National People’s Laws and Regulations Environment protection; Air pollution; 10 5 1 4 Congress Advertisement; Sports; Nationwide fitness program; Food safety; Social insurance; Occupational disease prevention; Health insurance for minors and elderly. State Council of China Regulations, Decisions, Medical reform; Population aging and 29 1 6 22 Opinions, Outlines, Reports, pension service; Health development plan- Plans, Rules, Notices ning; Planning for national economic and social development; Health service; Food and nutrition; Nationwide fitness program; Patriotic health movement; Mental health; Medical and healthcare service. Ministry of Health Decree of Minister, Opin- Medical reform; Disease prevention (CVD, 50 0 14 36 ions, Guidelines, Standards, Cancers, Diabetes, Dental hygiene, Mental Regulations, Protocols, health, Osteoporosis, etc.); Health literacy; Rules, Outlines, Bulletins, Health education; Tobacco control; Health Plans, Reports, Notices talent team construction; New rural cooperative medical system; Medicines; Medical rescuing; Elderly health; Floating population health. Other Ministries Decree of Minister, Opin- Medical reform; Medical insurance; ions, Notices, Standards, Medical rescuing; Aging pension; Student Outlines, Protocols, Rules, sports; Chinese medicine; Food safety; Regulations Tobacco control; Health of persons with disabilities; Community health; Nation- wide fitness program, and others. Total 164 13 51 100 Source: Wu et al. (2017) The “Healthy China 2030” strategy, issued in al. 2017). Relevant similarities between the two coun- 2016, proposed a basic roadmap of “work together tries include legislative and policy approaches, stan- and share together.” The plan involves efforts from dardized management of NCDs, and multisectoral the supply and demand sides and the involvement cooperation. However, differences between the coun- of societies, industries, and individuals, creating tries’ health systems and NCD control priorities have a powerful force to promote health. The plan will resulted in large divergences in epidemiological sta- greatly improve multisectoral cooperation for NCD tus and disease burden. The authors suggest that Chi- control in China. China has established a nearly na and Japan could benefit from increasing exchange perfect reporting system for infectious diseases, of ideas, experiences, knowledge, and technologies. but not for NCDs. For example, the country’s can- Singapore. In Singapore, the Healthy Living cer registry system covers only about 300 cities or Master Plan (HLMP) Taskforce was formed in Sep- counties; surveillance of behavioral risk factors has tember 2012 to facilitate integration of nationwide been carried out every three years since 2009; and efforts to enable healthy living (Figure 16.8). One of the CVD registry system was just recently initiated. the main aims of this Taskforce was to review the A comparative review of NCD control efforts in current health promotion and disease prevention China and Japan concluded that, while the two coun- landscape in Singapore, primarily in the context tries share certain similarities, differences in their of the country’s high rates of NCDs. The Taskforce respective approaches are reflected in results (Wu et sought to formulate a systematic plan that took into 474 CHAPTER 16 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E consideration the inputs of various public agencies encing behavior change for better health. and community organizations, along with public To achieve the vision of “healthy living every day,” consultation, and strengthened interagency collab- the strategy was to have seamless connections across oration by building on existing infrastructure and Singapore, in terms of the physical environment, be- ideas. The HLMP Taskforce adopted an inclusive tween people, and between the government and the and integrative approach by involving the communi- citizens. This was to be carried out through three an- ty in a co-creation process. All these elements led to chor strategies that influence the adoption of healthy a master plan that outlined strategies to ensure that behaviors, namely: (i) building a healthy workplace all Singaporeans have access to a healthy lifestyle, ecosystem through a worker-centric and sector-cen- so that healthy living becomes natural and effort- tric approach, facilitated by tripartite partnerships; (ii) less for everyone. The HLMP consolidated healthy integrating physical and social environments in the living initiatives and ideas from public and private community setting, by working across government agencies and the community, including the Regional and with the community; and (iii) creating greater Health Systems (RHS), to build momentum in influ- awareness for healthy living among the young. FIGURE 16.8 Singapore’s Healthy Living Master Plan Source: Presentation on “Whole-of Government Strategies for NCDs” by Prof. Yik Ying Teo of the Saw Swee Hock School of Public Health, National University of Singapore at the Prince Mahidol Award Conference (PMAC) 2019. While this study was unable to find any system- ure 16.9). This campaign highlights the important atic evaluations of the HLMP, it has been suggested role played by “champions” (in this case the Minister that Singapore’s concept of chronic disease manage- of Health) towards a specific goal. It sought to rally ment and the plan’s three-level theoretical frame- the entire nation to rein in Singapore’s diabetes and work are successful. The HLMP may provide useful obesity epidemics.  The government established a lessons and serve as a positive example for the rest multi-ministry task force to conceptualize and im- of the world (Su et al. 2019). plement a series of programs, focused mainly on In addition to the HLMP, in the face of an im- encouraging Singaporeans to take individual respon- pending diabetes epidemic, Singapore launched a sibility for their health while adopting a whole-of-so- new strategic campaign in 2016, announced by the ciety approach to promote regular exercise, healthy then Minister of Health in a parliamentary speech, dietary habits, early disease detection, and better dis- which was referred to as the “War on Diabetes” (Fig- ease treatment (Bee, Tai, and Wong 2022). COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 16 475 FIGURE 16.9 Singapore’s “War on Diabetes” Source: Presentation on “Turning Discovery into Healthier Communities” by Prof. Yik Ying Teo of the Saw Swee Hock School of Public Health, National University of Singa- pore at the Prince Mahidol Award Conference (PMAC) 2019. Singapore’s five-year effort in its “War on Diabe- Thailand. Thailand established its ThaiHealth tes” has not yet impacted the prevalence of diabetes health promotion program in 2001, the first ini- and obesity. However, overall, significant progress tiative of its kind in Asia. The effort followed the has been seen in many areas. The proportion of Sin- launch of similar organizations in some high-in- gaporeans engaging in regular leisure-time exercise come countries, notably Australia, New Zealand, increased from 29·4 percent in 2017 to 33·4 percent and Switzerland (Figure 16.10). Created under in 2020. Consumers have also shifted to buying Thailand’s Health Promotion Foundation Act 2001, healthier versions of food products, pushing up the and particularly relevant to the high NCD rates retail share of healthier food products by more than in the country, ThaiHealth’s mission is to support 1·5 times compared with five years ago. There have and develop health promotion processes leading to also been significant improvements in disease man- good health for Thai people and society. It pursues agement, with the rate for major lower extremity this goal by supporting rather than replacing groups amputations declining from 116 to 66·3 per 100,000 and organizations already working on public health patients from 2015 to 2019. Going forward, Singa- issues. ThaiHealth’s approach to health promotion pore’s “War on Diabetes” will be moving into a new is framed around a capacity building model in line phase, with a stronger focus on preventive health. with its enabling Act, which emphasizes building ca- Through the Healthier Singapore strategy3, the Sin- pacities in communities, government and non-gov- gapore Government will support upstream efforts ernmental organizations, public interest organiza- to keep individuals healthy and drive early interven- tions, state enterprises, and other agencies to plan, tion. For continuity of care, each resident will en- develop, and conduct their own health promotion roll with a family physician, who will support them programs (Caroll, Wood, and Tantivess 2007). throughout their life for different health care needs. 3   the Singapore MoH’s strategy for championing the national shift towards a population health approach, the details of which were an- nounced at the Committee of Supply 2022 by the Minister for Health, Mr. Ong Ye Kung. 476 CHAPTER 16 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E FIGURE 16.10 Thailand’s ThaiHealth Program Source: Presentation by Pairoj Saonuam on “ThaiHealth” at Prince Mahidol Annual Conference (PMAC) 2019. Pairoj Saonuam is the Assistant Chief Executive Officer, Thai Health Promotion Foundation (ThaiHealth) and the acting Director of Promoting Healthy Lifestyle Section, Thai Health Promotion Foundation. By all accounts, ThaiHealth has been able to ty, physical activity, and other health risk factors re- play an active role in supporting and accelerating vealed a significant level of activity being generated the commitment to health promotion espoused in by ThaiHealth; strategies included research, aware- the NHDP, Healthy Thailand, and the Joining Forc- ness and education, social mobilization, capacity es for Health Promotion Policy. An internal review building, and policy development. But the review noted that ThaiHealth is operating at a level com- also noted a tendency for ThaiHealth to merely mensurate with other such programs when they document success descriptively, without hard data, were five years old.4 The initiative is particularly pro- which made it difficult for the reviewers to gauge active and advanced in its workings with networks the actual impact of activities on health-related at- and partners, policy development, and engagement titudes, knowledge, intentions, and behaviors. This with community groups. In these areas, ThaiHealth issue was identified as an area warranting greater could be seen as a role model for others. Conversely, attention in the future. the areas of evaluation and health inequalities have been less comprehensively addressed by ThaiHealth Programs targeting NCD risk factors relative to its counterparts in other countries. Overall, the internal review concluded that Parry, West, and Laxminarayan (2009) present an it is difficult to isolate and quantify the results di- analytical and simulation framework for assessing rectly attributable to ThaiHealth. The reviewers the optimal levels and welfare effects of alcohol taxes confirmed that there have been notable downward and drunk-driver penalties, accounting for both ex- trends in a number of risk-related behaviors since ternalities and how policies interact with the broader the establishment of ThaiHealth, including the use fiscal system. Their main findings are that, across a of tobacco, and injuries and deaths associated with wide range of plausible scenarios, the fiscal compo- road accidents. An analysis of program activities in nent of the optimal (revenue-neutral) alcohol tax is relation to the issues of tobacco, alcohol, road safe- positive and quantitatively important. In many cases, 4   Presentation by Pairoj Saonuam at PMAC 2019. COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 16 477 it is larger than the externality-correcting component are compelling—although the studies are often spon- of the optimal tax, particularly if alcohol demand is sored by industry and must be validated objectively. fairly inelastic. Therefore, under most parameter Lowering the costs of healthy foods in supermarkets scenarios, fiscal considerations significantly rein- not only increases the quantity of fruits, vegetables, force the case for raising alcohol taxes. The authors and whole grains that people eat, but it also seems to claim that even modest increases in alcohol taxes can reduce people’s consumption of less nutritionally de- produce substantial welfare gains; they reckon, for sirable foods. The South African program offers en- example, that doubling the current tax from 12 to 24 couraging evidence that lowering the cost of healthi- percent of producer prices generates estimated annu- er foods can motivate people to substantially improve al welfare gains of around US$3 to 10 billion or more. their diets. Behavior changes also seem to be propor- A review on Colombia’s experience with to- tional to price changes (Strum and Yach 2013). bacco taxes found that tobacco tax hikes reduced Barton and Grant (2013) review the progress the number of smokers (from 4.51 to 3.45 million of the European Healthy Cities program in a report smokers) and smoking intensity, resulting in a drop entitled “Urban Planning for Healthy Cities” and em- in the number of cigarettes smoked in Colombia phasize the need for careful planning (Barton and (from 332.3 to 215.5 million 20-stick packs) – (Mal- Grant 2013). The study finds that the number of cities donado et al. 2022). Such reduction is expected to achieving a good level of understanding and activity decrease premature mortality, health care costs, in healthy urban planning has risen very substantially poverty, and people facing catastrophic expenditure over the study period. In particular, those achieving on healthcare, to increase health, income, and gen- effective strategic integration of health and planning der equity, and to strengthen domestic resource mo- have increased. The authors identify the development bilization even in the presence of illicit cigarettes. of planning frameworks which advance public health In a systematic review of 86 studies and a me- concerns in a spatial policy context driven often by ta-analysis of 62 studies, taxes on SSB were associ- market forces as a key challenge and posit that a ated with higher prices of targeted beverages (tax health-in-all-policies approach could be valuable. pass-through of 82 percent) and 15 percent lower Lencucha, Drope, and Chavez (2015) reviewed SSB sales, with a price elasticity of demand of −1.59 whole-of-government approaches to NCDs, specif- (Andreyeva et al. 2022a). No negative changes in ically with regard to tobacco control in the Phil- employment were identified. The authors opine ippines and found two salient challenges: (i) the that these findings suggest that while, SSB taxes inclusion of industry representation on Philippines’ may work as intended in reducing demand for SSBs Inter-Agency Committee (tobacco) and (ii) the at- through higher prices, further research is need- tempt to consolidate the responsibilities of different ed to understand their associations with diet and departments through a policy of “balance” between health outcomes and heterogeneity of consumer health and commercial interests. The authors con- responses. In another review of 54 studies and me- clude that whole-of-government approaches hold ta-analysis of 15 studies (Andreyeva et al. 2022b), promise for policy coherence across sectors, but subsidies for fruits and vegetables were associated that not all such approaches best serve health ob- with increased fruit and vegetable sales, with a price jectives. They also emphasize the importance not elasticity of −0.59, whereas changes in consumption only of structural features of whole-of-government were statistically insignificant. Evidence on other approaches but also the institutional culture and outcomes of food taxes and subsidies was limited, entrenched political and economic interests. The and the review suggested that additional research is authors strongly recommend that whole-of-gov- needed to ascertain the implications of such policies ernment approaches should exclude representatives for consumption, diet, and health outcomes. of tobacco, food, and alcohol industries in view of For the past four years, South Africa’s largest their conflicts of interest with the health objectives health insurer has operated an innovative program at hand; they point to the risk of co-option between called HealthyFood for its members. Participants re- the regulated and regulators to buttress their argu- ceive a 25 percent rebate on healthy foods (as defined ments. This clearly leads to a debate about specific by international dietary guidelines) in 800 supermar- aspects of NCD control where public-private part- kets nationwide. More than 300,000 middle-income nerships would be helpful, and what kinds of design South Africans are participating. So far, the results elements may need to be adopted in these cases. 478 CHAPTER 16 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E World Bank Group (WBG)-financed operations nities. In particular, the authors concluded that the Bank’s clients may not be fully served with the exist- In this section we examine the World Bank Group’s ing structure of sectoral oversight, budget incentives, support to addressing NCDs through its investment or knowledge systems. They therefore recommended projects and development policy lending. A review a shift from “accidental” or “opportunistic” multisec- of World Bank Group investments by Merrick et toral approaches to a more “strategic” approach. al. (2012) showed that in the preceding 22-year pe- The World Bank’s investment lending. A re- riod, the World Bank had approved 1,499 relevant view of the World Bank’s investment lending for this projects, of which 947 (63 percent) were in the non- study showed that, between FY16 to FY20, a total health sectors but supported health. It should be not- of 32 Bank-financed operations that focused on ed that at US$16.3 billion, financing for these non- tackling NCDs were initiated, amounting to a total health projects was equivalent to 36.5 percent of the commitment of over US$2 billion. The investment US$44.6 billion that was channeled to advance health volume and number of projects peaked in FY18 and goals through more conventional health-sector proj- FY19, respectively (Figure 16.11). Such investment ects during the period. The top five non-health sec- included both ‘hardware’ (e.g., infrastructure, equip- tors that received funding for health activities were: ment) and ‘software’ (e.g., training and capacity de- social protection, economic policy, transport, edu- velopment of health workforce). In such a macro cation, and the urban sector. The authors conclud- analysis of these investments, it is difficult to break ed that the Bank has both the instruments and the down the elements of investment, but it is import- demand for tackling health problems through other ant to stress that investments in the hardware would sectors. However, accompanying the many successes not be effective without concomitant strengthening that they observed, they also found missed opportu- of the software inputs, particularly human resources. FIGURE 16.11 World Bank investment lending for NCDs between FY16 and FY20 Source: Original figure for this publication based on a review of the World Bank’s portfolio of new IDA and IBRD commitments for FY16-FY20. COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 16 479 Investment Project Financing (IPF). Between cluding: (i) health systems strengthening, (ii) health FY16 and FY20, 10 IPF operations had a large – i.e., service delivery, (iii) NCD risk reduction, and (iv) greater than 50 percent of lending – focus on NCDs early detection and screening of NCDs. Health (see Table 16.4). Twenty-four percent of the IPF system strengthening and quality improvement in projects contributing to NCD control were in the healthcare services and delivery have been the over- Eastern Europe and Central Asia Region, 8 percent arching themes among the WBG’s NCD projects. in Africa, 7 percent each in Latin America and the There are also a handful of projects that explore Caribbean and the South Asia Regions, 5 percent in links between NCDs and nutrition, early childhood the Middle East and North Africa, and 4 percent in development, and service delivery in fragile, con- the East Asia and Pacific Region. flict-affected, and violence (FCV) settings. The projects focused on various activities in- TABLE 16.4 WBG IPF Operations with a >50 percent focus on NCDs Commitment Project ID Project Name Region Country Lead GP / Global Themes (US$ Million) P163721 Sri Lanka: Primary Health Care System South Asia Sri Lanka Health, Nutrition & 200.00 Strengthening Project Population P164452 Integrated Public Provision of Health Care Latin America & the Nicaragua Health, Nutrition & 60.00 Services Caribbean Population P163387 Jordan Emergency Health Project Middle East and North Jordan Health, Nutrition & 177.20 Africa Population P163345 Supporting Effective Universal Health Latin America & the Argentina Health, Nutrition & 300.00 Coverage in Argentina Caribbean Population P169677 Growing Up Healthy Together: Compre- Latin America & the El Salvador Health, Nutrition & 250.00 hensive Early Childhood Development In Caribbean Population El Salvador P167581 Andhra Pradesh Health Systems Strength- South Asia India Health, Nutrition & 328.00 ening Project Population P133193 Protecting Vulnerable People Against Non- Latin America & the Argentina Health, Nutrition & 303.82 communicable Diseases Project Caribbean Population P163476 Lebanon Health Resilience Project Middle East and North Lebanon Health, Nutrition & 95.80 Africa Population P162069 Nigeria- Accelerating Nutrition Results Africa Nigeria Health, Nutrition & 225.00 Population P166783 Saint Lucia Health System Strengthening Latin America & the St. Lucia Health, Nutrition & 20.00 Project Caribbean Population Source: Based on a portfolio review of World Bank investment lending between FY16 and FY20. Through these projects, the World Bank pro- Population (HNP) Global Practice, some NCD in- vided support to a broad range of activities and terventions were also included in projects led by interventions extending beyond the health sector, the Education, Transport, Macroeconomics, Trade including: (i) legislation for specific products, e.g., and Investment, and Finance, Competitiveness, tobacco and alcohol; (ii) performance-based fi- and Innovation (FCI) Global Practices. Many of nancing (PBF); (iii) cash transfers; (iv) communi- the non-health interventions have worked well, if ty-based approaches; (v) aging and aged care; and only on a small scale. It should be noted, however, (vi) taxation and excise duties. While the majority that although some of the World Bank operations of projects were led by the Health, Nutrition, and involved more than one sector ministry, few can be 480 CHAPTER 16 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E said to have adopted a deliberate WoS approach. improvement than is commonly thought; and (ii) The following sub-sections summarize some sa- significant progress in reducing cardiovascular risk lient aspects of the reviewed IPF operations focusing factors and other cost-effective NCD control mea- on NCDs, which may serve as useful case studies. sures has been achieved in advanced health systems with widely varying institutional designs for pri- Health mary care payment and organization. This points to the importance of providing the right clinical i. Primary health service expansion: interventions in the right setting to achieve impact, rather than focusing exclusively on macro-level in- • In India, the Andhra Pradesh Health Systems stitutional reform initiatives (World Bank 2018c). Project (P167581) drew in its design on the expe- rience of tackling NCDs in low- and middle- in- • In St. Lucia, the Health System Strengthening come countries. The project considered global scheme (P166783) focuses on strengthening evidence suggesting that patients with NCDs that NCD management at the primary care level. It require care over a long period can only be man- aims to strengthen the quality of care for diabe- aged equitably through health systems based on tes and hypertension based on standard care pro- primary health care. The project adopted a total tocols and provide financial rewards for health risk approach to screening for NCDs, enabling facilities according to the achievement of results early detection that is administered to those visit- (World Bank 2018d). ing primary health clinics. Cost-effective methods of screening, such as visual inspection with acetic • In Türkiye, NCD prevention and treatment have acid (for cervical cancers), are being adopted as the been integrated into primary care services, which appropriate option (World Bank 2019b; Kak 2019). are delivered at Family Medicine Centers and community health centers with support from the • World Bank support in Jordan (P170529) and Bank’s Health Systems Strengthening and Sup- Yemen (P167195), is supporting the countries’ port Project (P152799). While the system is set governments to improve coverage and quality up to be proactive and plan for NCD manage- of primary health care, including early detection ment, challenges also lie ahead on the coverage and preventive interventions that contribute to of NCD prevention, screening, diagnosis, and reducing the burden of disease of NCDs (World curative services. There is also a need to better Bank 2019c, 2019d). link NCD prevention with performance-based financing at the primary care level as well as to • In Sri Lanka, the Primary Health Care System integrate standard clinical pathways to deal with Strengthening Project (P163721) aims to increase NCDs in the e-health system (World Bank 2015). the utilization and quality of primary health care services, with an emphasis on detection and man- ii. Community-based approaches: agement of NCDs in high-risk populations. The Ministry of Health supports screening of women • In Tamil Nadu, India’s sixth most populous state, for cervical cancer, targeting women between ages NCDs account for nearly 70 percent of deaths 35 and 45 years. Two key lessons from the global and 65 percent of DALYs lost. The World Bank experience with addressing NCDs through prima- has provided a US$287 million loan to strength ry care that Sri Lanka brings to its program include: management of NCDs, with a focus on popula- (i) after tobacco cessation (an agenda on which Sri tion-based screenings of NCDs among the eligi- Lanka has long been ahead of the curve), the out- ble population (P166373, World Bank 2019a). patient management of cardiovascular risk factors such as hypertension and diabetes. This has been • In Anhui province, China, a Bank-financed the single most effective intervention to reduce project (P154716) also put emphasis on commu- mortality due to NCDs in advanced health systems, nity-based and home-based services and deliv- accounting for up to 25 percent of longevity gains ery and management of nursing care to ensure over several decades. In other words, hospital-based cost-efficiency, consumer satisfaction, and qual- care has played a much smaller role in overall health ity of care (World Bank 2018b). COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 16 481 iii. Performance-based financing: particularly in the context of a new excise tax policy that was to be introduced in January 2021. While • In Nicaragua, results-based financing would there have been hurdles to implementation, such support the improvement of maternal and child as the COVID-19 pandemic, the Government of health care services and promote the prevention Samoa has taken several steps on tobacco taxation, of NCDs (P164452). Diabetes, hypertension, and such as the current Tobacco Policy and Action Plan cervical cancer are the main investment priorities, (2014-2024), the Tobacco Excise Bill of 2018, as along with monitoring mechanisms to ensure well as the more recent Excise Bill Amendment ap- that diagnosed beneficiaries receive the neces- proved in July 2023 (Government of Samoa 2019, sary care. Partners have found that performance 2023; Sagapolutele 2018; Tong 2023). agreements are a useful tool for establishing a re- sults-oriented culture (World Bank 2018a). Political economy • In Tajikistan, the World Bank is supporting the i. Government stewardship: government’s performance-based financing scheme (P170358) to improve the coverage and quality of ba- • In El Salvador, the Promoting Human Capital in sic primary health services in rural health facilities in Health project (P169677) supports the expansion 10 districts. On the supply side, the Ministry of Health of the line of care for selected NCDs to all of the contracts rural health centers to enhance both the country’s health networks that tested in three net- quantity and quality of NCD services, providing cov- works under an earlier (and ongoing) NCD Project erage in about 450 health facilities to around 15 per- (P164356). A critical building block is strong po- cent of the country’s population (World Bank 2019f). litical commitment at the highest level of govern- ment and within the Ministry of Health to enhance iv. Evidence-based approaches: the efficiency of service delivery and improve the quality of health care (World Bank 2018e, 2020). • In Samoa, the development of evidence based NCD management has paved the way to guide health • In China’s Anhui Province, the World Bank fi- workers through the clinical decision-making pro- nanced its first-ever project focusing exclusively cess. The Samoa Health Strengthening Program for on the development of an elderly care service Results (PforR) (P164382) discussed the mobile-app system (P154716). Supporting development of patient tracking system and eHealth framework with government stewardship capacity is the first and the Ministry of Health. Consequently, a health data foremost requirement to expand social service aggregation and visualization tool was initiated in Sa- provision (World Bank 2018b). moa in 2020 and, despite delays due to the COVID-19 pandemic and changes in government, was made ful- ii. “Whole-of-government” mechanisms within the ly operational in mid-2021 (World Bank 2019e). “whole-of-society” approach: Economy and fiscal policy • We have already alluded above to the vital im- portance of fostering “whole-of-government” i. Taxation: approaches separately from the broader rubric of “whole-of-society” approaches. This principle has • In Samoa, the Government has committed to in- become abundantly clear in the implementation creasing the excise tax on tobacco products. Ac- of World Bank projects globally. cording to the latest (2017) WHO data, the total tax incidence of the most-sold cigarette brand in Development Policy Lending (DPL) operations. Samoa was 51.6 percent of the retail price, with Among the approved DPLs from 2005 to FY20 Q3, there the excise tax accounting for 38.5 percent of retail were four prior actions that directly addressed policy re- price. This is significantly below WHO’s recom- form and intervention on NCDs, ranging from public mendation of at least 70 percent excise tax share in health to budgetary and legal dimensions (Table 16.5). the final consumer price of a tobacco product. Con- The specific prior actions identified include launching currently, the World Bank has recommended that performance-based contracting at primary health care the Government of Samoa take nutritional content level, adjusting excise duty, and approving an Alcohol into account when designing excise taxes on food, Control Bill to minimize the harmful effects of alcohol. 