Report No. 8382-UNI Federal Republic of Nigeria Health Care Cost, Financing and Utilization (In Two Volumes) Volume l: Subsector Report October 18, 1991 Western Africa Department Population and Human Resources Operaticns Division FOR OFFICIAL USE ONLY e~ ' ' . . - , . > >.. , C, j)cuen ..ab WM; .~n -. .. .z * . ., ., . , . Theis ~cument f QaS Rtstriaed distribhution and rnay be used $y recipients. ° .oi1y in the p,erformance of their official dutes. Its contents may not othenwise ' b5w 4ido'c .X40.b ut Wortd Bank a horization. ,...................... .'...... .. CURRENCY EQUIVALENTS US$1.00 = N 8.74 (May 1991) ABBREVIATIONS AND ACRONYMS AHCS African Health Consultancy Services AVSC - Association for Voluntary Surgical Contraception BHSS Basic Health Services Scheme CEDPA Center for Education Development and Population Activities CHW Community Health Workers CIDA Caniadian International Developmenc. Agency DAE = Directorate of Adminstration and Finance CPI = Consumer Price Index DHPRS Department of Nptional Health Planning, Research and Statistics DNHP = Directorate of National Health Planning DPT = Diphtheria, Pertussis and Tetanus DRF Drug Revolving Fund ED = Essential Drug EDF = European Development Fund EDP = Essential Drugs Program EEC = European Economic Community EPI Expanded Program on Immunization FeHI = Family Health International FMFED = Federal Ministry of Finance and Econumic Development FMOH = Federal Ministry of Health FP Family Planning FPIA = Family Planning International Assistance HMD Hospital Management Board IDRC = International Development Research Center JHUPCS = john Hopkins University Population Communication Services JHUPIE = John Hopkins University Program for International Education LDC = Less Developed Country LGA = Local Government Authority MCH = Maternal and Child Health MFED Ministry of Finance and Economic Development MOH = Ministry of Health NEDP = Nigerian Essential Drugs Program NGO = Non-Governmental Organization NPB = National Population Bureau NPC = National Population Commission ODA - Overseas Develcpment Administration PHe = Primary Health Care PHCCU = Primary Health Care Coordinating Unit RMN World Bank Resident Mission, Nigeria SHMB = State Health Management Board SIDA = Swedish International Development Authority SMOH = State Ministry of Health TFR = Total Fertility Rate UNDP United Nations Development Program UNFPA = United Nations Fund for Population Activities UNICEF = United Nations Children's Fund USAID = U.S. Agency for International Development WHO = World Health Organization FISCAL YEAR January 1 - December 31 FOR OFFICIAL USE ONLY NIGERIA: HEALTH CARE COST, FINANCING AND UTILIZATION TABLE OF CONTENTS Page No. EXECUTIVE SUMMARY .... . . . . . . . . . . . . . . . . . . . . . . . i-vi I. RECENT PERFORMANCE OF THE NIGERIAN HEALTH SECTOR . . . . . . . . 1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 1 A. Population . . . . . . . . . . . . . . . . . . . . . . . . 2 B. Health . . . . . . . . . . . . . . . . . . . . . . . . . . 4 C. Utilization of Health Services . . . . . . . . . . . . . . 6 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . 12 II. INSTITUTIONAL FRAMEWORK AND POLICY DEVELOPMENT IN THE NIGERIAN PUBLIC HEA.TH SECTOR . . . . . . . . . . . . . 13 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 13 A. Background on National Health Policy . . . . . . . . . . . 13 B. Organization of Public Sector Health Care . . . . . . . . . 15 C. The Budget Process .16 D. Health Service Facilities . . . . . . . . . . . . . . . . . 17 E. Recent Developments in Health Planning . . . . . . . . . . 20 F. The Nigeria Essential Drugs Program . . . . . . . . . . . 20 G. Government Population Pol-cies and Programs .24 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 III. PUBLIC SECTOR HEALTH RESOURCES ... . . . . .... . . . . . . 26 A. The Flow of Health Sector Funds . . . . . . . . . . . . . . 27 B. Total Public Expenditure .30 C. Recurrent Expenditure . . . . . . . . . . . . . . . . . . . 33 D. Capital Expenditure . . . . . . . . . . . . . . . . . . . 37 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . 38 TV. HEALTH CARE IN O3UN STATE . . . . . . . . . . . . . . . . . . . 40 A. Description of Ogun State .40 B. State Financial Resources for Health . . . . . . . . . . . 45 C. Modern Health Services in Ogun State . . . . . . . . . . . 47 Conclusion ..56 This report wao prepared by Hazol Denton (senior project economist), with Mead Over (economist) as lead advisor who also wrote Annex E and developed the analysis In Chapter III. Chapter V: Part A was authored by Professor Annenarie Woutere (Johns Hopkins Univeraity); Part B by Professors John Akin and DOvid Guilkey (University of North Carolina); and Part C by Professor Ronald Vogel (University of Arizona). Annex V was prepared by Mrs. K. L. Mojidi (population and human resource specialist). The report drves extensively on the survey work of African Health Consultancy Services Ltd. Lagos (AHCS). Preparation of Chapter V was supported by funding from the World Bank Research Comittee. The report was edited by Richard Sacks. Citation of the report should be: by Hazel Denton, John Akin, David GuiIkey, Ronald Vogel and Annemarie Wouters. The authors particularly wish to express their appreciation to the Federal Ministry of Health for making available extensive date, plus providing acceas to the AHCS survey results. This document has a restricted distribution and may be used by recipients only in the performance of their official duties. Its contents may not otherwise be disclosed without World Bank authorization. Page No. V HEALTH CARE COSTS IN OGUN STATE . . . . . . . . . . . . . . . . 57 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 57 A. Facility Analysis .... . . . . . . . . ....... . 57 1. Financial Analysis of Public and Private Facilities, 1984-1986 . . . . . . . . . . . . . . . 60 2. Statistical Analysis of Efficiency in Public and Private Facilities: The Frontier Production Function 69 Cost Analysis Results . . . . . . . . . . . . . . . 71 B. Household Demand for Outpatient Health Se.vices . . . . . 74 1. Surveys and Samples Selected . . . . . . . . . . . . 75 2. The Socioeconomic Situation in Ogun State . . . . . . 76 (a) Demand Analysis ..80 (b) Estimation Results . . . . . . . . . . . . . . . 81 C. Health Care Cost-Recovery Simulations . . . . . . . . . . 84 (a) Choosing among 19 Policy Options . . . . . . . . 90 (b) The Administrative Costs of a Cost-Recovery Program 93 (c) Cost Analysis Results . . . . . . . . . . . . . . 93 Conclusion.. 95 Appeldix . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97 VI. SUMMARY AND CONCLUSIONS . . . . . . . . . . . . . . . . . . . . 106 Introduction ... ..106 A. Health System.. . .. 106 B. Health Status... .. 107 C. Health Problems ... .107 D. National Health Policy .....108 E. Financing . . . . . . . . . . . . . . . . . . . . . . . . . . 108 F. Ogun State ....109 G. The Analysis ...... . . .. 110 1. Households . . . . . . . . . . . . . . . . . . . . . . 111 (a) Data . . . . . . . . . . . . . . . . . . . . . . .111 (b) Methodology . . . . . . . . . . . . . . . . . . . 112 (c) Results .... . . . . . . . . ....... . . 112 2. Facilities .... . . .. . . . . . . . . . . . . . . 114 (a) Data .... . . . . . . . . . . ....... . . 114 (b) Analysis and Results . . . . . . . . . . . . . . . 115 3. Strategies for Cost Recovery . . . . . . . . . . . . . 116 4. Recommendations .. .117 REFERENCES MAPs IBRD 22161R NIGERIA: HEALTH CARE COST, FINANCING AND UTILIZATION NIGERIA HEALTH SECTOR REPORT Executive Summary A. Introduction 1. Nigeria has a commendably comprehensive National Health Policy which recognizes that "Primary Health Care is the key to attaining the goal of health for all the people." However, Nigeria's public health care services have serious shortcomings. Ironically, demana is decreasing while the number of potential users is rising. Staff do not adequately e-erve rural areas; facilities suffer from poor maintenance; and most public sector health care providers have few drugs for their patients. The report examines these problems in the context of public sector spending while noting that private health expenditure is also significantt individuals spend money on pharmaceuticals and on traditional healers, and in addition, the private sector is active at primary, secondary and tertiary levels. 2. The comparative efficiency of public and private facilities is examined in depth, using data from Ogun State. Against this background, the houselhold demand for health care in Ogun State, from both private and public sources, is analyzed. Opportunities for policy adjustment, including cost recovery, are then developed. The report concludes that although the basic components may be in place for an effective nealth care program, in Ogun State specifically and in Nigeria in general, adjusting these components could provide a more balanced and responsive program. To achieve its goal of 'Health for All by the Year 2000" the Government will need to recognize the complementary role of the private sector and insure that lin.,ted public sector health resources are used as efficiently as possible. 3. This study concentrates on Ogun State and treats it as a paradigm for the Nigerian health system. This stems from the fact that the Federal Ministry of Health has given the Bank access to the results of an unusually thorough and comprehensive survey of health care facilities and consumers that took place in the State in 1987. It must, however, be stressed that this study analyzes data from only one state. The findings may be applicable only to the extent Ogun State is similar to other regions. Nevertheless, the results highlight general areas for concern for other states to consider. B. Health System 4. Nigeria s health system has made great strides in improving access to health facilities. Since Independence in 1960, and especially since 1975, Nigeria h&s trained thousands of doctors, nurses, midwives, medical technicians, pharmacists, dentists and community health workers, and has rapidly increased the number of hospital beds. The resulting - ii - ratios of health professionals to population compa.;e very favorably with ratios in countries with per capita national incomes similar to Nigeria: the comparative data (for 1984) on population per nursing person in Sub- Saharan Africa show Nigeria had about three times the number of physicians as the regional average, and twice the number of nursing staff. 5. Many states have an adequate network of dispensaries, clinics, health centers, maternities, and hospitals of various sizes, at least in the urban areas. Many of these facilities are relatively new and have a sufficient conplement of equipment, although not necessarily in a good state of repair. In addition, people in many parts of the country can choose between private or publicly run health facilities. But the health status of the populatior is not as high as it could be in relation to investment in the health sector. 6. Health Status. Nigeria's health status indicators place it on a par with many of its poorer -eighbors. For example: (a) Infant mortality in 1988 was higher in Nigeria (103 per 1,000 live births) than in Togo (92) and Ghana (88). (b) One quarter of Nigerian babies had low birth weights in 1985, a proportion notably higher tlEan in Ghana (17 percent), Togo (20 percent), Benin (10 percent), and Niger (20 percent). (c) Life expectancy at birth for Nigerians (51 years in 1988) was average for Sub-Saharan Africa and shorter than life expectancy in Togo (53), Liberia (54), and Ghana (54). 