482 CHAPTER 16 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E TABLE 16.5 DPL operations relevant to NCDs supported by the World Bank Group, 2009-2019 FY Project ID Region Country Type Prior Action 2009 P106502 AFR Cabo Verde IDA Approved the “Strategy for Combating NCDs” as evidenced in the letter issued by the Recipient’s Minister of the Presidency and Council of Ministers, State Reform, and Defense. 2011 P116451 ECA Armenia IBRD/ IDA The Borrower has issued a Government Protocol Decree #3 dated January 29, 2010, on NCDs and launched performance-based contracting at primary health care level, including defined NCD services. 2015 P149963 EAP Tonga IDA The Recipient’s Ministry of Finance and National Planning has, through its 2014/2015 Budget Statement, adjusted the excise duty regime in line with the NCD health policy. 2019 P165928 EAP Samoa IDA The Recipient, through its Cabinet, has approved the Alcohol Control Bill, to minimize harmful effects from the consumption and abuse of alcohol. Source: Based on a portfolio review of World Bank’s Development Policy Lending between 2009 and 2019. IFC’s engagement with the private sector. The (this model could be replicated for social health in- International Finance Corporation (IFC), the pri- surance programs); (iii) in Asia, IFC has invested vate sector arm of the World Bank Group, supports with MiCare, a third-party administrator working public-private partnership (PPP) activities relevant with social and private health insurers; the firm is to NCD control and prevention. Examples include: using data to assess disease and population risks and (i) in Mexico, IFC brokered investment in Clínicas help plan a range of care and support interventions; Del Azúcar, which manage diabetes and hyperten- (iv) investment in one of the larger groups in Geor- sion for mid- to lower-income populations through a gia, where 85 percent of all health services are pro- growing network of clinics with digital overlay (Box vided by the private sector to the Government’s Na- 16.3); (ii) in South Africa, strong contacts have been tional Health Insurance Fund (Box 16.4); and (v) in developed with Discovery Health/Vitality, which is Spain, IFC has supported an innovative PPP model using data to track NCD risk for insured popula- in which the private sector was contracted to pro- tions, implementing supportive care programs, and vide comprehensive health services to a population, incentivizing the insured to live and eat healthily from primary care through to hospital services. BOX 16.3 Clínicas del Azúcar: a retail approach to diabetes care Bringing disruptive innovation to chronic disease management in Mexico In Mexico, diabetes is an epidemic, and it is the leading cause of death. In the State of Nuevo Leon, where Guadalupe is from, com- plications from diabetes have led to a spike in suicides, in part because the cost of treatment is high and getting care is challenging. Clínicas has pioneered an innovative concept for diabetes care in Mexico. Its one-stop-shop model incorporates a high-quality, multi-disciplinary health team under one roof. Coordinated care for diabetes and hypertension is provided by specialized doc- tors, nurses, nutritionists, and psychologists. Working together, they provide patients with a comprehensive diet and exercise strategy that helps to control blood sugar levels and prevent complications. The model has been successful, but to reach even more patients, Clínicas is integrating behavioral science, big data analytics, artificial intelligence technology, and a mobile app. This will help preserve its low-cost structure and reduce the financial burden for patients. The approach helps patients better manage complex conditions and results in fewer complications that require costly hospital treatment. As of July 2019, Clínicas had established 15 locations in four states, making it the largest private provider of specialized diabetes and hypertension care in Mexico. With a staff of 220 health specialists, of which more than two-thirds are women, it had treated more than 103,000 patients and prevented about 20,000 complications. More than 63 percent of patients had met their goals for maintaining blood sugar levels and had their diabetes under control. This is well above the rate in Mexico’s public sector and is comparable to international norms. Source: Casanova 2019. COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 16 483 BOX 16.4 Georgia Healthcare Group: universal health care delivered through the private sector Making modern health care affordable in a small country For 70 years, the healthcare system in Georgia was funded and delivered through the state. But after the collapse of the Soviet Union in 1991, investment, along with quality of care, declined sharply. The Soviet model of care was hospital-centric, and there was an oversupply of hospitals, beds, and doctors, yet there was a dramatic shortage of nurses. The market was fragmented. Medical staff were not well trained, and there were significant problems with quality and even simple hygiene. As a result, Georgia struggled for decades with high mortality rates. While mortality has improved in recent years, there continues to be a high incidence of hepatitis C and tuberculosis. Meanwhile, cardiovascular disease, cancer, and diabetes cases are rising and adding to the burden on the health system. Georgia Healthcare Group (GHG), a publicly traded company on the London Stock Exchange (LSE), was founded in 2015 as the holding company of the EVEX brand. The first EVEX hospital was opened in 2008. GHG has grown rapidly to become the largest health care provider in Georgia through its fully integrated network of hospitals, clinics, pharmacies, and health insurance. As of June 2018, GHG’s ecosystem covered 75 percent of the geographic areas it served. It operated 37 referral and community hospi- tals with 3,320 beds, accounting for about one-third of the market. Beginning in 2015, it started to ramp up clinics to promote efficient provision of primary care. It currently has 12 clusters of 17 district polyclinics and 24 express outpatient clinics. In 2016, it entered the pharmacy space. Through acquisitions, it has become the largest pharmaceutical retailer and wholesaler in the country with nearly 260 pharmacies. Through its insurance company, Imedi-L, it provides health insurance to 157,000 people, making it the second-largest health insurer in the country. In September 2016, IFC had an opportunity to play a catalytic role in GHG’s growth. With US$25 million of debt funding, IFC supported renovation of two acquired hospitals in Tbilisi and other facilities, as well as the expansion of outpatient services (polyclinics) across the country. The funding also supported introduction of new specialized services, including diagnostics, pediatrics, cardiology, and oncology. These investments promoted access to specialized services, some of which did not previ- ously exist in Georgia. In addition to overhauling facilities, GHG is improving its medical quality by standardizing processes and working on implementation of Joint Commission International (JCI) accreditation requirements to improve patient safety and quality. GHG is the largest employer in Georgia with about 15,500 full time employees, including 3,600 physicians and 3,300 nurses. It is training the next generation of doctors through 24 residency programs and is continually training its doctors, nurs- es, and other medical professionals. In 2017, GHG had GEL 747.8 million (US$298 million) in gross revenues. About 67 percent of gross revenues for the healthcare services segment were derived from universal health coverage (UHC). Source: Casanova 2019. OPPORTUNITIES AND CHALLENGES FOR A WOS APPROACH: WHAT HAVE WE LEARNED? Opportunities a. The COVID-19 pandemic has placed a spotlight b. Another stark fact that emerged from the pan- on the health sector and its weaknesses, showing demic is the need to tackle NCDs better, as how under-funded the sector has been global- almost all the comorbidities that led to hospi- ly. There is a growing realization that the health talizations and deaths were NCDs. This is also sector needs to be strengthened and made more an opportunity to emphasize the inadequate resilient if the world is to face future pandemics— investments that have gone into preventing, and more broadly acute or chronic health emer- controlling, and managing NCDs, for which— gencies from any cause—in a more effective man- over the medium to long term—health systems ner. This might well be an opportune moment, if around the world will pay a price; and used strategically and efficiently, for the health sector to demand and obtain more resources to c. Although WoS approaches are far from easy avoid the serious systemic failures faced by al- to implement, it is evident that, when a prob- most every country during the pandemic. lem affects the whole country/economy, and 484 CHAPTER 16 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E when the causes are multifaceted, there is no quite widely adopted, especially in African coun- alternative but for the society to stand up to tries in the late 1990s to early 2000s, strongly sup- fight them concertedly. Experiences from HIV/ ported by the World Bank. Multi-sector Nutrition AIDS, nutrition, climate change, and most re- Action Plans have also been put in place by several cently COVID-19 could provide appropriate countries around the world. Notwithstanding such platforms for NCD efforts to build on. efforts, getting different sectors to work together to- wards a common goal has been extremely challeng- Challenges ing, and achieving effective coordination among the sectors involved even more so. Experience shows that three factors usually im- The authors’ experience and involvement in pede inter-ministerial cooperation, which is critical some of these initiatives have yielded the following for governments to provide leadership to WoS ap- important lessons: proaches: a. It is important to ensure genuine involvement a. financial rules, which make it difficult for one and active participation of relevant sectors right ministry to share resources with another; from the beginning of the process, rather than a cursory attempt to involve the sectors at a late b. turf battles regarding who is responsible for the stage; multisector involvement should be en- domain in question, who gets the credit and sured in planning, implementation, and mon- resources for it, and who takes the lead. For itoring. Having a well-known and/or highly instance, it is often observed that nutrition is respected champion(s) considerably enhances “everybody’s responsibility but nobody’s busi- the likelihood of success in the WoS approach; ness.” NCDs are not quite in the same category, as they are clearly under the purview of minis- b. Identifying and leveraging a mutuality of inter- tries of health; however, effective NCD action est among societal actors can go a long way in requires coordinated efforts across various fostering cooperation and collaboration, par- ministries; and ticularly since there is no overriding authority that can mandate or enforce coordination. The c. the lack of domain knowledge in relevant min- rapidity with which a COVID-19 vaccine was istries. Ministries of health typically have the brought to the market is a striking example of technical knowledge on how to control NCDs, how mutuality of interest might/should be lev- but the other relevant ministries usually don’t eraged; have experts in that field, rendering efforts on their part suboptimal. c. Identifying clear roles and responsibilities for every sector, including the selection of a spe- Lessons Learnt cific sector as the lead sector, is critical to the success of the program, to avoid the situation The adage “think globally, act locally” can be aptly of “everybody’s business but nobody’s respon- modified for this context as: “think inter-sectoral- sibility”; and ly, act sectorally.” This is, however, easier said than done. Though examples of multisector actions and d. An appropriate institutional home for coordi- intersectoral coordination towards the goal of con- nation—physically as well as organizational- trolling NCDs are few and far between, valuable ly—is a key pre-requisite. Such an institutional lessons do emerge from experience with such ap- home may or may not be in the health sector. proaches toward other health problems, such as HIV/AIDS, malnutrition, early childhood develop- The lead agency or ministry. In most government ment, and disability management projects and pro- bureaucracies, there would be no difficulty in ac- grams. These health issues have long been recog- cepting the health ministry as the lead ministry nized—more so than other health problems—to be for programs with a health objective. However, for multidimensional in causation and impact. As a re- WoS interventions, the requisite coordination au- sult, several countries have experimented with ways thority is usually not vested with the health min- of achieving such coordinated action. MAPs were istry, because generally no single line ministry COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 16 485 would have the authority to convene and coordi- tries and non-state entities is a contentious one, as nate other line ministries. Similarly, funding flows highlighted above. Many governments look at the from one line ministry to another line ministry private sector and civil society with suspicion, and would be next-to-impossible in most bureaucra- vice versa. Although there are several examples of cies. Therefore, a ministry or body above the line public-private partnerships and government-NGO ministries—e.g., Planning and Finance, the Prime partnerships in the health, nutrition, and popula- Minister’s Office, or the President’s Office—might tion sector globally, many of these efforts have faced often be chosen to house the coordination function. serious challenges. Some have managed them suc- Such an approach, however, is double-edged. While cessfully, with others less so. vesting the coordination function in a higher-level In summary, the following challenges to WoS entity addresses the issue of convening power, coor- approaches for NCDs are worth noting: dination committees chaired by the Head of State or Head of Government may well be too high-ranking • Limited data/knowledge on NCDs/risk factors/ and may not meet frequently enough due to com- WoS approaches, hampering decision-making; peting demands on their time. Therefore, unless the problem is seen as a national one—as HIV/AIDS • Absence of political will and national/regional was in several African countries in the 1990s and as platforms for WoS approaches to NCD mitigation COVID-19 is today—and the Head of State/Head in many countries; of Government personally champions the cause of addressing it, coordinated action may remain more • Limited budgetary allocations and competition on paper than in practice. for resources limiting cooperation/collaboration; Coordination costs resources. Usually, a com- mittee is set up for coordination, but unless suffi- • Inadequate awareness among relevant sectors on cient resources – human and financial – are pro- roles and responsibilities; vided, e.g., in the form of an adequately staffed secretariat, these committees cannot function effec- • Limited MoH ability to coordinate both policy tively. There is a difference between committees and design and policy implementation; units/departments to fulfill the coordination/over- sight function. Committees are generally staffed by • Inadequate inclusion of, and interaction between, persons who have other full-time jobs, and there- all three branches of government, as typically even fore can give only partial attention to the task of the efforts to involve the “whole government” tend committee. Usually, therefore, it is better to establish not to include the legislative and judicial branches; a unit or department, rather than ad-hoc commit- tees, to oversee the program. For instance, Nepal set • Weak health systems, which constrain preven- up an autonomous multisectoral entity expressly for tion/management of NCDs; the coordination of its HIV/AIDS response, which included the active participation of civil society and • Limited social inclusion, with widespread geo- affected population groups. graphical, social, gender-based, and financial in- Engaging non-state actors. A special effort is equities in the accessibility, availability, and afford- needed to include non-state actors (i.e., the for-prof- ability of NCD services; and it private sector, non-profit organizations, civil so- ciety, and community-based groups). In addition • Differing motivations and interests among public to turf battles between government ministries, of- and private sectors and difficulties in “balancing” ten the relationship between government minis- health and commercial interests. 486 CHAPTER 16 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E THE WAY FORWARD WoS approaches need to be tailored to the context, because what works in one country or region may not be effective in another. Wide variations in context and scope/scale globally render any generic prescription impos- sible. A critical first step would be to recognize the importance of the economic burden of NCDs, as well as the bidirectional relationships between NCDs and all parts of any society and dealing with NCDs effectively as a global public health threat. Assuming such recognition exists, one can outline general principles and provide broad recommendations on the way forward. We formulate a set of high-level recommendations below. The way forward on how to adopt a WoS approach v. Investment in coordinated and concerted WoS to NCD prevention and control could be presented actions: Resources – human and financial – are as the following ten I’s: key for effective functioning of coordination platforms. To ensure that sufficient resources are i. Involvement of key stakeholders from the available to actualize WoS strategies and plans, outset, leveraging any congruence of inter- strong mobilization efforts will be required from ests, to build an effective alliance for change: both public and private sectors. A definitive way Comprehensive WoS national action plans or to ensure commitment of every sector to the strategies need to be developed at the country WoS approach could be to earmark a part of their level to combat NCDs. Key sectors and actors annual budget towards a common pool dedicat- should be part of planning, implementation, ed to the prevention and control of NCDs. and monitoring of WoS action plans, to en- sure broad-based ownership of the approach vi. Innovation: Constant innovation will be nec- while efficiently advancing shared objectives. essary to ensure the successful and sustained Dissenting groups must be consulted and their implementation of a WoS approach to NCDs, concerns acknowledged, but some selectivity overcoming the challenges outlined in earlier in early participation may be important for sections of this paper. In this regard, different in- success. The presence of one or more champi- novations have been described in the literature.5 ons or stewards can provide a significant impe- tus to WoS efforts. vii. Information sharing: This may proceed by establishing or leveraging institutionalized ii. Inclusion of non-state actors: NCD programs re- knowledge-exchange platforms — such as quire sustained efforts to include non-state actors North-South and South-South cooperation, (for-profits, non-profits, civil society, communi- twinning arrangements, and regional/global ty-based groups) within a mutually agreeable clearing houses — to exchange experiences, policy, operational, and regulatory framework. transfer technologies and skillsets, and mount/ sustain a concerted response against NCDs. iii. Identification of clear roles and responsibil- ities on NCDs for every sector and actor: A viii. Incorporation of lessons learned from global sector must be designated as the lead, to avoid experiences, so as not to “reinvent the wheel,” NCD action’s becoming “everybody’s business to adopt best practices, and to avoid repeating but nobody’s responsibility.” mistakes made by early adopters. iv. Institutionalization of mechanisms for coor- ix. Interventions to be prioritized: Prioritized imple- dination and cooperation: Effective coordina- mentation of interventions to address NCDs and tion requires an institutional home–physically risk factors—as part of an essential benefits pack- and organizationally. Without strong institu- age of health services—should be based on a sys- tional arrangements, WoS approaches would tematic burden of disease assessment, the cost-ef- remain rhetorical. fectiveness of the relevant interventions, and societal value choices to keep the effort tractable. 5   See for example: https://innoverce365.com/types-of-innovation-strategies/ COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 16 487 x. Indicators of success and failure need to be mon- ic interventions needs to be institutionalized, itored continuously: Any WoS effort is likely to with a view to recognizing and resolving im- involve a process of “learning by doing”; this plementation issues. In addition, impact eval- necessitates a strong emphasis on data and mea- uations need to be built into program design to surement for effective decision-making, with make sure that lessons are learned, documented, course corrections as necessary. Operational re- and considered on a continuing basis. search for periodic assessments of programmat- REFERENCES 1. Alleyne, George, and Sania Nishtar. 2013. “Sectoral cooperation 12. Bloom, David E., E.T. Cafiero, E. Jané-Llopis, S. Abrahams-Ges- for the prevention and control of NCDs.” In Noncommunicable sel, L.R., Bloom, S. Fathima, A.B. Feigl, et al. 2011. “The Global Diseases in the Developing World: Addressing Gaps in Global Policy Economic Burden of Non-Communicable Diseases.” Geneva, and Research. Edited  by Galambos, Louis, Jeffrey L. Sturchio, and Switzerland: World Economic Forum. https://world-heart-fed- Rachel Whitehead. Baltimore: Johns Hopkins University Press.  eration.org/wp-content/uploads/2017/05/WEF_Harvard_HE_ 2. The Lancet. 2020. “COVID-19: A New Lens for Non-Communi- GlobalEconomicBurdenNonCommunicableDiseases_2011.pdf. cable Diseases.” Editorial. The Lancet 396 (10252): 649. https:// 13. Bulc, Barbara, and Rohit Ramchandani. 2021. “Accelerating doi.org/10.1016/s0140-6736(20)31856-0. Global Health R&D: The Role of Product Development Part- 3. Agrawal, Utkarsh, Amaya Azcoaga-Lorenzo, Adeniyi Francis nerships.” Geneva: Switzerland: Global Development/Water- Fagbamigbe, Eleftheria Vasileiou, Paul Henery, Colin R Simpson, loo, Canada: Antara Global Health Advisors. https://www. Sarah J Stock, et al. 2021. “Association between Multimorbidity mmv.org/sites/default/files/uploads/docs/publications/Ac- and Mortality in a Cohort of Patients Admitted to Hospital with celerating-global-health-RD_The-role-of-product-develop- COVID-19 in Scotland.” Journal of the Royal Society of Medicine ment-partnerships_BBulc_RRamchandani_March2021.pdf. 115 (1): 22–30. https://doi.org/10.1177/01410768211051715. 14. Carroll, Addy, Lisa Wood, and Sripen Tantivess. 2007. “Many 4. Alexander, Eleanore, Derek Yach, and George A Mensah. 2011. Things to Many People: A Review of ThaiHealth.” Bangkok, Thai- “Major Multinational Food and Beverage Companies and Infor- land: Thai Health Promotion Foundation. mal Sector Contributions to Global Food Consumption: Impli- 15. Casanova, Ann M. 2019. “Case Study: Clinicas Del Azucar: Bringing cations for Nutrition Policy.” Globalization and Health 7 (1): 26. Disruptive Innovation to Chronic Disease Management in Mexico https://doi.org/10.1186/1744-8603-7-26. a Retail Approach to Diabetes Care.” Washington, D.C.: Interna- 5. Allen, Luke N, Brian D Nicholson, Beatrice Y T Yeung, and Fran- tional Finance Corporation. https://www.ifc.org/content/dam/ cisco Goiana-da-Silva. 2020. “Implementation of Non-Com- ifc/doc/mgrt/201906-a-retail-approach-to-diabetes-care.pdf. municable Disease Policies: A Geopolitical Analysis of 151 16. Chan, Juliana C. N., Lee-Ling Lim, Nicholas J. Wareham, Jon- Countries.” The Lancet Global Health 8 (1): e50–58. https://doi. athan E. Shaw, Trevor J. Orchard, Ping Zhang, Eric S. H. Lau, org/10.1016/s2214-109x(19)30446-2. et al. 2020. “The Lancet Commission on Diabetes: Using 6. Allen, Luke, and Ashley Bloomfield. 2016. “Engaging the Private Data to Transform Diabetes Care and Patient Lives.” The Lan- Sector to Strengthen NCD Prevention and Control.” The Lancet cet 396 (10267): 2019–82. https://doi.org/10.1016/S0140- Global Health 4 (12): e897–98. https://doi.org/10.1016/s2214- 6736(20)32374-6. 109x(16)30216-9. 17. Chen, Simiao, Michael Kuhn, Klaus Prettner, and David E. Bloom. 7. Anderson, Gerard F. 1992. “The Courts and Health Policy: 2018. “The Macroeconomic Burden of Noncommunicable Dis- Strengths and Limitations.” Health Affairs 11 (4): 95–110. eases in the United States: Estimates and Projections.” Edited by https://doi.org/10.1377/hlthaff.11.4.95. Muhammad Jami Husain. PLOS ONE 13 (11): e0206702. https:// 8. Andreyeva, Tatiana, Keith Marple, Samantha Marinello, Timo- doi.org/10.1371/journal.pone.0206702. thy E. Moore, and Lisa M. Powell. 2022. “Outcomes Following 18. Collins, Téa, Bente Mikkelsen, and Svetlana Axelrod. 2019. “In- Taxation of Sugar-Sweetened Beverages.” JAMA Network Open teract, Engage or Partner? Working with the Private Sector for 5 (6): e2215276. https://doi.org/10.1001/jamanetworko- the Prevention and Control of Noncommunicable Diseases.” pen.2022.15276. Cardiovascular Diagnosis and Therapy 9 (2): 158–64. https://doi. 9. Andreyeva, Tatiana, Keith Marple, Timothy E. Moore, and Lisa org/10.21037/cdt.2018.08.04. M. Powell. 2022. “Evaluation of Economic and Health Out- 19. Cox, Archibald. 1976. “The Role of the Supreme Court in Ameri- comes Associated with Food Taxes and Subsidies.” JAMA Net- can Government.” Harvard Law Review 103: 100–141. work Open 5 (6): e2214371. https://doi.org/10.1001/jamanet- 20. Cuadrado, Cristóbal, Alain Palacios, Tim Miller, and Branka workopen.2022.14371. Legetic. 2015. “Indicators for Monitoring the Socio-Econom- 10. Barton, Hugh, and Marcus Grant. 2012. “Urban Planning for ic Dimension of NCDs: A Pilot Study in Chile.” Working Paper. Healthy Cities.” Journal of Urban Health 90 (S1): 129–41. https:// Santiago: Ministry of Health, Chile, Pan-American Health Or- doi.org/10.1007/s11524-011-9649-3. ganization, and the Economic Commission for Latin America 11. Bee, Yong Mong, E. Shyong Tai, and Tien Y. Wong. 2022. “Singa- and the Caribbean . https://www.dcp-3.org/sites/default/files/ pore’s ‘War on Diabetes.’” The Lancet Diabetes & Endocrinology 10 resources/Indicators%20for%20Monitoring.pdf. (6): 391–92. https://doi.org/10.1016/S2213-8587(22)00133-4. 21. Dain, Katie. 2019. “A ‘Whole of Society’ Approach to Non-Com- 488 CHAPTER 16 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E municable Diseases Must Include Civil Society Organisa- Health Systems Strengthening Project - P167581 - Sequence No : tions.” The BMJ Opinion (blog). December 6, 2019. https:// 02 (English). Washington, D.C. : World Bank Group.  http://doc- blogs.bmj.com/bmj/2019/12/06/a-whole-of-society-ap- uments.worldbank.org/curated/en/754591576131288997/ proach-to-non-communicable-diseases-must-include-civ- Disclosable-Version-of-the-ISR-Andhra-Pradesh-Health-Sys- il-society-organisations/. tems-Strengthening-Project-P167581-Sequence-No-02 22. DeVon, Holli A., Catherine J. Ryan, Amy L. Ochs, and Moshe 35. Kankeu, Hyacinthe Tchewonpi, Priyanka Saksena, Ke Xu, and Shapiro. 2008. “Symptoms across the Continuum of Acute David B Evans. 2013. “The Financial Burden from Non-Com- Coronary Syndromes: Differences between Women and Men.” municable Diseases in Low- and Middle-Income Countries: American Journal of Critical Care: An Official Publication, Amer- A Literature Review.” Health Research Policy and Systems 11 (1). ican Association of Critical-Care Nurses 17 (1): 14–24; quiz 25. https://doi.org/10.1186/1478-4505-11-31. https://pubmed.ncbi.nlm.nih.gov/18158385/. 36. Kassa, Melkamu Dugassa, and Jeanne Martin Grace. 2022. “Non- 23. Ettman, Catherine K., Gregory H. Cohen, Salma M. Abdalla, communicable Diseases Prevention Policies and Their Imple- Laura Sampson, Ludovic Trinquart, Brian C. Castrucci, Rachel mentation in Africa: A Systematic Review.” Public Health Reviews H. Bork, et al. 2021. “Persistent Depressive Symptoms during 42 (February). https://doi.org/10.3389/phrs.2021.1604310. COVID-19: A National, Population-Representative, Longitudi- 37. Kerr, John, Costas Panagopoulos, and Sander van der Linden. nal Study of U.S. Adults.” The Lancet Regional Health - Americas 5 2021. “Political Polarization on COVID-19 Pandemic Response (100091): 100091. https://doi.org/10.1016/j.lana.2021.100091. in the United States.” Personality and Individual Differences 179 24. Galambos, Louis, Jeffrey L. Sturchio, and Rachel C. Whitehead, (179): 110892. https://doi.org/10.1016/j.paid.2021.110892. eds. 2013. Noncommunicable Diseases in the Developing World: 38. Kwon, Sung Hee, Jun-Pyo Myong, Hyoung-Ah Kim, and Kyeong Addressing Gaps in Global Policy and Research. Www.press.jhu. Yeon Kim. 2020. “Association between Morbidity of Non-Com- edu. Baltimore: Johns Hopkins University Press. https://www. municable Disease and Employment Status: A Comparison be- press.jhu.edu/books/title/11061/noncommunicable-diseas- tween Korea and the United States.” BMC Public Health 20 (1). es-developing-world. https://doi.org/10.1186/s12889-020-08883-3. 25. Government of Samoa. 2019. “Samoa National Tobacco Con- 39. Laryea, Dennis O., and Thomas B. Cueni. 2019. “Including the trol Policy and Plan of Action July 2019-June 2024.” Ministry of Private Sector in Partnerships to Tackle Non-Communicable Health, Government of Samoa. https://www.health.gov.ws/ Diseases.” The BMJ Opinion (blog). December 6, 2019. https:// wp-content/uploads/2023/01/Samoa-National-Tobacco-Con- blogs.bmj.com/bmj/2019/12/06/including-the-private-sec- trol-Policy-and-Plan-of-Action-Revised-11.11.2019.pdf. tor-in-partnerships-to-tackle-non-communicable-diseases/. 26. ———. 2023. Excise Tax Rate Amendment Act 2023, No. 5. 40. Lencucha, Raphael, Jeffrey Drope, and Jenina Joy Chavez. 2014. https://www.palemene.ws/PDFfiles/Excise%20Tax%20Rate%20 “Whole-of-Government Approaches to NCDs: The Case of the Amendment%20Act%202023%20-%20Eng_English.pdf. Philippines Interagency Committee—Tobacco: Table 1.” Health 27. Hancock, Christine, Lise Kingo, and Olivier Raynaud. 2011. “The Policy and Planning 30 (7): 844–52. https://doi.org/10.1093/ Private Sector, International Development and NCDs.” Global- heapol/czu085. ization and Health 7 (1): 23. https://doi.org/10.1186/1744- 41. Leppo, Kimmo, Eeva Ollila, Sebastián Peña, Matthias Wismar, 8603-7-23. and Sarah Cook. 2013. “Health in All Policies: Seizing Oppor- 28. Institute of Health Metrics and Evaluation (IHME). 2019. Global tunities, Implementing Policies.” Ministry of Social Affairs and Burden of Disease 2019. Database. https://vizhub.healthdata. Health, Finland. https://www.uab.edu/medicine/obesity/im- org/gbd-results/ ages/Health-in-All-Policies-final.pdf. 29. ———. 2020. “Disease Burden from Non-Communicable 42. Mahat, Kishori, and Badri Thapa. 2019. “Multi-Sectoral Ap- Diseases by Age.” Global Burden of Disease 2019. Our World proach to Non-Communicable Diseases Control: Easier Said in Data. https://ourworldindata.org/grapher/disease-bur- than Done.” Journal of Nepal Medical Association 57 (220). den-from-ncds-by-age?time=earliest..2019&facet=none. https://doi.org/10.31729/jnma.4729. 30. International Alliance for Responsible Drinking (IARD). 2019. 43. Maldonado, Norman, Blanca Llorente, Luz Myriam Rey- “Focus on Public–Private Partnerships: Case Studies.” https:// nales-Shigematsu, Belen Saenz-de-Miera, Prabhat Jha, and www.iard.org/science-resources/detail/Focus-on-Pub- Geordan Shannon. 2022. “Tobacco Taxes as the Unsung Hero: lic%E2%80%93Private-Partnerships-five-case-stu. Impact of a Tax Increase on Advancing Sustainable Develop- 31. Juma, Pamela A., Clarisse Mapa-tassou, Shukri F. Mohamed, Be- ment in Colombia.” International Journal of Public Health 67 atrice L. Matanje Mwagomba, Catherine Ndinda, Mojisola Olu- (March). https://doi.org/10.3389/ijph.2022.1604353. wasanu, Jean-Claude Mbanya, Misheck J. Nkhata, Gershim Asi- 44. Merrick, Tom and colleagues. 2012.  “What Can We Learn from ki, and Catherine Kyobutungi. 2018. “Multi-Sectoral Action in 22 years and US$ 16 Billion of Lending for Health Through Non- Non-Communicable Disease Prevention Policy Development Health Projects and Can We Do Better?” Presentation. Unpub- in Five African Countries.” BMC Public Health 18 (S1). https://doi. lished. World Bank. org/10.1186/s12889-018-5826-6. 45. Michael, Marmot, Jessica Allen, Peter Goldblatt, Eleanor Herd, 32. Kabayama, Mai , Kei Kamide, Kazue Sakakibara, and Kazuo Ha- and Joana Morrison. 2021. “Build Back Fairer: The COVID-19 yakawa. 2014. “The Role of Public Health Nurses in Japanese Marmot Review. The Pandemic, Socioeconomic and Health Long-Term Care Prevention Projects in the Community.” Jour- Inequalities in England.” Institute of Health Equity: London. nal of Nursing & Care 03 (03). https://doi.org/10.4172/2167- https://www.instituteofhealthequity.org/resources-reports/ 1168.1000166. build-back-fairer-the-covid-19-marmot-review. 33. Kagan, Robert A. 1991. “Adversarial Legalism and American 46. Murray, Christopher J L, Aleksandr Y Aravkin, Peng Zheng, Government.” Journal of Policy Analysis and Management 10 Cristiana Abbafati, Kaja M Abbas, Mohsen Abbasi-Kangevari, (3): 369–406. Foad Abd-Allah, et al. 2020. “Global Burden of 87 Risk Factors 34. Kak, Mohini. 2019.  Disclosable Version of the ISR - Andhra Pradesh in 204 Countries and Territories, 1990–2019: A Systematic COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 16 489 Analysis for the Global Burden of Disease Study 2019.” The and Enlightenment].” Zhonghua Yu Fang Yi Xue Za Zhi [Chinese Lancet 396 (10258): 1223–49. https://doi.org/10.1016/s0140- Journal of Preventive Medicine] 53 (12): 1198–1202. https://doi. 6736(20)30752-2. org/10.3760/cma.j.issn.0253-9624.2019.12.002. 47. NCD Alliance. 2020a. “Bridging the Gap on NCDs: From Global 60. Thakur, Jarnail Singh, Ria Nangia, and Sukriti Singh. 2021. “Prog- Promises to Local Progress.” Geneva: Switzerland: NCD Alliance. ress and Challenges in Achieving Noncommunicable Diseases https://ncdalliance.org/resources/bridging-the-gap-on-ncds- Targets for the Sustainable Development Goals.” FASEB BioAd- from-global-promises-to-local-progress-policy-brief. vances 3 (8). https://doi.org/10.1096/fba.2020-00117. 48. ———. 2020b. “New NCD Countdown 2030 Report Shows 61. The Lancet. 2020. “Non-Communicable Diseases—Level 1 Slow Progress towards UN SDG Target 3.4.” NCD Alliance (blog). Cause. Global Health Metrics. The Lancet 396 (10258): S42–43. September 4, 2020. https://ncdalliance.org/news-events/ 62. The National Academies of Medicine. 2022. Global Roadmap news/new-ncd-countdown-2030-report-shows-slow-prog- for Healthy Longevity. Washington, D.C.: The National Acade- ress-towards-un-sdg-target-34. mies Press. https://doi.org/10.17226/26144. 49. Office of National Statistics (ONS). 2020.  “Coronavirus (COVID-19) 63. Tong, Matai’a Lanuola Tusani T - Ah . 2023. “Bill on Tobacco Ex- Related Deaths by Disability Status, England, and Wales, 2020.” cise Scrutinised in House.” Samoa Observer, July 4, 2023. https:// Government of the United Kingdom. www.ons.gov.uk. www.samoaobserver.ws/category/samoa/104404. 50. Pan-American Health Organization (PAHO). 2011. “Non-Com- 64. Unwin, Nigel, T. Alafia Samuels, Trevor Hassell, Ross C. Brownson, municable Diseases and Gender.” Fact Sheet. Washington, and Cornelia Guell. 2016. “The Development of Public Policies to D.C.: Pan-American Health Organization/International Di- Address Non-Communicable Diseases in the Caribbean Coun- abetes Foundation. https://www3.paho.org/hq/dmdocu- try of Barbados: The Importance of Problem Framing and Policy ments/2011/gdr-ncd-gender-factsheet-final.pdf. Entrepreneurs.” International Journal of Health Policy and Manage- 51. Parry, Ian W. H., Sarah E West, and Ramanan Laxminarayan. ment 6 (2): 71–82. https://doi.org/10.15171/ijhpm.2016.74. 2009. “Fiscal and Externality Rationales for Alcohol Policies.” 65. Williamson, Elizabeth J., Alex J. Walker, Krishnan Bhaskaran, Seb The B.E. Journal of Economic Analysis & Policy 9 (1). https://doi. Bacon, Chris Bates, Caroline E. Morton, Helen J. Curtis, et al. org/10.2202/1935-1682.2133. 2020. “OpenSAFELY: Factors Associated with COVID-19 Death 52. Puska, Pekka, and Paresh Jaini. 2020. “The North Karelia Proj- in 17 Million Patients.” Nature 584 (July): 430–36. https://doi. ect: Prevention of Cardiovascular Disease in Finland through org/10.1038/s41586-020-2521-4. Population-Based Lifestyle Interventions.” American Journal 66. World Bank. 2015. Health Systems Strengthening and Support of Lifestyle Medicine 14 (5): 155982762091098. https://doi. Project (P152799). IBRD Project Appraisal Document (August 28, org/10.1177/1559827620910981. 2015). Washington, D.C.: World Bank. https://projects.worldbank. 53. Sagapolutele, Fili. 2018. “Lolo Signs Three Admin Bills into Law org/en/projects-operations/project-detail/P152799. — Cigarette Tax, Elderly/ Disabled Abuse Act, and Consumer 67. ———. 2018a. Nicaragua: Integrated Public Provision of Protection.” Samoa News, October 24, 2018. https://www.sa- Health Care Services (P164452). IDA Restructuring Paper moanews.com/local-news/lolo-signs-three-admin-bills-law- (March 15, 2018). Washington, D.C.: World Bank. https://doc- cigarette-tax-elderly-disabled-abuse-act-and-consumer. uments1.worldbank.org/curated/en/551171574707407442/ 54. Shibuya, Kenji, Hideki Hashimoto, Naoki Ikegami, Akihiro Ni- pdf/Disclosable-Restructuring-Paper-NI-INTEGRATED-PUB- shi, Tetsuya Tanimoto, Hiroaki Miyata, Keizo Takemi, and Mi- LIC-PROVISION-OF-HEALTH-CARE-SERVICES-P164452.pdf. chael R Reich. 2011. “Future of Japan’s System of Good Health 68. ———. 2018b. Anhui Aged Care System Demonstration Proj- at Low Cost with Equity: Beyond Universal Coverage.” The ect (P154716). IBRD Project Appraisal Document (May 29, 2018). Lancet 378 (9798): 1265–73. https://doi.org/10.1016/s0140- Washington, D.C.: World Bank. https://documents1.worldbank. 6736(11)61098-2. org/curated/en/357051529638290655/pdf/Project-Apprais- 55. Steenkamp, Lizelle, Robin Terence Saggers, Rossella Bandi- al-Document-PAD-disclosable-version-May-30-SECPO-com- ni, Saverio Stranges, Yun-Hee Choi, Jane S. Thornton, Simon ments-05312018.pdf. Hendrie, Deepak Patel, Shannon Rabinowitz, and Jon Patricios. 69. ———. 2018c. Sri Lanka: Primary Health Care System Strength- 2022. “Small Steps, Strong Shield: Directly Measured, Moder- ening Project (P163721). Restructuring Paper (June 27, 2018). ate Physical Activity in 65 361 Adults Is Associated with Sig- Washington, D.C.: World Bank. https://projects.worldbank.org/ nificant Protective Effects from Severe COVID-19 Outcomes.” en/projects-operations/project-detail/P163721. British Journal of Sports Medicine 56 (10): 568–76. https://doi. 70. ———. 2018d. Health System Strengthening Project org/10.1136/bjsports-2021-105159. (P166783). IDA Project Appraisal Document (August 14, 2018). 56. Sturchio, Jeffrey L., and Akash Goel. 2012. “The Private-Sector Washington, D.C.: World Bank. https://projects.worldbank.org/ Role in Public Health.” Washington, D.C.: Center for Strategic en/projects-operations/project-detail/P166783. and International Studies. https://www.csis.org/analysis/pri- 71. ———. 2018e. El Salvador: Addressing Non-Communica- vate-sector-role-public-health. ble Diseases Project (P164356). IBRD Restructuring Paper 57. Sturchio, Jeffrey L., Ilona Kickbusch, and Louis Galambos, eds. (September 13, 2018). Washington, D.C.: World Bank. https:// 2019. The Road to Universal Health Coverage. Baltimore: Johns documents.worldbank.org/en/publication/documents-re- Hopkins University Press. https://www.press.jhu.edu/books/ ports/documentdetail/822611584483599152/disclosable-re- title/12246/road-universal-health-coverage. structuring-and-or-additional-financing-paper-esv-address- 58. Sturm, Roland, and Derek Yach. 2013. “Eating Better for Less. To ing-non-communicable-diseases-p164356. Treat the Obesity Epidemic, Why Not Lower Prices on Healthier 72. ——— 2019a. Tamil Nadu Health Systems Reform Program. Food.” Modern Healthcare 43 (21): 27. https://pubmed.ncbi.nlm. IBRD Project Appraisal Document (February 22, 2019). Wash- nih.gov/23947103/. ington, D.C.: World Bank. https://projects.worldbank.org/en/ 59. Su, X. Y., M. Y. Si, Z. K. Zhu, Y. Jiang, and Y. L. Liu. 2019. “[the projects-operations/project-detail/P166373. Healthy Living Master Plan (HLMP): Singapore’s Experience 73. ———. 2019b. Andhra Pradesh Health Systems Strengthen- 490 CHAPTER 16 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E ing Project (P167581). IBRD Project Appraisal Document (April en, and the Tobacco Epidemic.” Edited by Jonathan M. Sa- 9, 2019). Washington, D.C.: World Bank. https://projects.world- mer and Soon-Young Yoon. Geneva, Switzerland: World bank.org/en/projects-operations/project-detail/P167581. Health Organization. https://iris.who.int/bitstream/han- 74. ———. 2019c. Yemen Emergency Health and Nutrition Proj- dle/10665/44342/9789241599511_eng.pdf?sequence=1. ect (P167195). IDA Project Paper (April 16, 2019) Washington, 82. ———. 2020a. “The Impact of the COVID-19 Pandemic D.C.: World Bank. https://projects.worldbank.org/en/proj- on Noncommunicable Disease Resources and Services: Re- ects-operations/project-detail/P167195. sults of a Rapid Assessment.” Geneva, Switzerland: World 75. ———. 2019d. Hashemite Kingdom of Jordan: Emergency Health Organization. https://www.who.int/publications/i/ Health Project Additional Financing (P170529). IBRD Project item/9789240010291. Paper (June 7, 2019) Washington, D.C.: World Bank. https:// 83. ———. 2020b. “World Health Statistics 2020 Visual Summary.” projects.worldbank.org/en/projects-operations/project-de- World Health Organization. 2020. https://www.who.int/data/ tail/P170529. gho/whs-2020-visual-summary. 76. ———. 2019e. Samoa Health System Strengthening Program 84. ———. 2021. “Noncommunicable Diseases.” Fact Sheet. Gene- (P164382). IDA Project Appraisal Document (November 19, va: Switzerland: World Health Organization. 2019). Washington, D.C.: World Bank. https://projects.world- 85. Wu, Fei, Hiroto Narimatsu, Xiaoqiang Li, Sho Nakamura, Ri bank.org/pt/projects-operations/project-detail/P164382. Sho, Genming Zhao, Yoshinori Nakata, and Wanghong Xu. 77. ———. 2019f. Tajikistan Health Services Improvement Proj- 2017. “Non-Communicable Diseases Control in China and Ja- ect. Second Additional Financing (P170358). IDA Project Paper pan.” Globalization and Health 13 (1). https://doi.org/10.1186/ (November 25, 2019). https://projects.worldbank.org/en/proj- s12992-017-0315-8. ects-operations/project-detail/P170358. 86. Yach, Derek. 2022. “Addressing the Impact of Multiple Chronic 78. ———. 2020a. Growing Up Healthy Together: Compre- Conditions Can Improve Health and Reduce Morbidity.” ReSolve hensive Early Childhood Development in El Salvador Project Global Health (blog). July 4, 2022. https://www.re-solveglobal- (P169677). IBRD Project Appraisal Document (February 27, health.com/post/addressing-the-impact-of-multiple-chron- 2020). Washington, D.C.: World Bank. https://documents. ic-conditions-can-improve-health-and-reduce-morbidity. worldbank.org/en/publication/documents-reports/doc- 87. Yach, Derek, Zoë A. Feldman, Dondeena G. Bradley, and umentdetail/136621583358514932/el-salvador-grow- Mehmood Khan. 2010. “Can the Food Industry Help Tackle the ing-up-healthy-together-comprehensive-early-childhood-de- Growing Global Burden of Undernutrition?” American Jour- velopment-in-el-salvador-project. nal of Public Health 100 (6): 974–80. https://doi.org/10.2105/ 79. ———. 2020b. “Europe and Central Asia Economic Update, ajph.2009.174359. Fall 2020: COVID-19 and Human Capital.” Washington, D.C.: 88. Zuidberg, Mark R J, Amanda Shriwise, Lisanne M de Boer, and World Bank. http://hdl.handle.net/10986/34518. Anne S Johansen. 2020. “Assessing Progress under Health 2020 80. ———. 2022. “World Development Report 2022: Finance for in the European Region of the World Health Organization.” an Equitable Recovery.” Washington, D.C.: World Bank. https:// European Journal of Public Health 30 (6): 1072–77. https://doi. www.worldbank.org/en/publication/wdr2022. org/10.1093/eurpub/ckaa091. 81. World Health Organization (WHO). 2010. “Gender, Wom- COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E C H A P T E R 17 491 Monitoring Indicators for Healthy Longevity Chapter 17 Healthy Longevity Initiative: A Performance Dashboard for Decision-Making in Low- and Middle- Income Countries Chapter 18 Assessing Human Capital, Non-communicable Diseases, and Healthy Longevity in Low- and Middle-Income Countries: Healthy Longevity Dashboard and the Case for Indian 17 492 C H A P T E R 17 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E Healthy Longevity Initiative A performance dashboard for decision-making in low- and middle- income countries Victoria Haldane a, Gisela M. Garcia b, Tahir Bockarie c, Daphne Wu d, Cristian A. Herrera b,e, Maria Luisa Latorre Castro f, Debapriya Chakraborty d, Beverly Essue a, Prabhat Jha a,d, Jeremy Veillard a,b a Institute of Health Policy, Management and Evaluation, University of Toronto b Health, Nutrition, and Population, Latin America and the Caribbean, World Bank c Warwick Medical School, University of Warwick d Centre for Global Health Research, Unity Health Toronto, University of Toronto e Department of Public Health, School of Medicine, Pontificia Universidad Católica de Chile f Fundación Universitaria Juan N Corpas COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E C H A P T E R 17 493 INTRODUCTION As the global population rapidly ages, there is a pressing need for approaches to promote health, well-being, and productivity across the life course (Beard and Bloom 2015; Bloom and Sousa-Poza 2013). The Healthy Longev- ity Initiative (HLI) sets an agenda that emphasizes how inaction towards healthy longevity will have profound and far-reaching impacts on human capital and productivity. The HLI takes a life course approach to consid- ering the linkages between human capital, non-communicable diseases (NCDs), healthy longevity, economic, and well-being outcomes. Increasing prevalence of NCDs, particularly amongst older adults, threatens both lives and livelihoods (GBD 2019 Ageing Collaborators 2022). Given changing demographics, urbanization, and increased exposure to complex risk factors, older adults living in low- and middle-income countries (LMICs) bear a growing and disproportionate burden of morbidity and mortality from NCDs (WHO 2014, Mathers and Loncar 2006). Yet, NCDs remain low on the policy agenda in many LMICs despite the accelerating threat they pose not only to individual health and well-being, but to human capital, productivity, and future growth more broadly (Miranda et al. 2008; Chaker el al. 2015; Allen et al. 2020). Advancing a healthy longevity agenda offers an op- and Sen 2014; Holzer and Yang 2004). Countries portunity to promote and monitor progress towards need a multi-sectoral measurement approach based human capital accumulation and preservation, on information that enables them to determine health, and well-being not only in younger age but their progress towards avoiding and adequate man- for current generations both young and old (World agement of NCDs, as well as other complementa- Bank 2020). The concept of healthy longevity de- ry goals that promote human capital accumulation scribes a process of ageing that promotes human and preservation across the life course. The Human capital across the life course and where NCDs, seri- Capital Index (HCI) is one example of how a mea- ous disability, and deaths are avoided in middle age, surement approach can be effective to motivate na- thereby compressing morbidity in older ages. This al- tional stakeholders to invest in people for greater lows older adults to maintain functional abilities and inclusive economic growth (Corral et al. 2021; Stein mental health, as well as social connectedness and and Sridhar 2019; World Bank 2020). productivity, into their eldest ages. Healthy longevity To advance a healthy longevity agenda globally, extends working lives and creates a more productive countries must bring together relevant and action- workforce through several interrelated channels that able data into a set of user-friendly resources that promote active and productive ageing and avoidance provide them with key information on health and of NCDs, underpinned by improvements in health human capital from across sectors to monitor per- and lifestyle (O’Keefe and Haldane 2024). From a na- formance, compare progress, guide investments and tional and global policy perspective, avoiding NCDs interventions, and drive strategy and planning. A is crucial to healthy longevity given the multifacto- common framework and harmonized benchmark- rial and interconnected ways these diseases impact ing approach for countries to develop a healthy lon- inclusive growth, including impacts on savings, fiscal gevity dashboard is urgently needed. effects, as well as accumulation or depreciation of hu- Here, we offer such a common framework and man capital (O’Keefe and Haldane 2024). harmonized approach to measurement of healthy Building and operationalizing such an agenda, longevity. We first contextualize and describe key however, requires strategic action and focused in- considerations for a healthy longevity dashboard. vestments, guided by local evidence, with progress We then present criteria for indicators to guide monitored over time. Measuring the effectiveness countries in making focused investments in NCDs of policies and investments and monitoring overall and healthy longevity across the life course more progress towards the goal of healthy longevity gives broadly. Three country case studies are presented visibility to key issues, enables benchmarking, and that illustrate a spectrum of dashboard possibilities offers important information to guide agenda set- considering differing data and information system ting, prioritization of resources and planning across contexts. Finally, we offer lessons learned through levels of action from the global to the local (Ba- the dashboard creation process to assist others in ranek, Veillard, and Wright 2011; Hulme, Savoia, creating national healthy longevity dashboards. 494 C H A P T E R 17 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CONTEXTUALIZING DASHBOARDS FOR PERFORMANCE MONITORING AND STRATEGIC ACTION Performance dashboards, hereafter ‘dashboards,’ are data visualization tools that bring together representations of indicators into a unified interface that allows stakeholders to monitor and assess performance and progress across selected dimensions (Helminski et al. 2022; Eckerson 2010). These dashboards are commonly used to monitor prog- ress and show impact in the private sector, where they are an integral part of evidence-based business performance measurement (Pfeffer and Sutton 2006). Performance dashboards for business are linked to key performance indi- cators and organizational targets. An integral part of business intelligence, dashboards support agile program de- livery linking current states to future growth as a part of ‘learning systems’ (Kaplan and Norton 1996). Increasingly, healthcare delivery organizations and health systems more broadly are adopting business intelligence and learning systems approaches to proactively measure, monitor, compare, and act based on metrics informed by ever evolving electronic medical record data and other health data sources (Veillard et al. 2010; Bonney 2013; OECD 2002). Frameworks and indicators have widely been used to statistics, demographic, and census data, as well as catalyze action around and monitor progress towards outcomes data across a range of diseases. By doing global goals. Notably, the Millennium Development so users are motivated to act, course correct, com- Goals (MDGs) presented a set of targets grounded pare progress against others, and make strategic in- in a common vision to reduce global poverty. The vestments to address global health challenges. MDGs saw the advent of a common framework for For example, during the COVID-19 pandemic promoting global development guided by indicators numerous dashboards have emerged that provide a to monitor progress (McArthur 2014). With a com- picture of the epidemiological situation and response mon goal and indicators to rally around, the MDGs effectiveness across an array of indicators (e.g., num- spurred global advancements in education, econom- ber of deaths, cases, hospitalization, test positiv- ic growth, and poverty reduction, with some of the ity rates, amongst others) – (Ivankovic et al. 2021). largest gains made in health, particularly in LMICs Many COVID-19 dashboards quickly adapted to (Narh-Bana et al. 2022; McArthur and Rasmussen include information on COVID-19 vaccine roll out. 2018). It’s estimated that the accelerated develop- These provided crucial information to inform local ment progress catalyzed by the MDGs saved an extra vaccination programs and prioritization strategies, as 20.9 to 30.3 million lives, approximately two-thirds well as monitor global vaccine equity (UNDP 2022). of these in Africa (McArthur and Rasmussen 2018). Another example of a dashboard linked to more ex- From national initiatives to reduce child mortality, in- plicit goals to end a global health threat is the global crease childhood education, and access to antiretro- tuberculosis program dashboard. Relying on data virals, to global partnerships such as the Global Fund collected by national tuberculosis programs across to Fight AIDS, Tuberculosis and Malaria, the MDGs a standard set of indicators, the dashboard presents spurred partnerships and initiatives that committed a comprehensive country- and regional-level dash- to measurement as the path towards greater health board to monitor progress towards the End TB Strat- and well-being globally (McArthur 2013). egy milestones (WHO 2004, 2022a). Similar dash- Building on this commitment to measurement, boards to track progress towards global health goals indicators are increasingly being aggregated into are available for malaria, polio, HIV, as well as wom- dashboards to readily and in near real time moni- en’s, children’s and adolescent health (RBM Partner- tor progress towards goals that advance health and ship to End Malaria 2022; WHO 2022b; AHEAD well-being (Annex 17.1). Dashboards leverage vital 2022; UNICEF 2022a, 2022b). A COMMON FRAMEWORK FOR HEALTHY LONGEVITY MEASUREMENT The healthy longevity dashboard is an ongoing effort to develop and refine a suite of indicators that bring together relevant data to measure and monitor country progress towards healthy longevity goals. Monitoring progress towards health and well-being in older age