7. The need for greater emphasis on preventive care is borne out by the figures. As in many sub-Saharan African countries, an overwhelming number of deaths in Nigeria are due to preventable infectious and parasitic diseases, in particular measles, malaria, pneumonia, and diarrhea. Childbearing women are especially at risk in Nigeria because of frequent pregnancies and closely spaced births starting at a very young age, and because of medically inadequate and hygienically inappropriate child delivery techniques. In addition, most mothers receive no form of modern pre- or postnatal care at all. 8. Underutilization. As stated above, the principal challenge is to improve the ability to provide useful medical services and thus increase the number of satisfied users. Ironically, demand for health services seems to be decreasing even as population grows. Some elements that contribute to underutilization in Nigeria's public health sector include: (a) the predominance of large, curative, hospital-based programs as opposed to smaller, more dispersed, preventive care programs; (b) poor management, with a fragmented system that lacks coordination and provides a weak referral network; and (c) lack of supplies and functioning equipment and, above all, a lack of drugs. - iii - 9. Financing. One underlying reality that exacerbates Nigeria's health care problems is the decreasing evailability of public funds at the federal, state and local levels. State and local spending in Nigeria is highly dependent on federal aid, but the Federal Government is itself highly dependent on oil revenue The downturn in the oil market in the early 1980s seriously reduced the resources available for public spending. The allocation to the public health sector has fallen sharply: the actual federal health budget (recurrent and capital) in 1990 was only 40 percent in real value that of 1981. By comparison, the total actual federal budget in 1990 was over 90Z of the real value of 1981, indicating that allocations to the health sector have been :ut back relative to other parts of the government. 10. The squeeze is felt through the system, with the fall in oil revenue in the middle 1980s. Examination of the state and local data for the 1980s shows that recurrent expenditure on items such as maintenance, medical supplies and drugs has been cut back to the barest minimum, while personnel costs have increasingly dominated recurrent outlays. 11. The drop in oil revenues helps explain the fiscal crisis in Nigeria's public health sector at a macro level. At the micro level, the public health sector faces a lowt allocation of resources and an inadequate system of cost recovery. Short of cash, the system provides services of poor quality. C. Findings 12. Households. Using a sample of 4,898 household interviews with over 22,000 individuals and of 1,763 adults who reported an illness in the four weeks prior to the survey, t'ie study analyzes how people sought treatment. For the first treatment, 68 percent headed to a public facility and 23 percent to a private one. The balance sought treatment at a pharmacy, or from a traditional healer or spiritualist. Twenty-niine percent of the total sample of ill people (including children and pregnant women) did not seek treatment of any kind. 13. The majority of the lowest income group seek health from public facilities while the highest income group makes more use of private facilities. But the more wealthy groups do not avoid public facilities: over half of the highest income groups who used a facility used a public fac4lity. As expected, of all income groups, self treatment is greatest in the lowest income groups. 14. Although price is significant in selecting health care, it is of small magnitude. Of more significance is the perceived quality of the facility. Measures of quality used were limited, but provide good indicators of basic issues: drug availability, physical condition of the facility, and investment on provision of care. The first two were highly significant, and the third significant at a low level. It appears that higher prices can be offset by higher quality: if revenues from higher prices at public facilities are used to increase care quality, especially with respect to availability of drugs and quality of the facilities, usage - iv - of these facilities (and of modern health care in general) might we'l increase. 15. Facilities. An invertory of the 445 modern health care facilities in Ogun State was undertaken. A detailed survey was made of 86 of these facilities to collect information on utilization, costs, revenues and cost-recovery programs. This survey included public and private facilities. It was found that for both inpatient and outpatient services facilities are operating with marginal cost less than average cost, especially for outpatient services, highlighting the opportunities for increasing returns to scale for admissions and nearly constant returns to scale for visits. The fact that even private sector providers were not achieving a minimum acceptable technical efficiency of 630 visits per health worker per year was surprising. The allocative analysis showed that public sector facilities are not using cost-minimizing combinations of health workers and support staff; the relative productivity of non-health workers is less than their relative wage, suggesting health care providers should re-examine their use of such staff to ensure that they are fully and optimally emple--rd. 16. The facility analysis concluded that costs in the public sector could be reduced: by changing the relative number of support staff and/or increasing the number of health workers (to the extent that the various types of staff are fungible); by eliminating underutilization of inpatient services by reducing the number of beds (possibly by privatizaLion); and by increasing demand (by improving the quality of care). 