is not a novel concept. Many countries have already committed to making progress on indicators relevant to healthy longevity such as life expectancy, NCD risk reduction, adult health, and gender equality. COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E C H A P T E R 17 495 The common framework for a healthy longevity ceptual grounding, which ensures that the life course dashboard described here allows for framing a se- approach proposed guides the choice and monitoring lected set of these indicators in a different way. It em- of key indicators. Proposed indicators map to the in- phasizes their role in promoting health and human tensive accumulation of human capital during early capital across the life course, while drawing atten- years, onwards to the period of further human capital tion to the multi-sectoral efforts needed to promote accumulation, deployment, and depreciation during healthy longevity, inclusive growth, and well-being. advanced schooling and working life, as well as the Importantly, there is no one-size-fits-all dash- declining human capital accumulation and accelerat- board for healthy longevity. Rather we propose a ed human capital depreciation of older adults (Figure common framework, grounded in a conceptual 17.1) – (O’Keefe and Haldane 2024). The framework framework, and complemented by a minimum set offers inroads for healthy longevity interventions and of indicators that countries can leverage to define policies that increase human capital accumulation, their own dashboard. Dashboards should be based minimize human capital depreciation, and impor- on key indicators and meaningful disaggregations tantly, preserve human capital in older ages to ward that are contextually relevant and supported by off human capital depletion as long as possible. Link- available data sources. These key indicators enable ing indicators to life course considerations provides a a harmonized benchmarking approach, so countries holistic and comprehensive view of population health. are well-placed to compare their progress against Given this, when developing a healthy longev- others with a similar level of development (e.g., sim- ity dashboard, three key actions and ten related do- ilar World Bank classifications) or other comparable mains must be prioritized, which map to the con- metrics (e.g., OECD countries). ceptual framework (Table 17.1). First, supporting The proposed common framework for a healthy enabling factors for human capital accumulation longevity dashboard takes a life course approach, focusing on actions related to the capacity of future grounded in a conceptual framework, and offers a aging cohorts. Human capital accumulation is in- suite of potential indicators for monitoring interven- tensive during early life, during schooling and early tions, progress towards goals, and relevant outcomes. career development. A life course approach, how- The conceptual framework charts human capital ever, emphasize that human capital continues to be across the life course. Dashboard creation and indi- accumulated and must be protected during work- cators selected are done on the basis of a solid con- ing life. Dashboards must consider indicators that FIGURE 17.1 Healthy Longevity Initiative conceptual framework of human capital (HC) across the life course School to work & Mid career to end of Early years Younger elderly Older elderly early career working life (0-14 years old) (70-79 years old) (80+ years old) (15-35 years old) (36-69 years old) Source: O’Keefe and Haldane 2024 496 C H A P T E R 17 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E reflect this intensive human capital accumulation ation and depreciation shaped by NCD risk factors, (e.g., including education outcomes, youth-focused prevention, and management. Third, creating sup- indicators, gender norms), as well as indicators that portive conditions for healthy and productive age- reflect disease prevention to protect human capital ing focusing on actions to improve the well-being of later in life. Second, ensuring adequate prevention vulnerable older adults. During older ages, human and control of NCDs, and reduced exposure to capital accumulation declines towards depletion but their risk factors, across the life course given their interventions and policies can support human capi- increasing prevalence, economic costs, and negative tal preservation and maintain it as long as possible. impacts on productivity, human capital deprecia- Indicators comprising a healthy longevity tion, and deployment. The working life phase of the dashboard should be actionable, pragmatic, and life course requires indicators that reflect the possi- comprehensive and represent a breadth of struc- bly divergent paths between human capital appreci- ture, process, and outcomes measures (Figure 17.2). TABLE 17.1 Key actions and related domains when developing a healthy longevity dashboard Intensive human capital Human capital accumulation, Declining human capital accumulation, Life course stage accumulation deployment, and depreciation accelerated depreciation and depletion 1. Promoting enabling factors for 2. Ensuring adequate prevention and 3. Creating supportive conditions for Key action human capital accumulation control of NCDs across the life course healthy and productive ageing • Disease prevention • NCD risk factors and behaviors • Healthy population Supporting • Education outcomes • NCD management • Productive ageing domains • Youth focus • Avoidable mortality • Well-being in old age • Gender norms FIGURE 17.2 Criteria for selecting healthy longevity indicators • Align with the HLI • Based on data that • Reflect a spectrum conceptual framework is readily available, of data infrastructure • Outcomes focused regularly reported and available in countries ideally comparable data collection and • Cover the entire life across countries monitoring course • Based on data that can • Provide meaningful • Policy relevant and be disaggregated by information based on aligned with existing region, income, race, available data resources national targets and gender across a variety of or international datasets committments • Easy to measure and not based on composite measures COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E C H A P T E R 17 497 CASE STUDIES: SIERRA LEONE, INDIA, AND COLOMBIA Considering the spectrum of data infrastructure differing NCD-related challenges and opportunities and potential indicators available, we engaged with for policy action towards healthy longevity. Impor- country stakeholders in the development of nation- tantly, each country represents a different data infra- al healthy longevity dashboards in three different structure context from which information for indica- country contexts: Sierra Leone, India, and Colombia tors can be drawn. Each case study offers an example (Boxes 17.2-17.4). While the country cases have been of the strengths of a progressive approach, whereby developed with in-country teams, they are at this countries leverage available data resources to make stage illustrative and would require more in-country purposeful and actionable dashboards. Sierra Leone participation to develop them further and build the illustrates how countries with nascent data infrastruc- ownership needed for them to be more than illus- ture can bring together available indicators based on trative. Table 17.2 provides a summary of indicators key investments in healthy longevity and NCDs in a across selected countries and forthcoming compan- meaningful way. India presents a case where data is ion papers discuss each dashboard, and the engage- more readily available, including from international ment process through which it was created, in greater sources, data is driven by national policy objectives, detail. Countries were chosen to represent different and can be used for benchmarking. Colombia offers contexts not only across income level, but also in an example of how countries can leverage more exten- their epidemiological and demographic profile. These sive datasets across multiple sectors, thereby offering unique mortality constructs in each country point to an example of a more comprehensive set of indicators. TABLE 17.2 Healthy longevity dashboard indicators for Sierra Leone, India, and Colombia. Supporting Indicators by country Key action domain Sierra Leone India Colombia Promoting Disease • Life expectancy at birth • Percentage of one-year olds who • Girls 14 years old vaccinated enabling prevention • Percentage of one-year olds received BCG vaccination; three doses against Human Papillomavirus factors for who received three doses of of diphtheria, tetanus, toxoid, and (HPV) human capital diphtheria, tetanus, toxoid, and pertussis (DTPS immunization); and two accumulation pertussis (DTPS immunization) doses of measles (MCV2) immunization • Percentage of population who received at least two doses of COVID-19 vaccination • Percentage of mothers whose last birth was protected against neonatal tetanus Education • Learning-Adjusted Years of • Learning-adjusted years of schooling • Learning Poverty outcomes Schooling at age 18 • Learning-Adjusted Years of • Literacy rate (adults, youth) • Participation rate of youth and adults in Schooling • Elderly illiterate population, 65+ formal and non-formal education and years (both sex, male, female) training in the previous 12 months Youth focus • Under-five mortality rate (per • Under-five mortality rate (per 1,000 • Adolescent fertility Rate (15-19 1,000 live births) live births) years) • Youth not in employment, • School age (age 5-14 years) and • Youth not in employment, education, or training youth and young adults (age 15-29 education, or training (NEET) • Youth and young adults (age years) mortality rate (per 100,000 15-29 years) mortality rate (per population) 100,000 population) • School enrolment rate • Youth unemployment Gender norms • Women’s participation in labor • Percentage of ever-partnered women • Gender gap in labour force force (15-64 years) who experienced intimate partner participation (15-64 years) • Contributing family worker, violence in the past 12 months • Gender Gap in time spent on female (%of female employment) • Percentage of women’s work time unpaid domestic work and care • Ratio of female to male labor (paid and unpaid) spent on unpaid force participation rate work • Percentage of ever-partnered • Ratio of female to male labor force women who experienced participation rate intimate partner violence in the • OECD Social Institution and Gender past 12 months Index (SIGI) 498 C H A P T E R 17 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E Supporting Indicators by country Key action domain Sierra Leone India Colombia Ensuring NCD risk factors • Mortality from CVD, cancer, • Stunting, wasting, and underweight • Stunting prevalence among adequate & behaviors diabetes ages 30 -70 years prevalence among children under children under 5 years prevention and • Occupational risk factors 5 years • Obesity in all ages control of NCDs attributable DALYs per 100,000 • Percentage of population aged 15+ • Daily smokers (50 + years) across the life capita years who use tobacco and/or smoke • Binge alcohol drinking aged course • Occupational risk factors cigarettes 14+ years attributable DALYs • Tobacco cessation represented by • Prevalence of overweight the ratio of former to current smokers among adults, BMI ≥ 25 (age- aged 45-59 years standardized estimate) (%) • Percentage of population aged 15+ • Prevalence of insufficient years who practice harmful use of physical activity among alcohol in the past 12 months adults aged 18+ years (crude • Percentage of population aged 20+ estimate) years with overweight or obesity • Raised blood pressure • Percentage of population aged 18+ (SBP>=140 OR DBP>=90) (age- years with current depression standardized estimate) • Percentage of population aged 20-79 • Total NCD Deaths (in years with type 2 diabetes thousands) NCD **No data found** • Percentage of population aged 30-79 • % Population diagnosed with management years with hypertension that are diabetes under treatment currently taking antihypertensive • % Population diagnosed with medication hypertension under treatment Avoidable • CVD and Diabetes mortality • Life expectancy at birth • Premature Mortality due to NCD mortality rate • Adult (aged 30-69 years) mortality (30-69 years) • NCD mortality rate rate (per 100,000 population) • Life expectancy at age 60 • Age-standardized mortality • Life expectancy at age 60 • Avoidable mortality due to rate by cause (per 100,000_- depression Cardiovascular • Life expectancy at age 60 (years) • Distribution of years of life lost by major cause group Age- standardized NCD mortality rate (per 100,000 population) • Probability (%) of dying between age 30 and exact age 70 from any of cardiovascular disease, cancer, diabetes, or chronic respiratory disease Creating Healthy • Deaths per 100000 population • Mortality rate at older ages (aged • Self-perception of overall health supportive population 70 years and above) (per 100,000 (60+) conditions for population) healthy and Productive • Labor force participation rate • Labor force participation rate among • Labor Force Participation rate productive ageing (50+) population aged 15 years and over (50+) ageing • % of pensionable age adults living in poverty Well-being in **No data found** • Population covered by at least • % of adults (60+) with limitations old age one social protection cash benefit to perform activities of daily (excluding health) living (ADL) • Public expenditure on social • Self-perception of happiness protection (excluding health) and on (60+) pension, as % of GDP COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E C H A P T E R 17 499 BOX 17.1 Case study of a healthy longevity dashboard in Sierra Leone NCD context Sierra Leone is a small country on the west coast of Africa and has a population of 8 million people. It is ranked 182 out of 189 countries on the World Bank’s Human Development Index (HDI), meaning that life expectancy, education, and per capital in- come indicators are very low compared to other countries on the HDI list. Recent data indicate that the average life expectancy is 54.3 years, while the average global life expectancy is 72.3 years. Communicable diseases still account for the majority of premature and avoidable deaths in this population, though this is changing. Indeed, a double burden of communicable and noncommunicable diseases is evident in the epidemiological profile. In 2019, 34.1% of deaths were caused by NCDs with car- diovascular and digestive diseases accounting for the majority (67%) of all NCD deaths. Planning context and opportunities Governance for health and the health sector is the responsibility of the Ministry of Health and Sanitation and guided by a Na- tional Health Sector Strategic Plan, as well as several supplementary sub-sector strategies. The health sector priorities align to the Sustainable Development Goals (SDGs), including SDG 3, which has provided a framework for prioritizing investments and interventions to reach the SDG targets by 2030. However, fragmentation of planning between sectors and implementation of initiatives, coupled with a lack of data have impacted the delivery of healthy longevity initiatives centered on human develop- ment and the reduction of NCDs. The ongoing development agenda, the changing disease burden, and strong commitment to accelerate health system recovery efforts from prior health emergencies (e.g., COVID-19 and Ebola) creates opportunities to leverage data in ways that can better support monitoring and planning for the current and emergent needs of the population. Dashboard development Sierra Leone is yet to make substantial investments in vital statistical and information systems. To guide the set of development measures which would provide reasonable baseline data on NCD mortality in relation to healthy longevity in Sierra Leone, we utilized data from the following: World Health Organization Global Health Observatory, Global Burden on Disease, IDF Diabetes Atlas, World Bank, International Labor Organization, Social Institute and Gender Index. Data from these international institutes was reviewed to allow for comparability of data across India and Colombia. The research team carried out discrete work to use the Comprehensive Mortality Surveillance for Action (COMSA) data to aggregate NCD mortality by socio-demographic groups and disease groups across the four provinces of Sierra Leone. We used Tableau to develop the dashboard. Sierra Leone’s dashboard is unique for four reasons. First, the dashboard is the first to present an interactive map visually to chart NCD mortality by region and across several sociodemographic indicators. Second, the dashboard has received positive feedback from Sierra Leoneans (in-county and aboard), including Sierra Leone’s Ministry of Health and Sanitation (MoHS). Sierra Leone’s MoHS have taken an interest in the dashboard being used as a tool for surveillance and to build on the dashboard to include other risk factor indicators. Third, the dashboard has received positive feedback from potential donor sponsors like the World Bank with the prospect of developing a country-wide healthy longevity initiative pack- age centered on NCDs. Lastly, the dashboard has stimulated broader discussions with policymakers, academics, and bilateral donors. Strengths of the approach proposed for Sierra Leone • Indicators are regularly collected and are specific to each of the four provinces. • Indicators collected thus far can be built upon to add other risk factor data to study associations of disease burden on so- cio-demographic indictors. Limitations • Lack of data on risk factors at the national or district/province level. • Reliability of in-country census data. • Employment indicators are hard to define; reliability of the labor force surveys to accurately capture the diverse structure of the labor market in SL and other LMICs is required moving forward. • Data analytics in the health sector in Sierra Leone is still in its infancy. • There is some promise that Sierra Leone will one day realize the power of data mining and integrate its routine demographic 500 C H A P T E R 17 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E BOX 17.1 Case study of a healthy longevity dashboard in Sierra Leone (continued) and population health management surveys. However, a barrier to actualizing the power of data analytics is a lack of data analysts trained in specialized software such as Power BI, Tableau and R. • To overcome this barrier, efforts to deliver training on this specialized software to individuals and organizations in Sierra Leone are needed. • Still, international knowledge exchange and capacity building is required to create a sustainable change in how Sierra Leone uses data from an analytics perspective to inform policy and planning of health interventions. Key enablers • The Sierra Leone 2019-2023 NCD Policy and Strategy, including recently published literature on NCDs in Sierra Leone, provid- ed a rough guide to our selection of indicators for the healthy longevity dashboard. • Engagement with key departments/directorates within Sierra Leone’s Ministry of Health and Sanitation; these include the Directorate of Non-communicable diseases and the Directorate of Planning and Policy. Future directions and opportunities • Investment in data infrastructure. • Training on use of available data to support planning efforts, including intersectoral planning and investment, as required by the HLI framework. • Public engagement / education on the importance and use of their data to support planning. • Training for policy analysts to generate, update and use the dashboard. • Discussions with policy makers on the comparators/benchmarks that are relevant in this context (e.g., all LMICs, low-in- come-countries). COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E C H A P T E R 17 501 BOX 17.2 Case study of a healthy longevity dashboard in India NCD context The population of India is rapidly ageing, and it is estimated that the number of people over the age of 60 in India will double by 2050, accounting for nearly 20% of the population. These growing numbers of older adults, however, also have a higher risk of developing NCDs and in turn, higher morbidity, and mortality. Based on the 2019 WHO Global Health Estimates, India has a premature mortality (death before age 70) rate of 422.20 per 100,000 population. Of the total deaths under age 70, NCDs account for about 59% in 2019, an increase of 136% from the 25% that NCDs accounted for in 2010. Planning context and opportunities India has a history of central planning, with decades of economic and development progress envisioned and charted through a series of Five-Year Plans. These have been created, executed, and monitored by the Planning Commission, and more recently the NITI Aayog, up until the most recent ‘12th Five-Year Plan.’ Under the strategic vision of these plans, the Central government has shaped health policy and planning through the Council of Health and Family Welfare, including setting targets and de- veloping schemes to further health and development goals. As the population rapidly ages, policy makers across sectors are faced with an urgent need for new social protection, health, and development programs that protect older adults from poverty, support their good health, and allow for further skill development across the life course. Dashboard development We conducted a review of data available from international institutions, including the World Health Organization Global Health Observatory, World Bank World Development Indicators, Organization for Economic Cooperation and Development (OECD) Statistics, ILOSTAT, and UNESCO Institute for Statistics. We only reviewed data from international institutions as this allows com- parability of data across countries. From the global data review, we gathered a list of plausible indicators that pertain to the accumulation, deployment, and depletion of human capital and healthy longevity, with a focus on NCDs. For each indicator in the dashboard, the country is assigned a score relative to all countries in the same income strata (based on the 2021 World Bank country classification). Thus, we aim to create a healthy longevity dashboard that can be used by LMICs where good quality health and economic data are usually sparse, to monitor their human capital and guide invest- ments towards NCDs and human capital. Strengths of the approach proposed for India • International comparability of results and benchmarking • Results can be reproduced easily over time Limitations: • Lack of data on adult education and health such as adult vaccination coverage and comparability across countries. • Most health data are reported for child health but less so for adult health where majority of deaths and disabilities are ac- counted by NCDs. Key enablers • The India 2017 National Health Policy provided a guide to our selection of indicators for the healthy longevity dashboard Future directions and opportunities • Better data systems for periodic data collection, monitoring, and surveillance of the health of adults, particularly prevalence of various NCDs and access to NCD care and management • Work with country governments on usability of the dashboard and provide necessary training to government personnel on how to use and tailor the dashboard to their needs as the needs change over time 502 C H A P T E R 17 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E BOX 17.3 Case study of a healthy longevity dashboard in Colombia NCD context Colombia is one of the countries in Latin America ageing at the fastest rate. The rapid ageing of the population requires the adaptation of a range of social and economic policies and sectoral transformation to respond to the needs of an ageing popu- lation. In Colombia, nearly eleven million people seek care for a chronic condition every year, of whom between 4.5 million and 6 million people have multiple NCDs. Mortality from the four main NCDS, diabetes, cardiovascular diseases, COPD, and cancer has increased in Colombia from 241.14 to 261.47 per 100,000 population between 2015 and 2020. Planning context and opportunities Colombia offers a rich tradition of central planning: by law, a national development plan must be developed by the national planning department and presented to Congress for discussion and vote after a new government is elected. The plan pro- poses a long-term perspective on development issues, is prioritized, costed, and presents targets for a range of sectoral and multi-sectoral development issues. In addition, the recently approved document Vision Colombia 20501 outlines the vision for the country acknowledging the challenges and opportunities of the rapid aging population Colombia 1 is facing. Dashboard development Colombia has made substantial investments in core statistical and information systems over the last decades. The health and so- cial sectors offer a wealth of quality information generated from vital statistics, census data, household surveys, clinical registries, electronic health records, and administrative datasets. The research team worked with leading academics and students to review potential data sources, map potential indicators to core domains of the HLI conceptual framework, select a parsimonious set of performance indicators (21 in total) meeting the criteria defined by the research team, define disaggregators and benchmarks or targets for each indicator (using to the extent possible existing national targets), and collect data at national and sub-national level. Further development of an online, automatically population electronic dashboard will offer key performance information to national policy makers, regional system planners and service delivery managers to understand progress towards better human capital outcomes for an ageing population and monitor the effectiveness of implementation of key interventions. Strengths of the approach proposed for Colombia • Indicators are available and are regularly collected (or can be calculated) • Indicators are specific to each of the three core domains of the healthy longevity initiative Limitations • Disaggregation is not always possible: about half of the indicators can be disaggregated at the subnational level only • One third of the indicators are self-reported, risk of bias and limitations • Targets are not easy to define, only about a third of the indicators have a national target and international comparisons are not always possible Key enablers • Prior investments in strengthening information systems across government • Culture of transparency and use of data to support decision-making • Long-term planning culture in government through the National Planning Department, the National Development Plan, the Ten Years Health Plan, and the Vision 2050 Future directions and opportunities • Need to keep on investing in better information data systems for monitoring and surveillance of the health of adults and older adults • Work with government at multiple levels to ensure that electronic dashboard is built for use and meets needs of different users: national policy makers, system planners, managers 1   Gobierno de Colombia, Departamento Nacional de Planeación. Visión Colombia 2050: Discusión sobre el país del futuro [Colombia Vision 2050: Discussion on the Country’s Future]. Bogotá, Colombia: Departamento Nacional de Planeación. 