17. Strategies for Cost Recovery. The results from assessing what people do when sick, ana how much it costs to provide health care facilities were combined in simulations of possible cost-recovery strategies. The results show the estimated relative change in magnitude and direction, rather than the actual net loss or surplus. The results are illuminating for policymakers, showing that if public sector prices are raised, the probability of seeking public sector care changes only margiaally depending on the response of the private sector, but the revenue impact is large. This increase in revenue, if invested in the facilities, could then result in an increase in public sector visits. However, price increases which are not subsequently or concurrently rewarded with quality improvements, are likely to result in a loss of demand that could overwhelm the revenue expectations. Handling cost recovery and investment effectively will require both commitment and good management. 18. These results highlight the major role that perceived facility quality plays in the decisions which patients make. But quality change can have very large cost implications, which could swamp revenue effects. Therefore the specific quality improvements need to be carefully selected, choosing those which are most cost-effective, providing the greatest demand response for the least anount of investment. Making this selection efficiently will require adequate information systems at the facility level. D. Recommendations 19. The analysis in this report of data from Ogun State suggests that the Federal Government of Nigeria could usefully consider the following actions/reforms for the public health sector: (a) Improve cost recovery through higher fees consistently applied. The analysis suggests that higher fees will not significantly discourage potential users. On the contrary, assuming that more funds are available to purchase supplies, maintain equipment, and to build an adequate inventory of drugs, usage will likely remain at present levels or iiiW;eae despite the higher fees. (b) Protection of the very poor. Even if usage of needed health care by the poor were not greatly affected by fee increases, the financial burden on the poor of payirng the fees would be important. Any mrove to increase charges for health care should L be accompanied by measures designed to reduce the financial impact of such charges on the poorest people. The most obvious approach would be directly to identify those who are poor and either to exempt them from charges or to subsidize them. (c) Encourage fee collection through retention at site. Enabling facilities to retain at least part of their fees at site (to use as investment funds and/or incentive payments) would provide flexibility to local managers and opportunities for improved responsiveness. (d) Ensure an adequate supply of drugs, funded through sustainable drug revolving funds on a cash-and-carry basis. Given the Uigh rate of 'self-medication" reported, it is recommended that efforts to provide safe and affordable drugs through public health facilities be strengthened and monitoring of public outlets be improved. A sustainable drug revolving fund needs a transparent payment system, with reimbursement for the poor rather than exemptions at the door. (e) Strengthen the referral system through clear price signaling. Patients should be encouraged to use the primary health care outlets first, paying less there than for treatment in the emergency room of a hospital. (f) Reduce share of recurrent costs spent on personnel. Personnel costs, which absorb up to 95 percent of recurrent budgets, need to be reduced in order to release funds for drugs, medical supplies and maintenance. The analysis indicates that the ratio of non-health workers to health workers is higher in public than in private hospitals, suggesting adjustments could be made here without hurting efficiency. Also, given the urban bias of staff, redeployment could provide better service to the currently underserved rural areas. - vi - (g) Coordinate public and private health sector supply. In both pcrsonnel and facilities, coordination could be improved to the general benefit of the public. Recognizing that many health personnel, especially physicians, are already working in both the public and private sectors, this practice should be regularized. With respect to facilities, rather than duplicating the provision of services, the public sector could concentrate more on preventive care and the private on curative care, which already appear to be their self-selected directions. (h) Establish effective management information systems. In order to implement, monitor and evaluate adjustments in health policy, basic data needs to be regularly collected and analyzed. E. Postscript 20. This report analyzes opportunities for reform, reviews the possible consequences, and confirms the general direction in which the Federal Ministry of Health has already moved. The report focuses on the situation in the mid-1980s. Since then, the National Health Policy has been formulated, and the public health budget has increasingly been reoriented toward primary health care. The Model Primary Health Care Program has been introduced, under which each LGA receives a one-time grant to upgrade its services. Personnel training has been scrutinized and some initial efforts at manpower planning have been undertaken. States have been assisted in the preparation of detailed health sector budgets to aid in financial planning. All these moves are in accord with the findings of this report. It is hoped that this analysis will assist policymakers in their continued efforts toward achicving "Health for All" through an effective public health care system. Additional resources--from consumers, from government, and from external agencies--can also assist in making these improvements. atexecsum (HD-NIRl) NIGERIA: HEALTH CARE COST, FINANCING AND UTILIZATION L. RECENT PERFORMANCE OF THE NIGERIAN HEALTH SECTOR Introduction 1.01 Mealth conditions in Nigeria improved in the decades following Indelendence in 1960, but they remain at a low level. While infant mortality and child death rates have been reduced, they are still high compared with countries at similar levels of development. Nigeria's public and private health care systems expanded rapidly during the oil boom of the late 1970s, bringing big improvements in the ratio of health facilities and personnel to population. But sharp cutbacks in public spending since 1981 (following the end of the oil boom) have caused a deterioration in the physical condition of health facilities and their utilization rates. 1.02 Aside from lowered public spending, Nigeria's health care system suffers from inefficient use of resources and from urban bias. Budget problems may be contributing to inefficiency. Since shrinking budgets are widely skewed in favor of personnel costs, --ewer resources are available for drugs, maintenance, equipment and supplies. When facilities lack supplies and drugs, people stay away. Lowered public spending also may abet the health system's urban bias. Private health providers tend to invest in cities, where potential patients are more numerous and profits are higher. If the government fails to build health facilities in rural areas, almost no one else does. Many of these problems are typical of developing countries. This report will attempt to place them in their Nigerian context and, it is hoped, contribute to an understanding of them as they arise in other African countries. Using state-level datal/ of household demand for health care and of facilities supplying health care, the issues are analyzed in derth to identify policy adjustments to address these problems. 1.03 The mat-srials used in preparing this report focus on the period of the mid-1980s. Since then the National Health Policy haa been formulated (1988) and the public health budget increasingly reoriented toward primary health care. States have been assisted in the preparation of detailed health sector budgets to aid in financial planning. These moves are in accord with the findings of t..is report, which highlight the need to continue such adjustments to meet the goal of 'Health for All by the Year 2000.f 11 The comprehensive survey work was undertaken in Ogun State at the request of the Federal Ministry of Health (FMOH) by African Health ronsultancy Services Ltd. (AHCS) of Lagos, who designed and 4auinistered the questionnaires and synthesized the results in three volumes. The analysis in this report is built upon the work of AHCS, which the FMOH made available to the World Bank. -2- A. Populatior. 1.04 Nigeria's large and rapidly growing population of 117.2 million2/ is complicating the country's goal of providing health care to all its citizens. Nigeria is Africa's most populous country and ra.iks among the ten most populous nations of the world.3/ Annual population growth is currently estimated at about 3.1 percent, up from 2.5 percent during 1965-73 and 2.7 percent during 1973-83. Ihe 1981-82 Nigeria Fertility Survey estimated the total fertility rate (TFR) at 6.3.4/ Table 1.1 summarizes Nigerian demographic and socio-economic indicators compared with some other countries. 1.05 High fertility and declining mortality have produced a very young population and a high dependency ratio. About 48 percent of Nigeria's population is less than 15 years old. Nigeria's dependency ratio (which compares the under-15 and over-64 population to the working-age population) is estimated at 0.98. This figure indicates that every working adult must support, educate, and provide health care for one other person. Declining death rates enable more and more children to survive to reproductive ages, building a wider base for further acceleration of population growth. 1.06 Population Prp,ections. Without strong measures to curb fertility, Nigeria's population will likely double by the year 2010. Conditions in Nigeria conducive to rapid population growth include high fertility rates, strong pronatalist attitudes, a low level of contraceptive knowledge and use, decreases in the duration of both breastfeading and postpartum sexual abstinence among urban and educated' -omen, and an increasing proportion of young people in the population. 1.07 Three alternative population projections for Nigeria up to the year 2320 appear in Table i.2. The "slow decline" projection assumes that fertility will fall slower than in the standzrd projection, with fertility declining to replacement levels by 2070. The "standard decline" piojection assumes that fertility will begin to fall in 1985-1990, then cont-ilue to 2/ 1990 staff estimates. 3/ The land area of Nigeria is 924,000 sq.km.; population density averages around 127 persons per sq.km. 4/ This figure, while lower than the Bank's 1988 estimate of 6.6, is still very high. Two countries with per capita incomes close to Nigeria's (Ghana and Sudan) have similar estimated total fertility rates (6.4). Preliminary data from the 1990 Nigeria Demographic and Health Survey estimates the TFR at 6.0. In Nigeria, regional fertility differences are more important than urban and rural differences. Table 1.1 COMPARATIVE INDICATORS OF POPULATION, HEALTH AND NUTRITION STATUS FOR NIGERIA AND SELECTED COUNTRIES (1988 Data Unleos Statoa Otherwise) Average Caloric Annual Rate Risk of Daily per Supply CNP Total of Population Life Crude Infant Dying Capita As Percent Per Capita Fortillty Growth Expectancy Death Mortality by Ago S Caloric of Total Country (USS) Rate (1980-88) at Birth Rato Rate Fem. dale Supply a/ Requirement ----------- -- -- --- -- -- --- -- -- --- …-…-- - - - - - - - - - - - - - - - - NIGERIA 290 6.8 b/ 3.8 61 16 103 161 170 214S 94 Other Ang lophone African Countr es: Ghana 400 6.3 3.4 64 13 88 126 144 1769 77 Kenya 370 6.9 3.8 69 11 70 96 110 2060 90 Lower-middle-Income Countries 1S80 8.9 2.3 86 8 67 84 75 2733 120 Other Countries with Similar Per Capita Income: Zambia 290 6.7 3.7 63 13 78 108 124 Uganda 280 7.3 3.2 48 17 101 147 187 2344 103 Niger 300 7.1 3.6 46 20 133 204 206 2432 108 Pakistan 360 6.8 3.2 56 13 107 139 128 2316 101 India 840 4.2 2.2 58 11 97 118 120 2238 98 Sub -Saharan Africa 330 6.7 3.2 61 16 108 164 172 2098 92 Soureo: OWorld Development Report 1990; PovertyO. Oxford University Press, New York, 19m0. Note: .