2022. https://colaboracion.dnp.gov.co/CDT/Prensa/Publicaciones/Documento_vision_colombia_2050.pdf. COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E C H A P T E R 17 503 DISCUSSION AND FUTURE DIRECTIONS In creating a healthy longevity dashboard, countries bring together informational resources that allow them to monitor progress across a range of goals that contribute to health longevity and better understand the effective- ness of core HLI interventions to reach these goals. Based on our work developing a common framework for a healthy longevity dashboard, as well as the country-level case studies presented, we offer key areas for action to assist country stakeholders in producing, presenting, using, and institutionalizing a healthy longevity dashboard tailored to their unique demographic and epidemiological context and data infrastructure. Producing dashboard data Access to reliable and regularly reported data is key information products, and data governance (PAHO to developing a healthy longevity dashboard. Pro- 2019). In countries with an emerging data ecosys- ducing this data in countries, however, often proves tem, the minimum set of indicators offers a starting challenging given the chronic underinvestment in point for measuring progress on healthy longevity. data systems and IT infrastructure (Global Com- As the data ecosystem develops and matures, other mission on Evidence to Address Societal Challeng- indicators can be added to scale up the dashboard es 2022). The COVID-19 pandemic underscored while retaining the minimum set of common indi- the need for timely data to inform decision-mak- cators (Table 17.4). ing at all levels and spurred global and national in- There are several key data sources that must be vestment in improvements to the coverage, quality, built up in countries to advance a minimum set of completeness, and capacity for data collection and indicators towards a more fulsome dashboard: vital sharing (de Bienassis et al. 2022). However, these statistics, household surveys, administrative data- efforts have emerged from longstanding informa- sets, electronic health records, and outcomes mea- tional inequity. Sparse and insufficient investment surement through patient-oriented surveys. Annex in data infrastructure have long undermined local 17.2 offers key data sources on ageing. Working in efforts to produce, use, and routinely share high partnership with local statistical offices and leverag- quality data in LMICs (Hoxha, Hung, and Grépin ing routine national data at the country level, how- 2020; Kumar et al. 2018). Piecemeal approaches ever, should be favored over reliance on internation- spurred by inconsistent donor and government al surveys. Efforts to create and sustain dashboards funding, coupled with a lack of standardized data are an important opportunity to promote robust production and limited human resources have, in and reliable statistical independence at a national many places, led to a patchwork of systems and data level. Indeed, dashboard development and enhanc- sources with limited governance or quality assur- ing data production overall should be viewed as a ance (Heeks 2022; Dehnavieh et al. 2019). part of local infrastructure building and equitable The approach proposed ensures that regardless information systems strengthening. These capacity of the data ecosystem and existing data infrastruc- building efforts must be based on approaches that ture, a set of meaningful indicators can be brought uphold privacy, ensure security, promote equitable together to guide monitoring and decision-making. and open sharing and dissemination, and consid- Drawing on the concept of maturity models, which er data as a public good, both at the local and the posit that people, organizations, processes, and sys- global level (OECD 2022). For example, the Dis- tems evolve in complexity and capacity through a trict Health Information System 2 (DHIS2) is an process of development over stages (Kolukisa et al. open-source web-based platform, the world’s larg- 2020; Vidal, Rocha, and Abreu 2019), Table 17.3 est health management information system, and characterizes the data-related capabilities across is used by 73 LMICs for data warehousing, visual- three infrastructure maturity levels – emerging, es- ization and real-time analysis of live data (DHIS2 tablished, and advanced. The infrastructure matu- 2022). The DHIS2 is not only a potential source of rity stage reflects capabilities and capacities across indicator data for countries with emerging infra- key domains adapted from the Information Sys- structure, but an important example of open data tems for Health Toolkit including IT infrastructure, to monitor and evaluate health and well-being goals namely data sources, quality and interoperability, (Dehnavieh et al. 2019). 504 C H A P T E R 17 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E TABLE 17.3 Data ecosystem features by domain and information infrastructure maturity level. Information ecosystem domain Emerging infrastructure Established infrastructure Advanced infrastructure IT infrastructure • Basic IT tools are generally available • Some advanced IT tools are available in • Advanced IT tools are ubiquitous (e.g., (e.g., hardware, software, internet well-resourced areas, but basic IT tools cloud systems) and innovative and are connectivity), but may be older or not are widely available. routinely evaluated and adopted. performing well. • Majority of data are digitally collected • Data sources are virtually all digitally • Paper-based data collection is from all key data sources, but some collected. common, with some digital data paper-based collection occurs. collection. Data sources • Key data sources have been identified • Key data sources are well established and • Data from all key data sources, as well (e.g., vital statistics, household there is evidence of some data collection as data from multiple and unstructured surveys, morbidity, mortality, and from multiple and unstructured sources, sources is routinely collected. surveillance data). such as social media and others. • There are established mechanisms to • Data may be routinely collected but • Data is routinely collected for the majority assess existing data sources that can challenges remain in ensuring data of key sources. improve them or integrate new ones. flow. Quality and • Some use of data standards, however • Standards have been adopted and some • Standards and national information interoperability few formalized standards are routinely sectors have a defined and documented architecture across sectors have been enforced. national information architecture. formally adopted and are periodically • There is limited or no interoperability • There is evidence of interoperability checked. between information systems. between some information systems. • There is significant interoperability across systems and across sectors including integrated national data repositories from multiple data sources. Information • Reports and other information • Reports and other information products • Stakeholders across sectors can products products are published but these are are routinely produced and tailored to effectively collaborate and engage with infrequent and resource intensive to meet stakeholder needs. data to produce real-time information produce. • Core datasets are readily available and products. • Data is not readily shared and requires shared. • All data is shared on principles of permission from the highest levels. OpenData. Data governance • Limited strategy and mechanisms • Data governance mechanisms have been • Formal and cross-sectoral data for data collection, use, sharing, and established among some stakeholders. governance mechanisms have been dissemination. • There is evidence of continuous established and are function effectively. • Limited privacy standards are improvement processes to monitor and • Continuous improvement processes developed. invest in data quality. are routine and drive innovation across • Privacy standards have been defined systems. at least for key data sources and • Data is considered as a public good enforcement occurs. with strong privacy standards and enforcement. Source: Adapted from PAHO Information Systems for Health (IS4H) Toolkit: https://www3.paho.org/ish/images/docs/about-IS4H-mm.pdf?ua=1 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E C H A P T E R 17 505 TABLE 17.4 Proposed healthy longevity indicators by information infrastructure maturity level Minimum Set of Indicators For countries Across Spectrum of Data Infrastructure Supporting Key action domain Limited Data Infrastructure Moderate Data Infrastructure Advanced Data Infrastructure Promoting Disease • % of one year old with full • Percentage of population who • % of population with full enabling prevention immunization, by sex received at least two doses of immunization according to factors for • Under-five mortality rate (per COVID-19 vaccine vaccination national immunization plan for human capital 1,000 live births) • Youth vaccinated against Human each age category accumulation Papillomavirus (HPV), by sex Education • Learning Poverty • Learning Poverty • Learning Poverty outcomes • Learning-Adjusted Years of • Learning-adjusted Years of schooling • Learning-Adjusted Years of Schooling Schooling Youth focus • School enrolment rate • Adolescent fertility Rate (15-19 years) • Adolescent fertility Rate (15-19 • Youth unemployment • Youth not in employment, education, years) • Adolescent fertility Rate (15-19 or training (NEET) • Youth not in employment, years) education, or training (NEET) Gender • Percentage of ever-partnered • Percentage of women’s work time • Gender gap in labor force norms women who experienced (paid and unpaid) spent on unpaid participation (15-64 years) intimate partner violence in work • Gender Gap in time spent on the past 12 months unpaid domestic work and care Ensuring NCD risk • Stunting, wasting, and • Stunting, wasting, and underweight • Stunting prevalence among adequate factors & underweight prevalence prevalence among children under children under 5 years prevention behaviors among children under 5 years 5 years • Obesity in all ages and control of • Prevalence of overweight • Percentage of population aged 20+ • Daily smokers (50 + years) NCDs across among adults, BMI ≥ 25 (age- years with overweight or obesity • Binge alcohol drinking aged the life course standardized estimate) (%) Percentage of population aged 15+ 14+ years years who use tobacco and/or smoke cigarettes NCD • Percentage of adult population • Percentage of population aged 15+ • % Population diagnosed with management with hypertension years with type 2 diabetes that are diabetes under treatment • Percentage of adult population currently taking medication to lower • % Population diagnosed with with type 2 diabetes blood glucose level hypertension under treatment • Percentage of population aged 30-79 years with hypertension that are currently taking antihypertensive medication Avoidable • Life expectancy at birth • Life expectancy at birth • Premature Mortality due to NCD mortality • Life expectancy at age 60 • Life expectancy at age 60 (30-69 years) (years) • Mortality from CVD, cancer, diabetes • Life expectancy at age 60 • Adult (aged 30-69 years) ages 30 -70 years • Avoidable mortality due to mortality rate (per 100,000 depression population) Creating Healthy • Deaths per 100,000 population • Mortality rate at older ages (aged • Self-perception of overall health supportive population 70 years and above) (per 100,000 (60+) conditions for population) healthy and Productive • Labor force participation rate • Labor force participation rate among • Labor Force Participation rate productive ageing among population (50+) population (50+) (50+) ageing • % of pensionable age adults living in poverty Well-being • Public expenditure on social • Population covered by at least • % of adults (60+) with limitations in old age protection (excluding health) one social protection cash benefit to perform activities of daily and on pension, as % of GDP (excluding health) living (ADL) • Self-perception of happiness (60+) 506 C H A P T E R 17 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E Presenting and using data in the dashboard Data visualization is an important strength of a goals that improve their health and happiness. dashboard-based measurement approach. A suc- To realize the potential of performance dash- cessful performance dashboard should be created boards, users must be able to engage with visual from a robust and user-centered design philosophy summaries to monitor progress, assess intervention grounded in accessibility, clarity, and meeting user effects, and think strategically about the healthy lon- needs (Kunjan, Doebbeling, and Toscos 2019; Gha- gevity commitments behind the indicators (Table zisaeidi et al. 2015). Users must be able to clearly 17.5). The public, including the media and civil soci- see the status of indicators and navigate through in- ety organizations, can use dashboards to participate creasing levels of detail depending on their unique in public policy making, monitor progress, and advo- informational needs. The dashboard, by meeting cate for greater action towards goals set at the high- these needs, providing open access to information, est level that impact health, happiness, productivity, and engaging with the political economy of mea- and the ability to age in place in communities. Policy surement and goal setting, can then support better makers will use the dashboard for performance im- policymaking and program integration. provement and to strengthen cross-sectoral cooper- There are four key dashboard users – the pub- ation. To do so, they require access to visualizations lic, policy makers, planners, and managers. The across a mix of outcomes and process indicators to creation and content of a dashboard represents a monitor the impact of interventions and determine tailored measurement approach to motivate action, their effectiveness to drive strategic decision mak- prioritization, and support decision-making by us- ing and investment. Planners will require indicators ers from the highest to most local levels of action describing outcomes, processes, and structures to (Kruglov, Strugar, and Succi 2021). Performance determine whether adequate and appropriate in- dashboards, however, also have important applica- frastructure and resources are in place to achieve tions for broader planning, local priority setting and goals and deliver programs. Whereas managers of programming, and are an important avenue for pub- programs that aim to promote or enhance healthy lic accountability (Veillard et al. 2016; Edwards and longevity in communities require dashboards capa- Thomas 2005). Performance dashboards, and the ble of visualizing granular details relevant to their exercises necessary to create them, put important is- context and programmatic goals for implementation sues on the agenda of decision-makers at all levels of and evaluation. These government stakeholders may influence, and are an important inroad for the public come from diverse sectors including national plan- to see, and advocate for, their data put to use towards ning departments, as well as ministries of finance, TABLE 17.5 Key dashboard users, their purpose, and types of measures needed by users User Purpose Types of measures needed General public • Transparency • Select outcome measures curated at their level • Participation in public policy making of involvement and relevant to their context and programmatic goals • Key performance indicators Policy maker • Transparency • Mix of outcomes, outputs and process indicators • Strategic goal setting • Indicators of relevance and efficacy of interventions • Performance improvement • Monitoring indicators • Cross-sectoral cooperation enhancement • Monitoring short term goals to achieve long term goals Planner • Transparency • Outcomes, process, and structure indicators • Benchmarking and best practices • Adequacy of infrastructure and resources to • Performance improvement achieve goals • Cooperation enhancement • Indicators of effectiveness and efficacy • Capacity building Manager • Transparency • Outcomes, process, and structure indicators • Performance improvement • Ability to access and visualize granular details • Cooperation enhancement relevant to their context and programmatic goals • Capacity building • Implementation Source: Adapted from Veillard et al. 2016. COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E C H A P T E R 17 507 health, and ageing amongst others. It is important the evidence-to-policy literature, institutionalization to recognize that dashboards must be built from the of such efforts can occur through governance, in oth- outset with flexibility in use to meet the needs of dif- er words, through rule-making and steering-related ferent users with different interests and needs. functions, structures or platforms that increase vis- Meeting these goals requires user journeys, ibility and protect it from ad hoc changes in politics information architecture, and visualizations that and contexts (Kuchenmuller et al. 2022).54 It can also are fit for purpose and prioritize usability to meet be achieved through the use of standards and rou- unique user goals, while ensuring a cohesive archi- tinized processes, that is products and processes that tecture. Dashboard design should be accessible and policy-makers trust and use, serving as institutional engaging, with enough explanatory information to memory and reducing reliance on individual people assure users of the quality, accuracy, and sources of with knowledge and skills. Partnerships and collec- the data underpinning the indicators. Functional- tive action are another way to institutionalize dash- ities should include user segmented journeys, maps boards, leveraging the extent to which stakeholders to show regional variations, and the ability to dis- interact as a mechanism for continued engagement play disaggregate data across several dimensions. and involvement for the same cause. Leadership and commitment to the dashboard through strong lead- ership with the ability to ensure long-lasting adop- Institutionalizing the dashboard tion through allocation of resources, encourage- ment, support, and mentorship can also add to its Once assembled, a healthy longevity dashboard en- sustainability. Dashboards can be institutionalized ables countries to establish contextually relevant, when supported with adequate resources, human, data-driven priorities to promote human capital financial, material, and information resources, in a and happiness across the life course and cannot sustainable and stable way. Finally, organizational be just a one-off exercise. Through the dashboard culture must align to ensure dashboards serve an on- creation process, countries clarify their healthy lon- going purpose. That is the basic values, assumptions, gevity goals, commitments, and investments, while artefacts, and beliefs that are considered valid and committing to ongoing goal monitoring. In doing promoted, support dashboard inputs and ongoing so, the dashboard can be a tool for change at the use. All these institutionalizing levers can be used to highest levels, shifting the power dynamic between ensure a healthy longevity dashboard is an enduring international funders and recipients and empow- tool as part of a cohesive measurement approach. In- ering countries to identify, monitor, and course stitutionalizing a dashboard offers a useful approach correct investments and interventions based on to long-term planning towards healthy longevity. It their own data availability and evaluation processes is also important to note enduring obstacles such as (Global Commission on Evidence to Address Soci- lack of sustainable investments and low human re- etal Challenges 2022). sources capacity that often result in management in- Institutionalizing a dashboard is critical to en- formation systems (of which dashboards are a sub- sure that the data informing the indicators is main- set) that are inadequately maintained or developed. tained and able to persist overtime. Learning from LIMITATIONS There are several limitations to the dashboard creation approach we describe here. First, when choosing indicators, we prioritized be- indicators on an enabling environment, where so- ing selective and parsimonious to ensure that the cieties and the built environment are built in such dashboards included the most relevant indicators a way that people are able to live healthy lives, ac- based on accessible and robust data. When consid- cumulate and protect human capital in their com- ering population-level data, data quality varies, in munities, eventually ageing in place. More work particular at sub-national levels and for indicators is needed to determine how best to measure and where the denominator is unknown. Second, our monitor this important aspect of healthy longevity framework and dashboards currently do not include (Wang et al. 2021; Lak et al. 2021; Kim et al. 2022). 508 C H A P T E R 17 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E Third, the dashboard is limited in its ability to track on a specific area, for example the key actions or current multi-sectoral investments and government domains we’ve highlighted, you risk losing focus on spending on healthy longevity. Future work should other unmeasured areas that need greater attention consider how to aggregate and monitor spending and investment (Kyeremanteng et al. 2019; Eijkenaar across the diverse interventions that preserve human 2013). This risk highlights the need for broader and capital and promote health longevity across the life interconnected monitoring efforts within and across course. Lastly, there may be risks to a healthy longev- countries, as well as the importance of contextual- ity dashboard that must be monitored. The pay for izing a healthy longevity dashboard within larger performance literature underscores that by focusing planning and programmatic efforts. CONCLUSION As the global population ages and the burden of NCDs grows, ensuring healthy longevity will be increasingly important to policymakers, planners, managers, and the public, particularly in LMICs. A healthy longevity dashboard is an important tool studies underscore that creating a healthy longevity firstly, to indicate progress across a range of goals dashboard is in reach across a spectrum of data in- that offer insights as to whether a country is pre- frastructure maturity levels and demographic con- paring for and adapting to an ageing population texts. A healthy longevity dashboard is ultimately and higher burden of NCDs; and secondly, to offer more than a monitoring tool. It represents a com- information on effectiveness of core interventions mitment to the health, productivity, and happiness to promote healthy longevity. In doing so, a healthy goals that underpin selected indicators and is an longevity dashboard can guide better systems plan- opportunity to use data for targeted interventions ning and management and determine multisec- and to promote strategic approaches to investing in toral interventions with a life course lens. Our case people across the life course. REFERENCES 1. AHEAD. 2022. Ending the HIV Epidemic in the U.S. America’s HIV ences 73 (February): 257–62. Epidemic Analysis Dashboard (AHEAD). https://ahead.hiv.gov 8. Chaker, Layal, Abby Falla, Sven J. van der Lee, Taulant Muka, Da- (accessed 14 October 2022). vid Imo, Loes Jaspers, Veronica Colpani, et al. 2015. “The Global 2. Allen, Luke N, Brian D Nicholson, Beatrice Y T Yeung, and Fran- Impact of Non-Communicable Diseases on Macro-Economic cisco Goiana-da-Silva. 2020. “Implementation of Non-Commu- Productivity: A Systematic Review.” European Journal of Epide- nicable Disease Policies: A Geopolitical Analysis of 151 Coun- miology 30 (5): 357–95. tries.” The Lancet Global Health 8 (1): e50–58. 9. Corral, Paul, Nicola Dehnen, Ritika D’Souza, Roberta Gatti, and 3. Baranek, Patricia, Jeremy Veillard, and John Wright. 2011. Aaart Kray. 2021. “The World Bank Human Capital Index.” In “Benchmarking Health Care in Federal Systems: The Canadian Measuring Human Capital, edited by Barbara Fraumeni. Aca- Experience.” In Benchmarking in Federal Systems, edited by Alan demic Press. Fenna. Forum of Confederations. 10. Dehnavieh, Reza, AliAkbar Haghdoost, Ardeshir Khosravi, Fa- 4. Beard, John R, and David E Bloom. 2015. “Towards a Compre- hime Hoseinabadi, Hamed Rahimi, Atousa Poursheikhali, Na- hensive Public Health Response to Population Ageing.” The hid Khajehpour, et al. 2018. “The District Health Information Lancet 385 (9968): 658–61. System (DHIS2): A Literature Review and Meta-Synthesis of Its 5. Bienassis, Kate de, Rie Fujisawa, Tiago C.O. Hashiguchi, Niek Strengths and Operational Challenges Based on the Experi- Klazinga, and Jillian Oderkirk. 2022. “ ences of 11 Countries.” Health Information Management Jour- 6. Bloom, David E., and Alfonso Sousa-Poza. 2013. “Ageing and nal 48 (2): 62–75. Productivity: Introduction.” IZA Discussion Papers, No. 7205. 11. DHIS2. 2022. District Health Information System 2 (DHIS2). Bonn, Germany: Institute of Labor Economics (IZA). www.dhis2.org. 7. Bonney, Wilfred. 2013. “Applicability of Business Intelligence in 12. Eckerson, Wayne W. 2010. Performance Dashboards: Measur- Electronic Health Record.” Procedia - Social and Behavioral Sci- ing, Monitoring, and Managing Your Business. 2nd ed. Hobo- COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E C H A P T E R 17 509 ken, N.J.: John Wiley. Kim, and Sunghwan Cho. 2021. “Measurement Indicators 13. Edwards, David, and John Clayton Thomas. 2005. “Developing of Age-Friendly Communities: Findings from the AARP a Municipal Performance-Measurement System: Reflections Age-Friendly Community Survey.” Edited by Suzanne Meeks. on the Atlanta Dashboard.” Public Administration Review 65 The Gerontologist 62 (1). (3): 369–76. 26. Kolukısa Tarhan, Ayça, Vahid Garousi, Oktay Turetken, Mehmet 14. Eijkenaar, Frank. 2011. “Key Issues in the Design of Pay for Per- Söylemez, and Sonia Garossi. 2020. “Maturity Assessment and formance Programs.” The European Journal of Health Economics Maturity Models in Health Care: A Multivocal Literature Re- 14 (1): 117–31. view.” Digital Health 6 (January): 205520762091477. 15. GBD 2019 Ageing Collaborators. 2022. “Global, Regional, and 27. Kruglov, Artem, Dragos Strugar, and Giancarlo Succi. 2021. “Tai- National Burden of Diseases and Injuries for Adults 70 Years lored Performance Dashboards—an Evaluation of the State of and Older: Systematic Analysis for the Global Burden of Disease the Art.” Peer J Computer Science 7 (October): e625. 2019 Study.” BMJ, March, e068208. 28. Kuchenmüller, Tanja, Laura Boeira, Sandy Oliver, Kaelan Moat, 16. Ghazisaeidi, Marjan, Reza Safdari, Mashallah Torabi, Mahboo- Fadi El-Jardali, Jorge Barreto, and John Lavis. 2022. “Domains beh Mirzaee, Jebraeil Farzi, and Azadeh Goodini. 2015. “Devel- and Processes for Institutionalizing Evidence-Informed Health opment of Performance Dashboards in Healthcare Sector: Key Policy-Making: A Critical Interpretive Synthesis.” Health Re- Practical Issues.” Acta Informatica Medica 23 (5): 317. search Policy and Systems 20 (1). 17. Global Commission on Evidence to Address Societal Challeng- 29. Kumar, Manish, David Gotz, Tara Nutley, and Jason B. Smith. es. 2022. “The Evidence Commission Report: A Wake-up Call 2017. “Research Gaps in Routine Health Information System and Path Forward for Decision-makers, Evidence Intermediar- Design Barriers to Data Quality and Use in Low- and Middle-In- ies, and Impact-oriented Evidence Producers.” Hamilton: Mc- come Countries: A Literature Review.” The International Journal Master Health Forum. of Health Planning and Management 33 (1): e1–9. 18. Heeks, Richard. 2022. “Digital Inequality beyond the Digital Di- 30. Kunjan, Kislaya, Bradley Doebbeling, and Tammy Toscos. 2018. vide: Conceptualizing Adverse Digital Incorporation in the Glob- “Dashboards to Support Operational Decision Making in al South.” Information Technology for Development 28 (4): 1–17. Health Centers: A Case for Role-Specific Design.” International 19. Helminski, Danielle, Jacob E Kurlander, Anjana Deep Renji, Jer- Journal of Human–Computer Interaction 35 (9): 742–50. emy B Sussman, Paul N Pfeiffer, Marisa L Conte, Oliver J Gada- 31. Kyeremanteng, Kwadwo, Raphaëlle Robidoux, Gianni D’Egi- bu, et al. 2022. “Dashboards in Health Care Settings: Protocol dio, Shannon M. Fernando, and David Neilipovitz. 2019. “An for a Scoping Review.” JMIR Research Protocols 11 (3): e34894. Analysis of Pay-For-Performance Schemes and Their Potential 20. Holzer, Marc, and Kaifeng Yang. 2004. “Performance Measure- Impacts on Health Systems and Outcomes for Patients.” Critical ment and Improvement: An Assessment of the State of the Art.” Care Research and Practice 2019 (1): 1–7. International Review of Administrative Sciences 70 (1): 15–31. 32. Lak, Azadeh, Parichehr Rashidghalam, S. Nouroddin Amiri, 21. Hoxha, Klesta, Yuen W Hung, Bridget R Irwin, and Karen A Phyo K. Myint, and Hamid R. Baradaran. 2021. “An Ecological Grépin. 2020. “Understanding the Challenges Associated Approach to the Development of an Active Aging Measure- with the Use of Data from Routine Health Information Sys- ment in Urban Areas (AAMU).” BMC Public Health 21 (1). tems in Low- and Middle-Income Countries: A Systemat- 33. Mathers, Colin D, and Dejan Loncar. 2006. “Projections of Global ic Review.” Health Information Management Journal, June, Mortality and Burden of Disease from 2002 to 2030.” Edited by 183335832092872. Jon Samet. PLoS Medicine 3 (11): e442. 22. Hulme, David, Antonio Savoia, and Kunal Sen. 2014. “Gover- 34. McArthur, John. 2013. “Own the Goals: What the Millennium nance as a Global Development Goal? Setting, Measuring and Development Goals Have Accomplished.” Foreign Affairs 92 (2). Monitoring the Post-2015 Development Agenda.” ESID Work- 35. McArthur, John W. 2014. “The Origins of the Millennium De- ing Paper No.32, University of Manchester. velopment Goals.” The SAIS Review of International Affairs 34 23. Ivanković, Damir, Erica Barbazza, Véronique Bos, Óscar Brito (2): 5–24. Fernandes, Kendall Jamieson Gilmore, Tessa Jansen, Pinar Kara, 36. McArthur, John W., and Krista Rasmussen. 