- = Not available. o/ Data are for 1988. 2,286 calorlos are considered 100%. b/ Estimated at 8.3 In Nigerian Fertility Starvey, 1981-82. c:nir (Jo) , . , l |~U - 4 - fall by the average rate for countries that have already undergone the transition to lower fertility, finally reaching raplacement levels by 2040. The third projection assumes a "rapid decline" in fertility, starting in 1985-1990, to replacement levels by 2015; this projection generates a TFR of about 4 in 2000, and therefore represents Nigeria's course of fertility decline under its declared policy target. Nigeria's youthful age struct- produces a momentum that will insure a population of more than 150 millic, by 2000 even using the most optimistic projection of fertility decline (the third scenario). Future Nigerians may have to deal with a population of 380 million by 2025 (assuming "slow" decline) or with a population of 224 million (assuming "rapid' decline). Table 1.2 NIGERIA: POPL&ATION PROJECTIONS 1986-2026 (In MillIons) Slow Fertility Decline a/ Standard Fertility Decline b/ Rapid Fertility Decline cV Growth Growth Growth Pop. Rate TFR d/ Pop. Rate TFR d/ Pop. Rato TFR d/ 1986 99.7 99.7 99.7 8.2 6.7 a.2 6.7 8.2 6.7 1990 117.2 117.2 117.2 3.4 6.0 8.1 6.2 2.7 5.5 1995 136.6 188.6 134.2 8.6 6.8 8.0 5.8 2.8 4.5 2000 166.8 169.0 160.8 3.6 f.6 8.0 6.8 2.1 8.7 2006 198.0 184.6 167.2 8.4 6.1 2.8 4.8 1.8 8.0 2010 284.7 212.8 182.8 8.8 6.8 2.6 4.8 1.6 2.4 2016 278.9 241.6 198.4 8.2 5.4 2.4 8.8 1.4 2.8 2020 826.6 271.7 210.1 8.2 6.0 2.1 8.8 1.3 2.2 2026 881.4 801.7 224.0 Source: Africa ReaIon POpulation ProlectionO. 1990-91 Edition, Working Paper WPS 598, World Bank. a/ Slow Fertility Decline: TFR slowly declines, starting 1985-90, to replement lovel (net reproduction rate = 1) by 2070. b/ Standard Fertility Doelino: TFR deelines beginning lmmdiately (1986-90) to replacement levol by 2040. Prediction based on hlstory of fertility decline In countries that have experioncod the transition. c/ Rapid Fertility Decline: TFR deelines beginning 1985-90 to replacement level by 2016. d/ Total Fertility Ratas (TFR) and growth rates aro for the quinquennium ending In tho year identifled. B. Health 1.08 Health Status Indicators -- Mortality and Morbidity. Nigerian fertility rates have stayed high while mortality rates have declined. The crude death rate was almost halved during 196.¶-87, from 27 to 15 per 1,000. Overall life expectancy at birth is around 51 years, up from 39 years in 1960. This figure is 14 years less than the average for lower-middle income countries but the same as that for Sub-Saharan Africa and closer to the levels prevailing among Nigeria's poorer neighbors, such as Benin (51 years), Togo (53 years) and Niger (45 years). Infant mortality is estimated at 105 per 1,000 live births (down from 140 in 1970), close to the average of 109 for Sub-Saharan Africa5/. In addition to poor health and inadequate coverage by the health services, the high fertility levels contribute to the observed high maternal mortality rate of 1,500 per 100,000 live births. 6/ Decreases in mortality will likely continue as medical and public health services expand in rural areas, where health facilities have historically been lacking and where about 70 percent of the population resides. 1.09 Four elements help explain these high mortality rates: (a) The toll of infectious and parasitic diseases is high: about 80 percent of such deaths are from preventable diseases.7/ (b) Probably fewer than 30 percent of pregnant women receive any form of modern pre- and postnatal care. (c) Traditional practices, particularly for child delivery, are often medically inadequate and unhygienic. (d) Frequent pregnancies and closely spaced births starting at very young ages damage the health and nutritional status of mothers and infants. 1.10 Nutrition. The already high levels of acute malnutrition, coupled with recent increases in food prices, are expanding the population at risk. This is a serious and growing problem in Nigeria. The data suggest that malnutrition is most prevalent among infants and young children. A 1984 survey found approximately one fifth of children under five with weight-for- height at least two standard deviations below the median of a reference population -- a high proportion in comparison to the past and to regional standards.8/ Similarly 25 percent of all babies have low birth weights (less than 2500 grams).9/ Such widespread malnutrition in an environment of high population growth, infection and illiteracy has serious implications for 5/ World Development Report, 1990. 6/ Federal Ministry of Health. 7/ The principal diseases are measles, malaria, pneumonia and diarrhea. Other common infectious diseases include gonorrhea, pertussis (whooping cough), schistosomiasis, chicken pox, and tuberculosis. All are diseases that could be readily treated by primary health care providers. (Data from the FMOH, Dept. of National Health Planning, Research and Statistics.) 