2018. “Change of et al. 2021. “Features Constituting Actionable COVID-19 Dash- Pace: Accelerations and Advances during the Millennium De- boards: Descriptive Assessment and Expert Appraisal of 158 velopment Goal Era.” World Development 105 (May): 132–43. Public Web-Based COVID-19 Dashboards.” Journal of Medical 37. Means, Arianna Rubin, Christopher G. Kemp, Marie-Claire Gwayi- Internet Research 23 (2): e25682. Chore, Sarah Gimbel, Caroline Soi, Kenneth Sherr, Bradley H. 24. Kaplan, Robert S., and David P. Norton. 2007. “Using the Balanced Wagenaar, Judith N. Wasserheit, and Bryan J. Weiner. 2020. “Eval- Scorecard as a Strategic Management System.” Harvard Busi- uating and Optimizing the Consolidated Framework for Imple- ness Review. July 2007. https://hbr.org/2007/07/using-the-bal- mentation Research (CFIR) for Use in Low- and Middle-Income anced-scorecard-as-a-strategic-management-system. Countries: A Systematic Review.” Implementation Science 15 (1). 25. Kim, Kyeongmo, Tommy Buckley, Denise Burnette, Seon 38. Miranda, J. J., S. Kinra, J. P. Casas, G. Davey Smith, and S. Ebra- 510 C H A P T E R 17 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E him. 2008. “Non-Communicable Diseases in Low- and Mid- cine-equity/ (accessed 14 October 2022). dle-Income Countries: Context, Determinants and Health Pol- 50. Veillard Jeremy, K Denny, B Tipper, N Klazinga. 2016. “Using Per- icy.” Tropical Medicine & International Health 13 (10): 1225–34. formance Measurement and Monitoring for Performance Im- 39. Narh-Bana, Solomon A., Mary Kawonga, Selase Adjoa Odopey, provement.” In Managing a Canadian Healthcare Strategy edited Frank Bonsu, Latifat Ibisomi, and Tobias F. Chirwa. 2022. “Fac- by S Carson and K Nossal. Kingston: Queen’s University Press. tors Influencing the Implementation of TB Screening among 51. Veillard, Jeremy, Tai Huynh, Sten Ardal, Sowmya Kadandale, PLHIV in Selected HIV Clinics in Ghana: A Qualitative Study.” Niek S. Klazinga, and Adalsteinn D. Brown. 2010. “Making BMC Health Services Research 22 (1). Health System Performance Measurement Useful to Policy 40. O’Keefe, Philip, and Victoria Haldane. 2024. “Towards a Frame- Makers: Aligning Strategies, Measurement and Local Health work for Impact Pathways between Non-communicable Dis- System Accountability in Ontario.” Healthcare Policy 5 (3). eases, Human Capital and Healthy Longevity, Economic and 52. Vidal Carvalho, João, Álvaro Rocha, and António Abreu. 2017. Wellbeing Outcomes,” in Unlocking the Power of Healthy Lon- “Maturity of Hospital Information Systems: Most Important gevity: Compendium of Research for the Healthy Longevity Initia- Influencing Factors.” Health Informatics Journal 25 (3): 617–31. tive. Washington D.C.: World Bank. 53. Wang, Chenghao, Diego Sierra Huertas, John W. Rowe, Ruth 41. Organization for Economic Cooperation and Development Finkelstein, Laura L. Carstensen, and Robert B. Jackson. 2021. (OECD). 2002. “Measuring Up: Improving Health System Perfor- “Rethinking the Urban Physical Environment for Century-Long mance in OECD Countries.” Paris: OECD Publishing. Lives: From Age-Friendly to Longevity-Ready Cities.” Nature Ag- 42. ———. Health Data Governance for the Digital Age: Imple- ing 1 (12): 1088–95. menting the OECD Recommendation on Health Data Gover- 54. World Bank. 2020. “The Human Capital Index 2020 Update: nance. Paris: OECD Publishing. Human Capital in the Time of COVID-19.” Washington, D.C: 43. Pan-American Health Organization (PAHO). 2019. “Information World Bank. Systems for Health Toolkit IS4H Maturity Assessment Tool,” 55. World Bank Group. 2020. “Human Capital Project: HCI Compass.” IS4H-MM2.0. Geneva, Switzerland: World Health Organization. Human Capital Project Washington, D.C.: World Bank Group. 44. Pfeffer, Jeffrey, and Robert I. Sutton. 2006. “Evidence-Based 56. World Health Organization (WHO). 2004. “Compendium of In- Management.” Harvard Business Review. January 2006. https:// dicators for Monitoring and Evaluating National Tuberculosis hbr.org/2006/01/evidence-based-management. Programs.” Geneva, Switzerland: World Health Organization. 45. RBM Partnership to End Malaria. 2022. Global Malaria Dash- 57. ———. 2014. “Global Status Report on Noncommunicable board. Copenhagen, Denmark: United Nations Office for Proj- Diseases 2014: Attaining the Nine Global Noncommunicable ect Services. https://dashboards.endmalaria.org/dashboard Diseases Targets, a Shared Responsibility.” Geneva, Switzerland: (accessed 14 October 2022). World Health Organization. 46. Stein, Felix, and Devi Sridhar. 2019. “Back to the Future? 58. ———. 2022a. Summary of Tuberculosis Data. Geneva, Swit- Health and the World Bank’s Human Capital Index.” BMJ, No- zerland: World Health Organization. https://worldhealthorg. vember, l5706. shinyapps.io/TBrief/?_inputs_&sidebarCollapsed=true&enti- 47. United Nations Children’s Fund (UNICEF). 2022a. Child Health and ty_type=%22country%22&iso2=%22AF%22&sidebarItemEx- Well-Being Dashboard. https://data.unicef.org/resources/child- panded=null (accessed 13 October 2022). health-and-well-being-dashboard/ (accessed 14 October 2022). 59. ———. 2022b. Polio Transition Programme Monitoring and 48. _____. 2022b. “The Countdown Country Profile: A Tool for Ac- Evaluation Dashboard. Geneva, Switzerland: World Health Or- tion.” https://data.unicef.org/countdown-2030/ (accessed 14 ganization. https://www.who.int/teams/polio-transition-pro- October 2022). gramme/monitoring-and-evaluation-dashboard (accessed 14 49. United Nations Development Programme (UNDP). 2022. Glob- October 2022). al Dashboard for Vaccine Equity. https://data.undp.org/vac- COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E C H A P T E R 17 511 ANNEX 17.1 EXAMPLES OF DASHBOARDS FOR GLOBAL GOALS Below we offer an overview of select dashboards used to monitor progress and chart strategic action towards global goals across the health, environment, as well as the development sectors. DASHBOARDS IN THE HEALTH SECTOR In the health sector, dashboards are descriptive and mainly used for tracking progress against global goals across vertical disease or stakeholder programming (such as tuberculosis, malaria, polio, maternal and child health) and are produced by international organizations (such as WHO) or consortia tasked with motivating progress towards established goals. Polio transition programme monitoring and evaluation dashboard Aims to be the central repository of output and pro- https://www.who.int/teams/polio-transition-pro- cess indicators to assess progress, identify areas for im- gramme/monitoring-and-evaluation-dashboard. provement, and document achievements. The dash- The dashboard offers a results chain of the po- board aims to ensure a transparent and regular way lio transition beginning with objects and following of informing stakeholders on progress towards polio inputs, outputs, and outcomes to achieve impact in transition goals, as well as monitoring the quality of four areas – a polio free world, reduced under five the polio immunization, surveillance, and outbreak child morbidity and mortality, and increased global response activities during the transition period, and health security. Indicators for further viewing can detecting areas for improvement during the transition be accessed across objectives: 1) sustain a polio-free period so as to take corrective action when needed: world after eradication; 2) strengthen immuniza- FIGURE 17A.1a Polio transition program monitoring and evaluation dashboard Source: https://www.who.int/teams/polio-transition-programme/monitoring-and-evaluation-dashboard 512 C H A P T E R 17 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E tion systems and surveillances; and 3) strengthen accessed and sorted by region and year. emergency, preparedness, detection and response The dashboard identifies priority countries capacity (through the International Health Reg- across WHO regions and is based on the compo- ulations); as well as number of process indicators. nents of the Strategic Action Plan on Polio Transi- Once an indicator is selected, visualizations can be tion (WHA 171_9)2. FIGURE 17A.1b Polio transition program monitoring and evaluation dashboard Source: https://www.who.int/teams/polio-transition-programme/monitoring-and-evaluation-dashboard 2   https://www.who.int/publications/i/item/A71-9 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E C H A P T E R 17 513 RBM partnership to end malaria global malaria dashboard The dashboard offers a comprehensive overview of The campaign dashboard provides a glob- efforts towards goals to end malaria globally. Under al overview on the number of planned campaigns the main dashboard are several sub-dashboards that by intervention and status by country. The dash- track progress on end malaria strategies and goals board includes indicators on distribution of insec- globally. Dashboard available include information ticide-treated mosquito nets (ITN), indoor residu- on campaigns, supply chain, global funding, coun- al spraying (IRS), as well as preventive treatments try support, malaria epidemiology, programmatic for children (SMC) and pregnant women. Data is and financial gaps, commodities forecasting, sur- drawn from the End Malaria Country/Regional veillance projects, and seasonal climate forecasts Support Partner Committee CRSPC country track- that offer probabilistic outlooks on temperature and er platforms. The map visual includes a colour indi- precipitation given the sensitive of malaria-carrying cator that reports the number of campaigns at risk mosquitos to environmental conditions. or off track within a country. FIGURE 17A.2 Global Malaria Dashboard – Campaign Dashboard Source: https://dashboards.endmalaria.org/dashboard The Supply Chain Dashboard aims to enable de- rapid diagnostic tests (RDTs) and treatment (Arte- cision-makers to visualize integrated data on stock lev- misinin-based combination therapy), as well as pre- els of key commodities across countries, as well as the vention tools. Data is drawn from the CRSPC tracker status of deliveries. It is offers information on targets and triangulated with information from main donors for essential commodities set by the WHO including including the Gates Foundation and PMI/USAID. 514 C H A P T E R 17 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E FIGURE 17A.3 Global Malaria Dashboard – Supply Chain Dashboard Source: https://dashboards.endmalaria.org/dashboard Countdown to 2030 for women’s, children’s and adolescents’ health Builds upon the successes of the Countdown to • Demographic information (population, birth, 2015 effort. It is a multi-stakeholder global move- death, and context domains); continuum of care ment that aims to accelerate the momentum to coverage and equity (pre-pregnancy, pregnancy, achieve the targets of the Sustainable Development birth, postnatal, infancy, childhood, and envi- Goals (SDGs) for ending preventable maternal, ronment domains); mortality and causes of death newborn, and child deaths; and catalyze efforts to (maternal, neonatal, child and adolescent mortal- achieve the vision of the Global Strategy for Wom- ity including causes of death). en’s, Children’s and Adolescents’ Health. ‘The Countdown Country Profile: A Tool For Ac- • Nutrition, breastfeeding practices, child nutrition- tion’ present the latest evidence to assess country prog- al status, maternal and newborn health, child and ress in improving women’s, children’s, and adolescents’ adolescent health, environment, equity indicators. health: https://data.unicef.org/countdown-2030/ The tool draws on data on data from multiple • Policies systems and financing indicators; service global databases informed by national and interna- delivery indicators. tional surveys and administrative sources. It pres- ents this data using visualizations and tables dis- The tool offers ways to use the evidence presented played on a country-specific interactive dashboard for action highlighting that all actors can leverage the across a range of indicators presented using pro- dashboard to improve accountability for women’s, gressive disclosure to reveal increasing complexity children’s, and adolescents’ health. More specifically, and detail including (organized by page): civil society organizations and advocates are advised that the tool can be used to highlight areas where COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E C H A P T E R 17 515 progress has been lacking. Accountability is empha- use in annual review cycles. It is also framed as a cat- sized and the dashboard offered as a tool to help users alyst for discussions around priority investments, and hold country government and partners accountable for strategic policy and programmatic prioritization. for fulfilling their commitments to the 2030 agenda. The dashboard is also offered to academic partners To country governments and policy makers, the dash- and research institutions as a database for comparative board is presented as a planning tool, particularly for analyses to reveal trends across countries and regions. FIGURE 17A.4 Page 1 of The Countdown Country Profile: A Tool For Action for Colombia Source: https://data.unicef.org/countdown-2030/country/Colombia/1/ 516 C H A P T E R 17 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E Child health and well-being dashboard The dashboard was designed through a consulta- green signifying good progress or that the CRC target tive process with the Children in All Policies 2030 has been met, yellow indicates that moderate progress (CAP2030) Data and Learning Working Group, has been made or the target is on track, and red giving through which consensus was reached on basic a warning that the target needs attention or accelera- dashboard parameters through a three-step pro- tion to meet goals. Countries can be sorted by SDG, cess.3 The dashboard is an additional accountability WHO, or UNICEF regions, as well as by income. Data mechanism to assess progress on the UN Conven- can be filtered by age, domain, and country with op- tion on the Rights of the Child (CRC). The us- tions to view the numbers instead of the stoplight vi- er-friendly dashboard helps countries to regularly sualization, as well as colour-blind friendly formatting. monitor their progress on child health and wellbe- The dashboard presents indicators across four ing and make evidence-based decisions about pri- domains – survival, protection, development, and ority areas for action and resource allocation. participation. Thirty-seven indicators are visual- The Child Health and Well-being Dashboard ized, four indicators per age bracket are offered aims to make data available to improve child and ado- across six age brackets ranging from birth to 19 lescent well-being and health: https://data.unicef.org/ years, as well as context and policy indicators. resources/child-health-and-well-being-dashboard/ The dashboard highlights that the data and vi- The tool draws on data from the UNICEF data sualizations allow policymakers, governments, and warehouse comprised of data sets from national and organisations to easily monitor and compare a se- international surveys, as well as administrative data. It lection of indicators by region, country, age group, presents the data using a stoplight visualization with domain and income. FIGURE 17A.5 Section of the Child Health and Well-being Dashboard Source: https://data.unicef.org/resources/child-health-and-well-being-dashboard/ 3   Requejo J, Diaz T, Park L, Strong K, Lopez G, Borrazzo J, Costello A, Dalglish SL, Doherty T, Felici C, Gram L. Child health and wellbeing dash- boards: accountability for children’s rights. The Lancet. 2022 May 14;399(10338):1847-9. COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E C H A P T E R 17 517 DASHBOARDS IN THE ENVIRONMENT SECTOR In the environment sector dashboards are both descriptive and predictive. Dashboards, such as the Climate Action Tracker (CAT), are extensively used to call for action towards climate goals and emphasize the effects of insufficient action towards meeting climate goals set in the Paris Agreement. FIGURE 17A.6 Climate Action Tracker dashboard for Canada Source: https://climateactiontracker.org/countries/canada/ 518 C H A P T E R 17 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E The climate action tracker (CAT) An independent scientific analysis that monitors sectoral analysis to illustrate required pathways for government climate action and measures it against meeting global temperature goals. the Paris Agreement aim of “holding warming well The CAT offers a colour coded scale to rank below 2°C and pursuing efforts to limit warming to a country’s performance against global emissions 1.5°C.” Created in 2009 by Climate Analytics and goals, while also offering visualizations of historical the NewClimate Institute, the CAT quantifies and emissions and future pathways depending on poli- evaluates climate change mitigation targets, poli- cies and actions that lead to different temperature cies, and action. The dashboard aggregates country ranges derived from global least-cost models. The action to the global level and using the MAGICC dashboard also offers a descriptive synthesis of the climate model determines likely temperature in- data presented and outlines policy developments creases during the 21st century. The CAT also offers that could impact country-level ratings. DASHBOARDS IN THE DEVELOPMENT SECTOR The Sustainable Development Goals (SDGs) are key overarching global development goals emerging from prior collective goal-setting efforts including the Millennium Development Goals. The SDGs have several associated dashboards to monitor progress towards their multiple, cross-cutting, and ambitious aims. To realize these goals, it has been emphasized that decision-makers need ongoing access to data that is accurate, timely, sufficiently disaggregated, relevant, accessible, and user friendly. FIGURE 17A.7 Section of the Open SDG Data Hub Country Profile: India Source: https://unstats.un.org/sdgs/dataportal/countryprofiles/IND COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E C H A P T E R 17 519 The open SDG data hub To ensure this level of data availability and quality, To fully implement and monitor progress on the the Open SDG Data Hub makes SDG-related data Sustainable Development Goals, decision makers ev- available for exploration and analysis: https://un- erywhere need data and statistics that are accurate, stats.un.org/sdgs/dataportal. timely, sufficiently disaggregated, relevant, accessible The goal of the datahub is to enable data provid- and easy to use. The Open SDG Data Hub promotes ers, managers, and users to discover, understand, and the exploration, analysis, and use of authoritative SDG communicate patterns and interrelationships from data sources for evidence-based decision-making and the SDG data and statistics. Data is organized into advocacy. Its goal is to enable data providers, manag- SDG indicators organized by goal. The database also ers and users to discover, understand, and communi- offers SDG analytics and country profiles that allow cate patterns and interrelationships in the wealth of users to track country progress towards the 17 SDGs. SDG data and statistics that are now available. ANNEX 17.2 TABLE 17A.1 Available Data to Better Understand Population Ageing Type Example Internationally administrated surveys • WHO Surveys on Ageing and Health • Longitudinal Social Protection Survey • World Value Survey Regionally administrated surveys • Survey of Health, Ageing and Retirement in Europe • Household Finance Consumer Survey Nationally administrated surveys • Health and Retirement Study (US) • Longitudinal Study of Ageing (UK, Brazil, China, India, Ireland, Japan, Mexico, and the Republic of Korea) • Luxembourg Wealth Study Administrative data from health • Vital Statistics data and social systems • Clinical registries • Electronic health records 18 520 CHAPTER 18 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E Assessing Human Capital, Non- communicable Diseases, and Healthy Longevity in Low- and Middle-Income Countries Healthy longevity dashboard and the case for India Daphne C. Wu a, Jeremy Veillard b, Victoria Haldane c, Seemeen Saadat d, Prabhat Jha a a Centre for Global Health Research, Unity Health Toronto, University of Toronto, Canada b Health, Nutrition, and Population, Latin America and Caribbean, World Bank c Institute of Health Policy, Management and Evaluation, University of Toronto d Health, Nutrition, and Population Global Practice, World Bank COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 18 521 INTRODUCTION The global population is rapidly aging. By 2050, one in six people will be over the age of 65 (United Nations, Department of Economic and Social Affairs, Population Division 2022). An aging population comes with a com- plex set of risks to health, well-being, and productivity. Non-communicable diseases (NCDs) are one such risk that increasingly threatens the lives and livelihoods of older adults. Indeed, 74% of all deaths people annually are due to NCDs (WHO 2022a). Low- and middle-income countries (LMICs) bear a disproportionate and growing burden of NCD-related morbidity and mortality, with 77 percent of all NCD deaths occurring in LMICs (WHO 2022a). Yet, in many LMICs, healthy aging and NCDs are a nascent area of policy action. Coordinated efforts are needed globally to emphasize quired with progress monitored over time. the importance of healthy longevity on both health The HLI offers a common framework and har- and well-being. The concept of healthy longevity de- monized benchmarking approach to measuring and scribes a process of ageing that promotes human capi- monitoring healthy longevity, while also recognizing tal across the life course and where NCDs, serious dis- that differing data and information system contexts ability, and deaths are avoided in middle age, thereby across countries will inform the mix of indicators compressing morbidity into the older ages. The World and their presentation in performance dashboards Bank Healthy Longevity Initiative (HLI) sets an agenda (hereafter ‘dashboards’). Dashboards bring together that promotes human capital accumulation and pres- representations of indicators into a unified interface ervation across the life course, while also considering that uses data visualization tools to enable perfor- the linkages between these, NCDs, healthy longevity, mance monitoring across selected performance di- economic, and well-being outcomes (World Bank mensions. Here, we describe one such healthy lon- 2024). To realize such an agenda, strategic action, and gevity dashboard proposed for use by countries with focused investments, guided by local evidence, are re- moderate data infrastructure maturity. METHODS Selecting indicators Rationale for selecting HLI indicators The healthy longevity dashboard is an ongoing effort is informed in part by country context as well as by to develop and refine a suite of indicators that bring the data infrastructure maturity of the country devel- together relevant data to measure and monitor coun- oping the dashboard. As such, indicators must be se- try progress towards healthy longevity. As part of lected that not only correspond to different stages of these efforts, a common framework for healthy lon- the life course but that are also are actionable, prag- gevity and harmonized approach has been proposed matic, and comprehensive and represent a breadth of under HLI (World Bank 2024, O’Keefe and Haldane structure, process, and outcomes measures. 2024). Under this approach, indicators were selected that map to the overarching HLI conceptual frame- Indicators of context work and that can be distilled across three key actions and ten related domains to be prioritized when de- First, we selected contextual indicators. These indi- veloping a healthy longevity dashboard (Table 18.1). cators were selected to provide crucial information This approach allows us to identify indicators that on basic social, demographic, and economic char- map to data infrastructure maturity in a given coun- acteristics of a country that enables policymakers try, while ensuring comparable and consistent con- and other dashboard users to correctly interpret the ceptual underpinnings. What indicators are used to setting from which other performance-related indi- carry out these actions and represent these domains cators are drawn. 522 CHAPTER 18 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E TABLE 18.1  Key actions and related domains when developing a healthy longevity dashboard Declining human capital Intensive human Human capital accumulation, accumulation, accelerated Life course stage capital accumulation deployment, and depreciation depreciation and depletion Key action 1. Promoting enabling 2. Ensuring adequate prevention 3. Creating supportive factors for human capital and control of NCDs across the conditions for healthy accumulation life course and productive ageing Supporting • Disease prevention • NCD risk factors and behaviors • Healthy population domains • Education outcomes • NCD management • Productive ageing • Youth focus • Reducing avoidable mortality • Well-being in old age • Gender norms The indicators of context are as follows: HLI indicators i. Gross domestic product (GDP) per capita and We then used the conceptual framework to identify World Bank income category based on the 2021 healthy longevity indicators that draw from available World Bank country classification: low income, data that are contextually relevant, readily available lower-middle income, upper-middle income, or and would inform policymakers and others on key high income; elements for monitoring and progress. To identify such indicators, we conducted a review of data avail- ii. GINI index; able from international institutions, including the World Health Organization (WHO) Global Health iii. Poverty gap at $3.65 a day (2017 PPP), defined Observatory (WHO 2022c), World Bank World as the mean shortfall in income or consumption Development Indicators (World Bank 2022), Orga- from the poverty line of $3.65 a day, expressed as a nization for Economic Co-operation and Develop- percentage of the poverty line (World Bank 2022); ment (OECD) Statistics (Organization for Econom- ic Co-operation and Development (OECD) 2022), iv. Percentage of population aged 60 years and older; ILOSTAT (International Labour Organization (ILO) 2022), and UNESCO Institute for Statistics (UNES- v. Percentage of population of working age defined CO Institute for Statistics 2022). We only reviewed as ages 15 to 64; data from international institutions as this allows comparability of data across countries which is an vi. Labour force participation rate, defined as the important objective of the dashboard. We did not in- proportion of the population ages 15 and older clude data from the Institute for Health Metrics and that is economically active; Evaluation’s Global Burden of Diseases, Injuries, and Risk Factors Study (GBD), as most estimates in the vii. Ratio of female to male labor force participa- GBD are generated from complex models using in- tion rate; puts from country surveys, vital statistics, and other data sources, where available (Vos and others 2020). viii. Mean years of schooling among population aged For LMICs where data collection and availability is 25 years and over; sparse, the input sources used in the GBD model are ambiguous and the estimates generated are thus de- ix. Percentage of total deaths due to NCDs; batable. As such, we excluded data from the GBD. From the global data review, we gathered a list x. UHC coverage index, defined as the average cov- of plausible indicators that can be classified into erage of essential services based on interventions 10 broad domains pertaining to the accumulation, in reproductive, maternal, newborn and child deployment, and depletion of human capital and health, infectious diseases, NCDs and service healthy longevity, with a focus on NCDs. These capacity and access, among the general and the domains are: (i) life expectancy, (ii) mortality at dif- most disadvantaged population (WHO 2022b). ferent life stages (under five years, school age (age COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 18 523 FIGURE 18.1 Criteria and process of indicator selection 5-14 years), youths and young adults (age 15-29 d. The indicator must be easy to measure and not years), adults aged 30-69 years, and older ages (aged based on composite indicators, 70 years and above)), (iii) child health, (iv) adult immunizations, (v) risk factors, (vi) access to care e. The indicator must be based on data that are re- for NCDs and NCD management, (vii) education, ported on a regular basis at the country level and (viii) gender equality, (ix) labor force, and (x) social comparable across countries, and protection. We did not have a domain on morbidity as the only cross-country sources of morbidity data f. The indicator must have data that are based on are the WHO and the GBD, and WHO estimates are country surveys and not generated chiefly from drawn primarily from the GBD. epidemiological or econometric models. The list of plausible indicators was further streamlined using the following criteria: The steps used to select the indicators are presented in Figure 18.1. a. The indicator must be outcome-focused, The list of selected indicators along with the b. The indicator must be available for most coun- domains are presented in Table 18.2. Where sex tries, particularly for LMICs where data are usu- and income disaggregated data are available, they ally sparse, were included. This list of indicators may, however, differ by country, depending on data availability for c. The indicator must align with national or inter- the country. national goals, 524 CHAPTER 18 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E TABLE 18.2  Indicator domains and indicators in the HLI dashboard for setting with moderate data infrastructure, and data sources for the India HLI dashboard Domains Indicators Life expectancy • Life expectancy at birth • Life expectancy at age 60 • Healthy life expectancy at birth • Healthy life expectancy at age 60 Mortality at different life stages • Under-five mortality rate • School age (age 5-14 years) mortality rate • Mortality rate among youths and young adults (age 15-29 years) • Adult (aged 30-69 years) mortality rate • Mortality rate at older ages (aged 70 years and above) Child health Childhood vaccinations • BCG immunization coverage among one-year olds • Diphtheria tetanus toxoid and pertussis (DTP3) immunization coverage among one-year-olds • Polio (Pol3) immunization among one-year-olds • Measles-containing-vaccine second-dose (MCV2) immunization coverage by the nationally recommended age Healthy growth • Stunting prevalence among children under 5 years • Wasting prevalence among children under 5 years • Underweight prevalence among children under 5 years Adult immunizations • Tetanus vaccination coverage for pregnant women • COVID-19 vaccination (two doses) coverage for all ages Access to care for NCDs • Prevalence of treatment (taking medicine) for hypertension among population aged 30-79 with and NCD management hypertension NCD risk factors Tobacco use • Tobacco use prevalence among population aged 15 years and over • Cigarette smoking prevalence among population aged 15 years and over • Tobacco use prevalence among youths aged 15-24 years • Tobacco cessation, represented by the ratio of former to current smokers aged 45-59 years Harmful alcohol use • 12-month prevalence of harmful alcohol use among population aged 15+ years Obesity • Obesity prevalence among population aged 20 years and over Type 2 diabetes • Prevalence of type 2 diabetes among population aged 20-79 Hypertension • Prevalence of hypertension among population aged 30-79 years Mental health • Prevalence of current depression among population aged 18 years and over Education • Learning-adjusted years of school at age 18 Gender equality • Percentage of ever-partnered women aged 15-49 years who experienced physical and/or sexual violence by current or former partner in the previous 12 months • Percentage of women’s (aged 15-64) work time (paid and unpaid) spent on unpaid work • Ratio of female to male labor force participation rate • OECD Social Institution and Gender Index (SIGI)* Labour force • Labour force participation rate among population aged 15 years and over Social protection • Population covered by at least one social protection cash benefit (excluding health)† • Public expenditure on social protection (excluding health), as % of GDP • Public expenditure on pension, as % of GDP * The OECD Development Centre’s Social Institutions and Gender Index (SIGI) is a cross-country measure of discrimination against women that takes into account country laws, social norms, and practices around women’s rights and access to resources (including discrimination in the family, restricted physical integrity, restriction access to productive and financial resources, and restricted civil liberties for women) (OECD 2019). Although this index is not based on data reported periodically by countries, it serves an important information source on “Whether or not legal frameworks are in place to promote, enforce and monitor gender equality and women’s empowerment” (SDG Indicators 5.1.1). † Proportion of the total population receiving at least one contributory or non‑contributory cash benefit, or actively contributing to at least one social security scheme. COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 18 525 Performance score a country is calculated using the standard formula: To assess the performance of a country relative to z=((x - µ))/σ other countries with respect to an indicator, we normalized the data across countries to calculate where ‘x’ is the value for the study country, µ is the the score based on two approaches: percentile rank mean value across all countries, and ‘σ’ is the stan- approach and z-score approach. Details of the two dard deviation. approaches are described below. In both approaches, Based on the z-score, we then used the z-table to the study country is compared with other countries obtain the percentage of countries that are perform- that fall under the same income strata as the study ing below the study country. This percentage is used country, based on the 2021 World Bank country as the score for indicators where higher the value, classification (World Bank 2021), and have a popu- better the performance, such as life expectancy. For lation of more than seven million (or 0.1 percent of indicators where lower the value, better the perfor- the world population) in 2021, based on the United mance, such as mortality rate, the score is further Nations’ World Population Prospects 2022 (Unit- calculated as one minus calculated percentage. The ed Nations, Department of Economic and Social score calculated using this approach is denoted by ‘Z’. Affairs, Population Division 2022). A score of 100 Based on the scores calculated from the two ap- percent indicates best performance, 50 percent indi- proaches, we assigned the study country into one of four cates average performance, and 0 percent indicates quartiles: <25%, 25 - <50%, 50% - <75%, and ≥75%. worst performance relative to the other countries. Performance dashboard results for india Percentile rank approach Drawing on these indicators and scoring approach, In the percentile rank approach, the score for an in- we present the HLI dashboard for India as an ex- dicator is represented by its percentile rank. The per- ample of settings with moderate data infrastructure. centile rank is calculated using the standard formula: India is a lower-middle income country, accord- ing to the 2021 World Bank country classification Percentile rank=(M+(0.5*R))/Y (World Bank 2021), with a population of about 1.39 billion in 2021 (United Nations Department where ‘M’ is the number of values below the value of Economic and Social Affairs Population Division for the study country, ‘R’ is the number of values 2022). Thus, on the dashboard, India is ranked rela- equals to the value for the study country, and ‘Y’ is tive to 33 lower-middle income countries. The list of the total number of values. 33 countries is provided in Appendix Table A18.1. For indicators where a higher value indicates Appendix Table A18.2 presents the data sources for better performance, such as life expectancy and each indicator included in the India dashboard. For employment rate, the percentile rank is directly in- each indicator, we used most recently available data terpreted as the score, whereas for indicators where between 2015 and 2019. We did not use data older a lower value indicates better performance, such as than 2015 as they may not reflect the current sce- mortality and morbidity rate, the score is further nario in the country. We did not use data after 2019, calculated as 1 – percentile rank. We use ‘P’ to de- as they may not be comparable to those before the note the score calculated based on this approach. COVID-19 pandemic. Z-score approach Alignment with national and international goals While the percentile rank approach provides the performance of a country relative to all other coun- In 2015, all Member States of the United Nations tries, it does not take into account the distribution of (UN) adopted the Sustainable Development Goals the indicator across countries. To take into account (SDGs). The SDGs comprise of an ambitious set of 17 the indicator distribution, including the mean and goals and 169 targets addressing the global challenges standard deviation, a score is computed based on of poverty, inequality, climate change, environmental the z-score. This approach assumes that the indica- degradation, peace, and justice (United Nations Gen- tor values are normally distributed and a z-score for eral Assembly 2015). The HLI indicators broadly ad- 526 CHAPTER 18 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E dress a number of areas in the SDGs, including Goal life expectancy, child mortality, mortality from major 2 on zero hunger, goal 3 on good health and well-be- NCDs (cardiovascular diseases, cancer, diabetes, and ing, goal 4 on quality education, goal 5 on gender chronic respiratory diseases), child immunizations, equality, and goal 8 on decent work and economic access to medications among those with NCDs, tobac- growth (United Nations General Assembly 2015). co use, stunting, and gender-based violence (Ministry The HLI indicators chosen for this dashboard of Health and Family Welfare, Government of India are also relevant to the current 2017 Indian Nation- 2017). Given this, the HLI dashboard could serve as a al Health Policy, including universal health coverage, guide to achieving many of the policy objectives. RESULTS Context Table 18.3 presents contextual information on eco- higher proportion of population aged 60 and over nomic, social, and health indicators for India and than the average for lower-middle income countries compares them, where applicable, with the average but lower than for the world (10.2 percent in India for all lower-middle countries and the world, using vs 9.1 percent in lower-middle income countries vs most recently available data between 2015 and 2019. 13.5 percent globally). It also has a higher propor- In 2019, the GDP per capita for India sits just below tion of working age population; however, the labor the average for all lower-middle income countries at force participation rate is much lower. The ratio of US$2,047 or Int$6,887 (US$2,386 or Int$7,566 for female to male labor force participation rate and lower-middle income countries). It has an adequately mean years of schooling in India are also lower than equal income distribution with a GINI index of 35.7, that in lower-middle income countries and globally. and an estimated poverty gap measured at US$3.65 Finally, in terms of health profile and service cov- per day (2017 PPP) of 12 percent, making it compara- erage, India has a higher proportion of deaths that ble to that for lower-middle income countries (11 per- are attributable to NCDs and higher UHC coverage cent), but higher than the global average of 8 percent. compared to the average of lower-middle income In terms of demography, India has a slightly countries but lower than the global average. TABLE 18.3  Indicators of context– India, lower-middle income countries and world in 2019 Indicators of context India Lower-middle income countries World GDP per capita, in US$ (Int$) 2,047.2 2,386.5 11,320.9 (6,887.5) lower-middle income (7,566.2) (17,607.6) GINI index 35.7 - - Poverty gap at $3.65 a day (2017 PPP), as % of the poverty line 12 11 8 Percentage of population aged 60 years and older (%) 10.2 9.1 13.5 Percentage of population of working age (%) 66.9 64.9 64.2 Labour force participation rate (%) 49.3 54.4 60.2† Ratio of female to male labor force participation rate 28.9 43.6 65.8† Mean years of schooling among population aged 25 years and over 6.7 7.3 8.9 Percentage of deaths due to NCDs 65.9 62.6 73.6 UHC service coverage index * 61 58 67 * UHC (universal health coverage) service coverage index is defined as the average coverage of essential services based on interventions in reproductive, maternal, newborn and child health, infec- tious diseases, NCDs and service capacity and access, among the general and the most disadvantaged population (WHO 2022b). † Modelled ILO estimate. - N/A COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 18 527 Health indicators (Domains 1-6) Table 18.4 presents the data and scores based on per- dicators (stunting, wasting, and underweight). On centile rank approach (P) and z-score approach (Z) the other hand, where gender disaggregated data are for the HLI health indicators for India. For a majority available, we find that males score higher than females of the health indicators, P and Z are closely correlat- except in harmful alcohol use and prevalence of hy- ed. For most indicators, India falls in the categories of pertension. In other words, males in India fare better 25 - < 50% (orange) and 50 - < 75% (yellow) among than females when compared to their counterparts lower-middle income countries. India fares better on in other lower-middle income countries. Scores are obesity than most other lower-middle income coun- not available for mental health (current depression) tries but lags substantially on the healthy growth in- due to lack of comparable data from other countries. TABLE 18.4  HLI health indicators (domains 1-6) and scores based on percentile rank approach (P) and z-score approach (Z) for India Domain Indicators Both sexes Males Females P (out of Z (out of P (out of Z (out of P (out of Z (out of Value 100) 100) Value 100) 100) Value 100) 100) Life expectancy Life expectancy at birth - - - 69.5 66 68 72.2 47 55 (years) Life expectancy at age 60 - - - 18.1 84 74 19.5 66 51 (years) Healthy life expectancy at - - - 60.3 53 57 60.4 41 38 birth (years) Healthy life expectancy at - - - 13.0 69 54 13.5 16 23 age 60 (years) Mortality at Under five years (per 1,000 33 47 60 - - - - - - different life live births) stages School age (5-14 years, per 55 47 66 - - - - - - 100,000 population) Youths and young adults (15- - - - 127 63 69 113 38 37 29 years, 100,000 population) Adults aged 30-69 years (per - - - 813 59 59 595 47 51 100,000 population) Older ages (70 years and above, per 100,000 - - - 7,512 81 77 6,830 56 60 population) Child health Childhood immunizations BCG immunization coverage 85 25 41 - - - - - - among one-year olds (%) DTP3 immunization coverage 85 50 59 - - - - - - among one-year-olds (%) Polio (Pol3) immunization coverage among one-year- 85 56 61 - - - - - - olds (%) MCV2 immunization coverage by nationally 81 68 62 - - - - - - recommended age (%) Healthy growth Stunting prevalence among 34.7 8 17 - - - - - - children under 5 years Wasting prevalence among 17.3 0 0 - - - - - - children under 5 years Underweight prevalence among children under 5 33.4 0 0 - - - - - - years 528 CHAPTER 18 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E Domain Indicators Both sexes Males Females P (out of Z (out of P (out of Z (out of P (out of Z (out of Value 100) 100) Value 100) 100) Value 100) 100) Adult Tetanus vaccination coverage 90 24 48 - - - - - - immunizations for pregnant women (%) COVID-19 vaccination (two doses) coverage for all ages 67 81 83 - - - - - - (%) Access to care for Prevalence of treatment NCDs and NCD (taking medicine) for management hypertension among - - - 17 69 66 25 56 55 population aged 30-79 with hypertension (%) Risk factors Tobacco use prevalence among population aged 15 - - - 23 59 66 3 44 63 years and over (%) Cigarette smoking prevalence among - - - 10 90 85 1 59 66 population aged 15 years and over (%) Tobacco use prevalence among youths aged 15-24 - - - 12 63 73 2 50 64 years (%) Ratio of former to current - - - 0.27 16 27 0.33 22 35 smokers aged 45-59 years 12-month prevalence of harmful alcohol use among - - - 2 56 70 0.1 81 83 population aged 18 years and over (%) Obesity prevalence among population aged 20 years - - - 3 94 82 5 91 88 and over (%) Prevalence of type 2 diabetes among population aged 20- 10 25 40 - - - - - - 79 years (%) Prevalence of hypertension among population aged 30- - - - 25 41 48 24 78 74 79 years (%) Prevalence of current depression among - - - 2.4 3.0 population aged 18 years and over (%) Score of <25% Score of 25-<50% Score of 50-<75% Score of ≥75% Not available Not analyzed/not applicable Social and economic indicators (Domains 7-10) Table 18.5 shows the P and Z scores for the HLI social violence. Furthermore, it performs worse than 75 and economic indicators for India. For many of these percent of other lower-middle income countries in indicators, India falls in the 50 - < 75% category. How- female labor force participation rate and the ratio of ever, India performs worse than 50 percent of other female to male labor force participation. lower-middle income countries on intimate partner COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 18 529 TABLE 18.5  HLI social and economic indicators (domains 7-10) and scores based on percentile rank approach (P) and z-score approach (Z) for India Domain Indicators Both sexes Males Females P (out of Z (out of P (out of Z (out of P (out of Z (out of Value 100) 100) Value 100) 100) Value 100) 100) Education Learning adjusted years of school by age 18 (out - - - 7.1 76 58 57 72 57 of 14) Gender Percentage of ever-part- equality nered women aged 15-49 years who experienced physical and/or sexual - - - - - - 18 30 28 violence by current for former partner in the previous 12 months (%) Percentage of women’s (aged 15-64) work time - - - - - - 65 56 59 (paid and unpaid) spent on unpaid work (%) Ratio of female to male labor force participation - - - - - - 35 16 12 rate SIGI index - - - - - - 34 67 72 Labour force Labour force participation rate among population - - - 76 80 61 26 16 24 aged 15 years and over Social protec- Population covered tion by at least one social 24 52 51 - - - - - - protection cash benefit (excluding health)* Public expenditure on so- cial protection (excluding 1.4 - - - - - - health), as % of GDP Score of <25% Score of 25-<50% Score of 50-<75% Score of ≥75% Not available Not analyzed/not applicable * Proportion of the total population receiving at least one contributory or non‑contributory cash benefit, or actively contributing to at least one social security scheme. DISCUSSION The performance measurement literature identifies four pathways to improvement through public reporting (Berwick, James and Coye 2003; Hibbard, Stockard and Tusler 2003; Hibbard 2008): the change pathway (pro- viders of services use comparative information to improve performance); the selection pathway (health system users apply comparative information to change care consumption from poor to good performers); pay-for-per- formance (providers who achieve standards or targets receive financial rewards); and reputational damage (pro- viders who perform poorly suffer damage to their public reputation from regular public reports). These four pathways are consistent with the public sector literature on performance management, which classifies perfor- mance improvement pathways depending on their internal or external source of control and on the supportive or punitive actions derived from the controls (Boland and Fowler 2000; Veillard et al. 2005). The use of global performance dashboards as a Globally, performance dashboards have been devel- monitoring tool has been increasing in recent years oped and used by governments and multilateral or- owing to their ease of use and effectiveness in com- ganizations to monitor status and progress in global municating information about complex systems. development or for specific sectors such as health, 530 CHAPTER 18 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E education, or other fields. Numerous dashboards be relevant to many high-income settings. exist in global health, such as the SDG Dashboard The current HLI dashboard relies on data read- (Sachs and others 2023), the UNICEF Child health ily available in the public domain. As such, LMICs and well-being dashboard (UNICEF 2022), the ILO can replicate the dashboard for their countries with- World Social Protection Data Dashboards (ILO out the need for additional data collection. Howev- 2020), and the World Bank Human Capital Index, er, relying only on readily available data restricts the all performance dashboards serving specific pur- range of indicators included in the dashboard. Due poses. However, to the best of our knowledge, no to this, there are, for example currently no indicators performance dashboards have been developed with available internationally on important performance a focus on health longevity, which is defined by the dimensions relevant to human capital and healthy WHO as “the process of developing and maintaining longevity, such as adult immunizations (e.g., human the functional ability that enables wellbeing in older papillomavirus immunization among adolescent age” (WHO 2020). Here, we present an HLI dash- girls, flu immunization at all ages, and herpes zos- board that LMICs can use to monitor their status ter at older ages), ongoing education among adults, related to human capital, NCDs, and healthy longev- adult population that are free of NCDs, and public ity. The dashboard presented in this paper follows expenditure on long term care and on pensions. Al- the change pathway, providing opportunities for though performance for these indicators is expect- government to identify priorities for action derived ed to be very low in LMICs, collection of such data from international comparisons with peer countries. and regular surveillance is essential to monitor the In addition to focusing on healthy longevity, the status of human capital and healthy longevity. novelty of the HLI dashboard lies in its focus on in- Further, not all the indicators selected in the ternational comparisons. In the HLI dashboard, for performance dashboard can be calculated at na- each country and indicator, two scores are assigned tional level, limiting the understanding of variation to indicate the country’s performance relative to all in performance at a sub-national level. This limits other countries in the same income strata. These the use of data at sub-national level, in the highly scores are calculated based on the percentile rank decentralized politico-administrative context of approach and the z-score approach, respectively. India. Complementing the indicators selected with The first, percentile rank approach, has been pre- additional or substitutive indicators that can be cal- viously used in the SDG Index to compare perfor- culated at sub-national level and be made useful to mance across all countries (Schmidt-Traub and oth- decision-makers able to act on findings will be an ers 2017). The second z-score approach of scoring important next step in tailoring this performance adds an additional dimension to understanding the dashboard to the needs of policymakers in India. cross-country data by accounting for the distribu- Finally, improving the usability of the dash- tion of the indicator across countries. This z-score board for decision-making requires understanding approach of scoring has not been used by any previ- in greater detail the nature of actions by government ous global health performance dashboard, to the best and other actors (including the private sector) re- of our knowledge. The dashboard is an Excel-based quired to improve performance. This is of particular tool which does not require intensive technology or importance for indicators requiring multi-sectoral capacity training. Hence, it can be readily adopted interventions to drive better performance, such as by LMICs. Due to its focus on LMICs, some of the nutrition indicators for example. indicators, such as children underweight, may not CONCLUSION This performance dashboard offers a compelling, simple, and intuitive approach to measure performance on healthy longevity, a critical challenge to many middle-income and high-income countries, and an emerging challenge for many low-income countries. The dashboard uses existing data and is grounded their essential data infrastructure to produce per- in performance measurement systems implement- formance information of sufficient granularity to ed at scale in most countries. It should not however enable relevant decision-making at the appropriate preclude countries from investing in building up governance level, with information of high quality COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 18 531 and of relevant timeliness and granularity. In addi- public on indicators selection and general design of tion, further research is required in understanding the performance dashboard will be a condition for the drivers of healthy longevity, improving its mea- usefulness of the approach and strengthening ac- surement, and producing performance dashboards countability for better results (Veillard et al. 2015). of use to policymakers. Engaging with the general REFERENCES 1. Berwick, Donald M., Brent James, and Molly Joel Coye. 2003. 14. United Nations, Department of Economic and Social Affairs, “Connections between Quality Measurement and Improve- Population Division. World Population Prospects, Online Edition, ment.” Medical Care 41 (Supplement): I–30I–38. 2022. [Accessed September 28, 2022]. 2. Boland, Tony, and Alan Fowler. 2000. “A Systems Perspective of 15. UN General Assembly. 2015. “Transforming our world : the 2030 Performance Management in Public Sector Organisations.” In- Agenda for Sustainable Development, 21 October 2015,  A/ ternational Journal of Public Sector Management 13 (5): 417–46. RES/70/1.” https://www.refworld.org/docid/57b6e3e44.html. 3. Hibbard, Judith H., Jean Stockard, and Martin Tusler. 2003. [Accessed October 4, 2022]. “Does publicizing hospital performance stimulate quality im- 16. UNICEF. 2022. Child health and well-being dashboard. https:// provement efforts?”. Health Affairs 22 (2): 84–94. data.unicef.org/resources/child-health-and-well-being-dash- 4. Hibbard, Judith H. 2008. “What can we say about the impact board/. [Accessed October 20, 2022]. of public reporting? Inconsistent execution yields variable re- 17. Veillard, Jeremy, François Champagne, Niek S. Klazinga, Vahé sults.” Annals of Internal Medicine 148 (2): 160–61. Kazandjian, Onyebuchi A. Arah, and Ann.-Lise Guisset. 2005. “A 5. International Labour Organization (ILO). 2020. World Social Pro- Performance Assessment Framework for Hospitals: The WHO tection Data Dashboards. https://www.social-protection.org/ Regional Office for Europe PATH Project.” International Journal gimi/WSPDB.action?id=19. [Accessed October 13, 2022]. for Quality in Health Care 17 (6): 487–96. 6. International Labour Organization (ILO). 2022. ILOSTAT - The 18. Vos, Theo, Stephen S. Lim, Cristiana Abbafati, Kaja M. Abbas, leading source of labour statistics. https://ilostat.ilo.org/. [Ac- Mohammad Abbasi, Mitra Abbasifard, Mohsen Abbasi-Kange- cessed September 28, 2022]. vari, et al. 2020. “Global Burden of 369 Diseases and Injuries in 7. Ministry of Health and Family Welfare, Government of India. 204 Countries and Territories, 1990–2019: A Systematic Analy- 2017. “National Health Policy 2017.” https://www.nhp.gov.in/ sis for the Global Burden of Disease Study 2019.” The Lancet 396 nhpfiles/national_health_policy_2017.pdf. [Accessed Sep- (10258): 1204–22. tember 28, 2022]. 19. World Bank. 2021. World Bank Country and Lending Groups. 8. O’Keefe, Philip, and Victoria Haldane. 2024. “Towards a frame- https://datahelpdesk.worldbank.org/knowledgebase/arti- work for impact pathways between non-communicable dis- cles/906519-world-bank-country-and-lending-groups. [Ac- eases, human capital and healthy longevity, economic and cessed September 28, 2022]. wellbeing outcomes,” in Unlocking the Power of Healthy Lon- 20. _____. 2022. World Development Indicators. Washington, D. C.: gevity: Compendium of Research for the Healthy Longevity Initia- World Bank. https://data.worldbank.org/indicator. [Accessed tive. Washington D.C.: World Bank. September 28, 2022]. 9. Organization for Economic Co-operation and Development 21. _____.2024. Unlocking the Power of Healthy Longevity: De- (OECD). 2019. “SIGI 2019 Global Report: Transforming Challeng- mographic Change, Noncommunicable Diseases, and Human es into Opportunities, Social Institutions and Gender Index.” Capital. Washington, DC: World Bank. Paris: OECD Publishing. https://doi.org/10.1787/bc56d212-en. 22. World Health Organization (WHO). 2020. Healthy ageing and 10. _____. 2022. OECD Statistics. https://stats.oecd.org/. [Accessed functional ability. https://www.who.int/news-room/ques- September 28, 2022]. tions-and-answers/item/healthy-ageing-and-functional-abili- 11. Sachs, Jeffrey D., Guillaume Lafortune, Grayson Fuller, and ty [Accessed August 2, 2022]. Eamon Drumm. 2023. “Implementing the SDG Stimulus.” Sus- 23. _____. 