8/ Nigerian Integrated Survey of Households, Federal Office of Statistics, 1984. 9/ World Development Report, 1990. - 6 - the survival of chi'dren. However, to the extent that the foods selected are nutritionally poor and intrahousehold food distribution favors adult males, some improvement could be achieved through education in nutrition. C. Utilization of Health Services 1.11 Service Quality and Lack of Drugs. Nigeria's health delivery system consists of an extensive network of relatively well-staffed public and private facilities, with impressive expansion since Independence. About 80 percent of these facilities are in the public sector; the private facilities are concentrated in the South. Approximately 75 percent of health facilities are in urban areas, serving about 30 percent of the population. The paramount issue facing the health sector is the declining quality and use of public health services. This situation stems from the financial constraints imposed on public spending by the decline in Nigeria's oil revenues. While the system expanded in the late 1970s and early 1980s, federal health expenditure fell by 60 percent during 1981-90 in real terms (based on actual budgets). Sharp declines are also thought to have occurred at the state and local levels. Because the cuts initially led to cancellations of capital projects, the impact on health services was limited. As the cuts became more severe, personnel costs, which tend to be inflexible, began to claim an increasing and disproportionate bLAgetary share.10/ 1.12 Drugs. Shrinking budgets also caused deep cuts in operational spending, especially for drugs. An adequate supply of appropriate drugs is essential to any health service delivery system. Public sector health facilities in particular have persistently low drug stocks, and commercial pharmacies often have very limited supplies. Lack of drugs reduces the productivity of he%lth personnel. Wlen the public becomes aware that public health facilities have no drugs, people bypass their services and attempt to obtain drugs directly.ll/ 1.13 Health Personnel. Nigeria has made great efforts to expand health services, especially in training and deploying health manpower. For instance, the number of doctors quadrupled to 16,145 during 1975-87. The resulting doctor-population ratio of 1:6,603 compares well with that of other low-income 10/ For instance, in Oyo State nonpersonnel recurrent costs decreased from 35X to under 7Z during 1981-84. In some LGAs, less than 1U of the budget was available for nonpersonnel expenditure. This shrinking of nonpersonnel recurrent expenditure is discussed further in Chapter III. 11/ This was confirmed by a preparatory study for the Population and Health Project for Imo State wbich found that lack of drugs and other supplies and higher user fees were the main reasons people cited for not using public health services. In Kano State, outpatient attendance fell from 104 visits per 100 persons to 40 per 100 during 1981-85. Chapter V.B finds lack of availability of drugs a crucial factor in determining clinic attendance in Ogun State. - 7 - countries.121 During the same period, the number of nurses tripled from 17,904 to 56,120 (a ratio of 1:1,900), and the number of midwives more than tripled, from 13,101 to 45,852. Hospital beds almost doubled during 1975- 87.13/ The resulting ratios of health professionals to population compare very favorably with ratios in countries with per capita national incomes similar to Nigeria: the comparative data (for 1984) on population per nursing person in Sub-Saharan Africa show Nigeria had about three time the number of physicians as the regional average, and twice the number of nursing staff.14/ 1.14 Despite these undeniable advances, the system is plagued by low service quality and poor distribution of facilities. Many government health facilities (particularly rural facilities at the primary level) are undersupplied, inadequately staffed, and in poor physical condition. Although the health sector has recorded impressive reductions in mortality rates, as noted in paragraph 1.09, many deaths are still due to preventable diseases. 1.15 Inadequate and imbalanced allocations of funds are clearly affecting the operation of health care facilities. A study to identify the reasons for severe underutilization of public sector health facilities in Imo State cited "limited access, in time and cash relative to the service provided, lack of supplies and equipment, and high costs." 15/ 1.16 The poor service stemming from inadequately supplied facilities is a priority problem. In addition, four other constraints have adversely affected the public health sector: (i) an inadequate emphasis on preventive and primary health care; (ii) institutional fragmentation and duplication; (iii) poor manpower planning; and (iv) inadequate cooperation between the public and private sectors. These points are discussed below. (i) Inadequate emphasis on preventive and primary health care has been decried in health policies since the mid-1970s. Nonetheless, analysis of recent state health budgets reveals that the emphasis on curative care, which has always been high, grew in the early 1980s. In eight states for which data are available, curative care increased from 72 percent to 81 percent of the health budget during 1981-85. See Table 1.3. 12/ This is somewhat below the average for middle-income countries (1 doctor per 4,800 people). Of the doctors, 18Z were foreigners. 13/ These figures are from the FMOH, which provides a detailed series and covers several categories of manpower and facilities. 14/ World Development Report, 1990 15/ Attah, E. B. 