2022a. Noncommunicable diseases. https://www.who. tainable Development Report 2023. Paris: SDSN, Dublin: Dublin int/news-room/fact-sheets/detail/noncommunicable-diseas- University Press. https://doi.org/10.25546/102924. es. [Accessed August 16, 2023]. 12. Schmidt-Traub, Guido, Christian Kroll, Katerina Teksoz, David 24. _____. 2022b. UHC service coverage index (3.8.1). https:// Durand-Delacre, and Jeffrey D. Sachs. 2017. “National Baselines www.who.int/data/gho/indicator-metadata-registry/imr-de- for the Sustainable Development Goals Assessed in the SDG tails/4834. [Accessed October 13, 2022]. Index and Dashboards.” Nature Geoscience 10 (8): 547–55. 25. _____. 2022c. Global Health Observatory. Geneva, Switzerland: 13. UNESCO Institute for Statistics. 2022. UIS Statistics. http://data. World Health Organization. https://www.who.int/data/gho/data/ uis.unesco.org/. [Accessed September 28, 2022]. indicators/indicators-index. [Accessed September 28, 2022]. 532 CHAPTER 18 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E ANNEX 18.1 TABLE 18A.1  List of 33 lower-middle income countries with population of more than 7 million in 2020 Algeria Ghana Myanmar Tunisia Angola Haiti Nepal Ukraine Bangladesh Honduras Nigeria Tanzania Benin India Pakistan Uzbekistan Bolivia Indonesia Papua New Guinea Viet Nam Cambodia Iran Philippines Zimbabwe Cameroon Kenya Senegal Côte d’Ivoire Lao People’s Democratic Republic Sri Lanka Egypt Morocco Tajikistan TABLE 18A.2  Indicator data sources for the HLI India dashboard Indicators Data sources GDP per capita World Bank 2022. GINI index Poverty gap at $3.65 a day (2017 PPP), as % of the poverty line Percentage of population aged 60 years and older United Nations Department of Economic and Percentage of population of working age Social Affairs Population Division 2022. Labour force participation rate World Bank 2022. Ratio of female to male labour force participation rate Mean years of schooling Global Data Lab 2013-2023; Smits and Per- manyer 2019. Percentage of total deaths due to NCDs WHO 2020. UHC service coverage index WHO 2022a. Life expectancy at birth WHO 2022b. Life expectancy at age 60 Healthy life expectancy at birth Healthy life expectancy at age 60 Under-five mortality rate (per 1,000 live births) WHO 2020. School age (age 5-14 years) mortality rate (per 100,000 population) Mortality rate among youths and young adults (age 15-29 years) (per 100,000 population) Adult (aged 30-69 years) mortality rate (per 100,000 population) Mortality rate at older ages (aged 70 years and above) (per 100,000 popula- tion) Percentage of one-year olds who received three doses of diphtheria tetanus WHO 2022b. toxoid and pertussis (DTP3) immunization. Percentage of one-year olds who received three doses of polio (Pol3) immu- nization. Percentage of one-year olds who received two doses of measles (MCV2) immunization Stunting prevalence among children under 5 years WHO 2022b. Wasting prevalence among children under 5 years Underweight prevalence among children under 5 years Percentage of mothers whose last birth was protected against neonatal WHO 2022b. tetanus Percentage of population that received at least two doses of COVID-19 Mathieu and others 2022 vaccination COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E CHAPTER 18 533 Indicators Data sources Percentage of population aged 15+ years who currently use tobacco Institute for Health Metrics and Evaluation Youth tobacco use (percentage of population aged 15-24 years who use (IHME) 2022. * tobacco) Ratio of former to current smokers aged 45-59 years Percentage of population aged 15+ years who smoke cigarettes WHO 2022b. Percentage of population aged 15+ years who practice harmful use of alco- hol in the past 12 months Percentage of population aged 20+ years who have overweight or obesity Percentage of population aged 20-79 years who have type 2 diabetes International Diabetes Federation 2021. Percentage of population aged 30-79 years who have hypertension WHO 2022b. Percentage of population aged 18 years and older who currently have Gururaj and others 2016. depressive disorders Percentage of population aged 30-79 years with hypertension that are cur- WHO 2022b. rently taking antihypertensive medication Learning-adjusted years of schooling at age 18 World Bank 2021. Prevalence of intimate partner violence among ever-partnered women in the WHO 2019. past 12 months Percentage of women’s work time (paid and unpaid) spent on unpaid work ILO 2018. Ratio of female to male labor force participation rate World Bank 2022. Social Institution and Gender Index (SIGI) OECD 2019. Labor force participation rate among population aged 15 years and over World Bank 2022. Population covered by at least one social protection cash benefit (excluding ILO 2020. health) † Public expenditure on social protection (excluding health), as % of GDP ILO 2020. * We included estimates of tobacco use from the GBD because they were generated from nationally representative surveys on tobacco use. †Proportion of the total population receiving at least one contributory or non‑contributory cash benefit, or actively contributing to at least one social security scheme. ANNEX 18.1 REFERENCES 1. Global Data Lab. Subnational HDI. The Netherlands: Global care jobs for the future of decent work”. Geneva, Switzerland: Data Lab, Institute for Management Research, Radboud Uni- International Labour Organization. versity; 2013-2023. https://globaldatalab.org/shdi/download/. 6. _____. 2020. World Social Protection Data Dashboards. https:// [Accessed April 4, 2023]. www.social-protection.org/gimi/WSPDB.action?id=19. [Ac- 2. Gururaj, G., M. Varghese, V. Benegal, G. N. Rao, K. Pathak, L. K. cessed October 13, 2022]. Singh, et al. 2016. “National Mental Health Survey of India, 7. Mathieu, Edouard, Hannah Ritchie, Esteban Ortiz-Ospina, Max 2015-16: Prevalence, patterns and outcomes.” Bengaluru, Na- Roser, Joe Hasell, Cameron Appel, Charlie Giattino, and Lucas tional Institute of Mental Health and Neurosciences, NIMHANS Rodés-Guirao. 2021. “A global database of COVID-19 vaccina- Publication No. 129. tions.” Nature Human Behaviour 5 (7): 947–53. 3. Institute for Health Metrics and Evaluation (IHME). 2022. “The 8. Organization for Economic Cooperation and Development Global Burden of Disease Study 2019 (GBD 2019) Results.” Seat- (OECD). 2019. “SIGI 2019 Global Report: Transforming Challeng- tle, Washington: IHME, University of Washington, 2022. https:// es into Opportunities, Social Institutions and Gender Index.” vizhub.healthdata.org/gbd-results/. [Accessed October 13, 2022]. Paris: OECD Publishing. https://doi.org/10.1787/bc56d212-en. 4. International Diabetes Federation. 2022. Diabetes estimates 9. Smits, Jeroen, and Iñaki Permanyer. 2019. “The Subnational (20-79 y): Age-adjusted comparative prevalence of diabetes, Human Development Database.” Scientific Data 6 (1): 190038. %. https://diabetesatlas.org/data/en/indicators/2/. [Accessed 10. United Nations, Department of Economic and Social Affairs, October 13, 2022]. Population Division. World Population Prospects, Online Edition, 5. International Labour Organization (ILO). 2018. “Care work and 2022. [Accessed September 28, 2022]. 534 CHAPTER 18 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E 11. World Bank. 2022. World Development Indicators. Washington, mortality-and-global-health-estimates/global-health-esti- D. C.: World Bank. https://data.worldbank.org/indicator. [Ac- mates-leading-causes-of-dalys. [Accessed October 13, 2022]. cessed September 28, 2022]. 14. _____. 2022a. UHC service coverage index (3.8.1) https:// 12. _____. 2021. “The Human Capital Index 2020 Update: Hu- www.who.int/data/gho/indicator-metadata-registry/imr-de- man Capital in the Time of COVID-19.” Washington, D.C.: World tails/4834. [Accessed October 13, 2022]. Bank; 2021. 15. _____. 2022b. Global Health Observatory. Geneva, Switzer- 13. World Health Organization (WHO). 2020. Global Health Es- land: World Health Organization. https://www.who.int/data/ timates 2020: Deaths by Cause, Age, Sex, by Country and by gho/data/indicators/indicators-index. [Accessed September region, 2000-2019. Geneva, Switzerland: World Health Or- 28, 2022]. ganization. https://www.who.int/data/gho/data/themes/ COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E APPENDIX 1 535 Appendix 1: Acknowledgments by Chapter Chapter 1 summarizes findings from the pre-cursor The team is especially grateful to Adriana to the Healthy Longevity Initiative - the Non-Com- Blanco (PAHO), Erica Di Ruggiero (University municable Diseases and Human Capital Research of Toronto), Daniel Dulitzky (World Bank), Sue Initiative, led by the World Bank Human Capital Horton (University of Waterloo), Dean Jamison Project, in collaboration with partners, in 2019. The (UCSF), and Rachel Nugent (RTI), who served initiative was coordinated by Sir George Alleyne as co-chairs or formal respondents during work- (Director Emeritus, Pan American Health Organi- shop sessions. Valuable comments and sugges- zation (PAHO)), Timothy Evans (McGill Universi- tions were received on earlier drafts of this report ty), Prabhat Jha (University of Toronto), and Jeremy from the following colleagues: Francisca Akala, Veillard (World Bank). This chapter was drafted by Adriana Blanco, Erica Di Ruggierro, Dean Jami- the initiative coordinators and Alexander Irwin (In- son, Aakash Mohpal, Rachel Nugent, Miriam dependent Global Health Writer and Researcher). Schneidman, and Stéphane Verguet. The authors It is based on the proceedings of the Non-Commu- would particularly like to thank the University of nicable Diseases and Human Capital Analytic Work Toronto colleagues who provided administrative, and Key Messaging Workshop, hosted in Toronto by logistical, and technical support during the July the University of Toronto Dalla Lana School of Pub- 9-10, 2019, Toronto workshop, especially Desirée lic Health on July 9-10, 2019, and sponsored by the Bernard and Yvonne Edwards. Access Accelerated Trust Fund. The team is thankful to the workshop partici- Chapter 2 summarizes the literature on the impacts pants for their contributions to the workshop and of health and longevity on growth, using macro- the discourse on healthy longevity. Authors whose economic, growth accounting, and microeconomic work was presented at the workshop included approaches and presents a framework for channels Dean T. Jamison (University of California at San through which NCDs interact with these outcomes. Francisco (UCSF)); David Watkins (University of The chapter is authored by Phillip O’Keefe (Uni- Washington); Prabhat Jha, Ryan Hum, Daphne versity of New South Wales) and Victoria Haldane Wu (University of Toronto); Aayush Khadka and (University of Toronto). Stéphane Verguet (Department of Global Health The team is grateful to Sue Horton (University and Population, Harvard T.H. Chan School of of Waterloo) and Michele Gragnolati (World Bank) Public Health); Sanam Roder-DeWan (Harvard for their peer review. The chapter also benefited University); Jean-Louise Arcand, Seoni Han, and from feedback from Sir George Alleyne (PAHO), Daniele Rinaldo (The Graduate Institute, Geneva); Paul Isenman (World Bank/OECD (retired)), Pra- Jeremy Addison Lauer (World Health Organiza- bhat Jha (University of Toronto), Beverley Essue tion (WHO)); Ojaswi Pandey and Aakash Mohpal (University of Toronto), Stephane Verguet (Harvard (World Bank). The team is also grateful for the con- T.H. Chan School of Public Health), Linda Fried tributions of Daniel Sellen (University of Toron- (Columbia University Mailman School of Public to); Francisca Akala, Daniel Dulitzky, and Miriam Health), Gabriel Demombynes (World Bank), John Schneidman (World Bank); Sarbani Chakraborty Giles (World Bank), Elena Glinskaya (World Bank), (Access Accelerated); Adriana Blanco (PAHO); Pe- Norbert Schady (World Bank), Michael Weber dro Conceicao (UNDP); Vivek Goel and Erica Di (World Bank), Jeremy Veillard (World Bank), Gise- Ruggiero (University of Toronto); Sue Horton (Uni- la Garcia (World Bank), Sameera Altuwaijri (World versity of Waterloo); Alexey Kulikov (World Health Bank), Seemeen Saadat (World Bank), and John Organization); Rachel Nugent (Research Triangle Piggott (University of New South Wales). Institute (RTI) International) at the workshop. 536 APPENDIX 1 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E Chapter 3 estimates the avoidable mortality from cuela de Negocios, Universidad Adolfo Ibáñez), Pra- non-communicable diseases by world regions be- bhat Jha (University of Toronto), William Savedoff tween 2000 and 2050. It was authored by Angela Y. (Social Insight), and Alan Fuchs (World Bank). The Chang (University of Southern Denmark), Gretch- chapter benefited from peer review feedback, which en A. Stevens (Independent Researcher), Diego S. was provided by Cornelis Peter Van Walbeek (Uni- Cardoso (Purdue University), Bochen Cao (WHO), versity of Cape Town) and Patrick Petit (Interna- and Dean T. Jamison (UCSF). Peer reviewers in- tional Monetary Fund (IMF)). Valuable inputs were clude Usha Ram (International Institute of Popula- also received from Jeremy Veillard (World Bank), tion Sciences) and Mukesh Chawla (World Bank). Kate Mandeville (World Bank), Ceren Ozer (World The authors have dedicated the chapter to Professor Bank), Danielle Bloom (World Bank), Hana Ross James W. Vaupel (1945-2022) who inspired them (University of Cape Town), Owen Smith (World and many others with his intellect, vision, and pas- Bank), and Paul Isenman (World Bank/OECD). sion for great research. The authors are also thankful to Daniel Araya for his valuable research assistance. Chapter 4 provides estimates of avoidable mortality by cause of death for 2000 to 2050. It was written Chapter 8 examines the burden of key non-commu- by Stéphane Verguet (Harvard T.H. Chan School nicable diseases for older adults ages 45 years and of Public Health), Sarah Bolongaita (Harvard T.H. above by sex and reviews evidence on main chal- Chan School of Public Health), Angela Y. Chang lenges for these populations in access to adequate (University of Southern Denmark), Diego S. Car- healthcare. The chapter was written by Seemeen doso (Purdue University), and Gretchen A. Ste- Saadat (World Bank), Meriem Boujadja (World vens (Independent Researcher). The team received Bank), and Sameera Altuwaijri (World Bank). The valuable feedback from Usha Ram (International chapter benefited from peer review inputs from Institute of Population Sciences), Dean T. Jamison Tanima Ahmed (World Bank), Amparo Gordil- (UCSF), Prabhat Jha (University of Toronto), Bev- lo-Tobar (World Bank), and Faiza Benhadid (In- erley Essue (University of Toronto), and partici- dependent Expert on Gender and Human Rights). pants at the Health Systems Research Symposium Earlier feedback from Gisela Garcia (World Bank) (Bogota, November 2022) on an earlier version of was also helpful in shaping the chapter. the chapter. Chapter 9 explores the relationship between caring Chapter 5 presents data on the rates of mortality de- for older parents and parents-in-law and the labor cline globally between 2000 and 2019. It was drafted supply of women and men aged 40 to 59 years in by Daphne C. Wu (University of Toronto), Debapri- Colombia, Indonesia, and Poland, as well as that ya Chakraborty (University of Toronto), Ryan Hum of an individual’s parents’ disability status and labor (University of Toronto), Prabhat Jha (University of supply in Egypt. It was authored by Roberta Gatti Toronto), and Dean T. Jamison (USCF). (World Bank), Daniel Halim (World Bank), Allen Hardiman (University of Illinois), and Shuqiao Sun Chapter 6 focuses on the role of behavioral science (World Bank). The chapter benefitted from peer re- in addressing non-communicable diseases. The view feedback from Tanima Ahmed (World Bank) chapter was developed by Ana Maria Rojas (World and discussions with seminar participants at the Bank), Ana Maria Munoz Boudet (World Bank), El- Healthy Longevity Initiative Workshops. len Moscoe (World Bank), Julian Jamison (Univer- sity of Exeter), and Carlos Riumallo Herl (Erasmus Chapter 10 discusses key challenges in health care University). The authors are grateful to Damien De and long-term care as populations age. Developed Walque (World Bank) and Akshar Saxena (Nanyang initially as a background paper for the World Bank’s Technological University) for their peer review; and Independent Evaluation Group’s (IEG) review on to Sir George Alleyne (PAHO) and Paul Isenman aging, it was drafted by Natalia Aranco Araújo (World Bank/OECD), for their valuable feedback. (World Bank) and Gisela M. Garcia (World Bank). The authors are very grateful to Ian Forde (World Chapter 7 reviews the evidence on role fiscal in- Bank) and Cristian Hererra (World Bank) who pro- struments in changing behaviors for better health vided valuable peer review inputs. outcomes. It was authored by Guillermo Paraje (Es- COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E APPENDIX 1 537 Chapter 11 focuses on the demand and supply of who prepared the calculations for the household long-term care in low- and middle-income coun- survey analysis, and to Patrick Hoang-Vu Eozenou tries. The chapter was developed by Elena Glinskaya (World Bank) and Marc-Francois Smitz (World (World Bank), Xiaohui Hou (World Bank), Zhan- Bank) for their guidance and access to micro-da- lian Feng (RTI International), Marco Angrisani ta from the HEFPI database. Christoph Kurowski (University of Southern California), Guadalupe (World Bank), Gonzalo Reyes (World Bank), and Suarez (Research Triangle Institute), Jigyasa Shar- Paul Isenman (World Bank/OECD) provided very ma (World Bank), Drystan Phillips (University of useful comments. Southern California) , Jenny Wilkens (University of Southern California), Jinkook Lee (University of Chapter 15 focuses on prioritizing interventions for Southern California), Yeeun Lee Yoo (University of addressing NCDs and health system investments. Southern California), Samuel Lau (University of The chapter is authored by David Watkins (Uni- Southern California), Hae Yeun Park (University of versity of Washington), Sali Ahmed (University of Southern California), and Yizhou Chen (University Washington), and Sarah Pickersgill (University of of Southern California). The chapter benefited from Washington). The chapter is based on earlier work peer review feedback from Indu Bhushan (Partner- by the authors as part of the larger Disease Control ship for Impact) and Feng Zhao (World Bank). Priorities Project, 3rd edition (DCP3), which has been extensively reviewed separately. Chapter 12 presents a review of the evidence on the associations between labour market outcomes and Chapter 16 discusses whole of government and major risk factors for non-communicable diseases. whole of society approaches for addressing NCDs It was written by Debapriya Chakraborty (Universi- bringing in experiences from across the world. The ty of Toronto), Daphne Wu (University of Toronto), chapter is authored by Ramesh Govindaraj (World and Prabhat Jha (University of Toronto). The au- Bank) and Sundararajan Srinavasa Gopalan (Inde- thors are thankful to Beverly Essue (University of pendent Researcher and former World Bank staff Toronto) for her valuable advice on the chapter and member). The authors are grateful for the contribu- to David Lightfoot (Unity Health) for his support tion of Nora Wu (World Bank), who provided ex- with literature searches. cellent research support and was responsible for the analysis of the World Bank’s portfolio of NCD proj- Chapter 13 identifies labor market and related ects, and the valuable inputs provided by Charles policies that foster productive longevity i.e., ex- Dalton (IFC) on examples of public-private collab- tending productive working lives and increasing oration on NCDs. The chapter benefited from the labor force participation and productivity among feedback from many internal and external reviewers, mature workers in L/MICs. It is authored by Sara including Sir George Alleyne (PAHO), Prabhat Jha Johansson de Silva (World Bank) and Indhira San- (University of Toronto), Sameera Maziad Altuwaijri tos (World Bank). Victoria Strokova (World Bank) (World Bank), Jeremy Veillard (World Bank), Derek peer reviewed the chapter and Gustavo Demarco Yach (Independent Global Health Consultant), Jeff (World Bank), Elena Glinskaya (World Bank), and Sturchio (Rabin Martin), Patrick Osewe (Asian De- Victoria Levin (World Bank) provided very helpful velopment Bank), Shambhu Acharya (World Health comments and suggestions on an early version. Organization), VR Muraleedharan (Indian Institute of Technology Madras), and Seemeen Saadat (World Chapter 14 discusses adequacy of pensions and Bank), who provided incisive comments. The au- other non-contributory models for ensuring access thors would also like to thank the Human Capital to healthcare for older and aging adults, particularly Project of the World Bank, and the Healthy Longev- in low- and middle-income countries. The chapter ity Initiative, for the financial support and encour- is authored by Gustavo Demarco (World Bank), Jo- agement provided for undertaking this study. hannes Koettl (World Bank), Miglena Abels (World Bank) and Andrea Petrelli (World Bank). The au- Chapter 17 discusses the importance of data mon- thors are thankful to Claudia Rodriguez Alas (World itoring and presents a common framework and Bank)and Ana Sofia Martinez Cordova (World harmonized benchmarking approach to measur- Bank) for their support on the household survey ing healthy longevity. The chapter was authored by data analysis; to Danilo Aristizabal (World Bank), Victoria Haldane (University of Toronto), Gisela M. 538 APPENDIX 1 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E Garcia (World Bank), Tahir Bockarie (University All chapters benefited from feedback from members of Warwick), Daphne Wu (University of Toronto), of the core HLI team at various stages and from par- Cristian A Herrera (World Bank), Maria Luisa La- ticipants at the Healthy Longevity Initiative Techni- torre Castro (Fundación Universitaria Juan N Cor- cal Workshop held in Mexico City in May 2022 host- pas), Debapriya Chakraborty (University of Toron- ed by Escuela de Salud Pública De México, World to), Beverly Essue (University of Toronto), Prabhat Bank, and University of Toronto, and the Healthy Jha (University of Toronto), and Jeremy Veillard Longevity Initiative’s Author Workshop held in Oc- (World Bank). tober 2022 at the World Bank in Washington D.C. A special thanks to Sir George Alleyne and Paul Isen- Chapter 18 details the data monitoring dashboard man for their thorough review of all chapters. for healthy longevity in India. The chapter was con- ceived by Daphne Wu (University of Toronto), and The findings, interpretations, and conclusions ex- Prabhat Jha (University of Toronto), with inputs pressed in this paper are solely those of the authors from Jeremy Veillard (World Bank) Victoria Hal- and do not necessarily reflect the views of the Board dane (University of Toronto), and Seemeen Saadat of Executive Directors of the World Bank or the (World Bank). governments they represent. COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E APPENDIX 2 539 Appendix 2: Workshops and Consultations 1. NCDs and Human Capital Workshop ya, Ramesh Govindaraj, Daniel Halim, Nedim Ja- ganjac, Dean Jamison, Julian Jamison, Venus Jaraba, December 6–7, 2018, in Washington D.C. Spon- Blanca Llorente Anaas, Hugo López Gatell, Claudia sored by the World Bank. Macias, Norman Maldonado, Laura Vivian Mendo- za Ardila, Ellen Moscoe, Ana Maria Munoz Bou- Chair: George Alleyne and Tim Evans det, Phillip O´Keefe, Truman Packard, Guillermo Participants: Jean-Louis Arcand, Kathryn Gilman Paraje, María Luisa Latorre Castro, Eduardo Lazca- Andrews, Zelalem Debebe, Michele Gragnola- no Ponce, Luz Myriam Reynales, Seemeen Saadat, ti, Dean Jamison, Prabhat Jha, Aart Kraay, Jeremy Belen Saenz de Miera Juarez, William Savedoff, Lauer, Aakash Mohpal, Rachel Nugent, Dena Rin- Gretchen Stevens, Florence Theodore, Angela Vega gold, Sanam Roder-DeWan, Rosa Sandoval, Miriam Landaeta, Jeremy Veillard, Stéphane Verguet, David Schneidman, Jeremy Veillard, and Stéphane Verguet Watkins, Daphne Wu, and Feng Zhao 2. Non-Communicable Diseases and Human 4. HLI Dashboards Workshop Capital Analytic Work and Key Messaging Workshop September 20–22, 2022 in Bogotá, Colombia. Spon- sored by the World Bank. July 9–10, 2019, at the Dalla Lana School of Public Health at University of Toronto. Sponsored by the Co-Chairs: Gisela Garcia and Jeremy Veillard Access Accelerated, the University of Toronto, and Participants: Sameera Altuwaijri, Debapriya the World Bank. Chakraborty, Beverley Essue, Victoria Haldane, Cristian A. Herrera, Prabhat Jha, Maria Luisa La- Co-Chairs: George Alleyne, Daniel Dulitzky, Timo- torre Castro, Seemeen Saadat, Gretchen Stevens, thy Evans, and Rachel Nugent Angela Vega, and Daphne Wu Participants: Francisca Akala, Jean-Louis Arcand, Adriana Blanco, Sarbani Chakraborty, Pedro Con- 5. Healthy Longevity Initiative Technical ceicao, Erica Di Ruggiero, Vivek Goel, Sue Horton, Workshop II Alexander Irwin, Dean Jamison, Prabhat Jha, Alex- ey Kulikov, Jeremy Lauer, Aakash Mohpal, Miri- October 26–28, 2022, in Washington D.C. Spon- am Schneidman, Daniel Sellen, Jeremy Veillard, sored by the World Bank. Stéphane Verguet, and Daphne Wu Chair: Sameera Altuwaijri 3. Healthy Longevity Initiative Technical Participants: George Alleyne, Meriem Boujadja, Workshop I Angela Chang, Debapriya Chakraborty, Gisela Gar- cia, Elena Glinskaya, Sundararajan Srinivasa Gopal- May 18–20, 2022, in Mexico City. Sponsored by Insti- an, Ramesh Govindaraj, Victoria Haldane, Daniel tuto Nacional de Salud Pública and the World Bank. Halim, Anselm Hennis, Alexander Irwin, Paul Isen- man, Prabhat Jha, Toni Joe Lebbos, Ellen Moscoe, Co-Chairs: George Alleyne, Sameera Altuwaijri, Ana Maria Munoz Boudet, Philip O’Keefe, Guiller- Michele Gragnolati, and Prabhat Jha mo Paraje, Seemeen Saadat, Gretchen Stevens, Jere- Participants: Tonatiuh Barrientos, Luis Benveniste, my Veillard, Stéphane Verguet, and David Watkins Diego Cardoso, Debapriya Chakraborty, Angela Chang, Beverly Essue, Gisela Garcia, Elena Glinska- 540 APPENDIX 2 COM PEN D IUM O F R E SE ARCH FO R T H E H E ALT H Y LO N G E V I T Y I N I T IAT I V E 6. HLI Analytic Meeting 7. Economic Value of Avoidable Mortality February 8–9, 2023, in Washington D.C. Sponsored March 16–17, 2023, in Toronto. Sponsored by the by the World Bank. World Bank and University of Toronto Chair: Sameera Altuwaijri Co-Chairs: Dean Jamison and Prabhat Jha Participants: Rythia Afkar, George Alleyne, De- Participants: George Alleyne, Sarah Bolongaita, De- bapriya Chakraborty, Daisy Demirag, Gisela Garcia, bapriya Chakraborty, Angela Chang, Ryan Hum, Anselm Hennis, Alexander Irwin, Paul Isenman, Alexander Irwin, Paul Isenman, Gretchen Stevens, Prabhat Jha, Bente Mikkelsen, Seemeen Saadat, If- Stéphane Verguet, and Daphne Wu fath Sharif, and Michael Weber The Healthy Longevity Initiative Demographic change, aging populations, and the rising burden of non-communicable diseases (NCDs) pose formidable challenges worldwide. The drastic shifts in global demographics underway include an increased population of adults, higher mortality and hospitalizations, and heightened caregiving burdens, particularly impacting women. The World Bank’s Healthy Longevity Initiative (HLI) has undertaken comprehensive analyses to offer s olutions, t urning d emographic challenges into opportunities. Key HLI recommendations focus on life course investments to improve health, reduce poverty, address gender inequity, enhance productivity, and increase overall wellbeing. A holistic, country-led approach is crucial, emphasizing the interdependence of responses to demographic shifts, pandemic threats, and climate change. Governments worldwide can prioritize life course HLI investments to control NCDs, delivering tangible benefits that grow dramatically and swiftly over time. Cost-effective measures, such as expanding low-cost treatments for common vascular diseases and cancers, can have a profound impact on households and societies. Fiscal interventions such as higher excise taxes on tobacco yield remarkably quick benefits. The HLI proposals involve increasing healthcare personnel, including essential nurses and doctors, along with upgrading primary care facilities, which yield substantial returns on investment. The HLI builds upon the remarkable achievements in reducing childhood, maternal and infectious disease mortality. The HLI emphasis on research and development, coupled with global public goods, provides a means to “bend the cost curve” for NCDs. At its core, the report underscores the imperative for country-driven initiatives to invest in Healthy Longevity. This presents a viable route to alleviate poverty and elevate well-being, utilizing the strength of the entire life course. The life-course approach can increase the impact of human capital, enabling progress and prosperity for all societies.