'Underutilization of Public Sector Health Facilities in Imo State, Nigeria," PHN Technical Note 86-1 (Washington, D.C.: World Bank, January 1986). Also flagged were "harsh and uncaring attitudes on the part of health care providers, nepotism in the delivery of services, diversion of services and supplies from public to private facilities, and the imposition of unofficial 'surchargesl for service." Table 1.3 CURATIVE CARE AS A PERCENTAGE OF HEALTH RECURRENT EXPENDITURES AT ALL GOVERNMENT LEVELS IN NIGERIA 1981 1982 1983 1984 1985 Federal 85.5 86.2 85.2 83.1 85.7 States a/ Anambra n.a. n.a. 81.8 81.8 73.6 Cross River 67.1 63.5 63.9 64.2 83.1 Imo 79.6 80.0 75.4 82.5 82.5 Lagos n.a. n.a. 74.0 84.9 n.a. Ogun 79.4 77.2 79.6 87.3 n.a. Ondo 70.7 71.4 74.3 78.0 n.a. Oyo 83.6 85.6 83.9 78.7 83.5 Sokoto 50.6 57.0 n.a. n.a. n.a. Avg. States 71.8 72.5 76.1 79.6 80.7 LGAs b/ Oyo, Oyo 74.0 72.7 37.3 59.1 63.9 Ibadan- Oyo 22.9 20.8 20.9 25.8 33.1 Ibarapa, Oyo n.a n.a 67.0 69.0 71.5 Avg. LGAs 48.5 46.8 41.7 51.3 56.2 Sources Annex F, Tables 9, 10 and 25. a/ Recurrent expenditure is considered as curative at the state level. b/ Within LGAs, curative is discrete from other expenditure categories. (ii) Institutional fragmentation and duplication are particularly serious at the state level, where the functions of State Ministries of Health (SMOH) and State Health Management Boards (SHMB) (see para 2.10) often = overlap. In additioxa, many states have launched programs without regard to services already provided (if inadequately) by LGAs. Local authorities generally lack the technical and managerial capacity to fulfill their mandate of providing PHC services. They have begun to receive state assistance only recently. Yet the states have limited ability to provide the needed technical and logistical support. Their capacity to provide in-service and management training and health education is divided among various SHOH and SHMB programs. More recently, the model PHC program of the FMOH has been targeting the needs of the LGAs, of which about one-quarter have now received federal assistance. That federal, state and local health providers often compete for the same patients reflects the weakness of the referral system. In addition, the positive contribution which the private sector can make is poorly recognized or utilized. (iii) Manpower planning is limited. Little effort has been made to link national priorities in preventive and primary health care with the appropriate training of health manpower. Priorities badly need to be redirected. Vast sums have been spent to train health service professionals (doctors and nurses) whose - 9 - main interest is to work in urban hospitals. Community health workers (CHW), especially at the village level, need to be retrained and reoriented away from facility-based care. Current efforts to revise the clinical manual should assist the CHI to improve the general quality of care. (iv) Effective cooperation with the private sector (modern or traditional) is already a feature of health systems in many African countries. Not so in Nigeria, where the private health sector is dominated by church-based providers. The general attitude of the public sector to the private appears to be one of distrust and skepticism, yet many public providers illegally work in both sectors. Likewise the public health sector has mostly failed to develop relations with traditional care providers, who rem&in the preferred choice for health care in many places.16/ 1.17 Access to Health Care. About 15 percent of urban and 30 percent of rural Nigerians do not have access to modern health care. Since most public and privately owned health facilities are located in urban areas, about 70 percent of health personnel serve only about 30 percent of the population. Patients are further hampered by the poor road network and weak transportation systems. Population per health facility ratios range from about 200:1 in Lagos State to 129,000sl in Benue State. For eight states the ratio exceeds 20,000:l; in five states it exceeds 50,000:1. Health facility beds are concentrated in the southern zones, which have twice as many beds as the northern (Table 1.4). The southwestern zone, which includes the capital, Lagos, has the highest number of government beds. Between 1984 and 1987, the total number of beds increased by 11 percent; of this increase, over three-quarters were in the Southern part of the country. 1.18 Maternal and Child Health Services. Maternal and child health (MCH) is a major thrust of government health policy and programs, but service is still inadequate. Even where primary health care services are available, MCH services often are not (e.g., the MCH equipment may be unavailable or broken). This poor mix means that fewer rural mothers and children receive primary health care than the total number of facilities might suggest. For example, in 1984 it was estimated that more than 70 percent of pregnant women received no prenatal care at all; some 80 percent of births occurred outside health facilities; and around only 5 percent of pregnant women received inoculations against tetanus, a major cause of neonatal mortality. With a major campaign, the rate of inoculation had risen to 16 percent by 1987. A shortage of trained midwives worsens an already bad situation. This shortage is particularly acute in the North, where cultural patterns and low female eCucation standards produce very few local candidates for training programs. 16/ Selective subsidization -- by providing commodities for free or at reduced prices -- may be one way to encourage those healers who pass certain established criteria, such as training courses. - 10 - Table 1.4 DISTRXBUTION OF BEDS IN GOVERNMENT AND NONGOVERNMENT HEALTH FACILITIES (1907) Zone Stato GovGrnment Nongovornmont Total E of Bode a/ Bede b/ Bode Nongovernment BOwe Southeast Anambra 4,876 4,205 8,580 49.0 sonue 1,844 1,660 8,604 47.4 Crose River c/ 2,670 8,492 8,882 54.9 Iso 8,108 7,172 10,280 69.8 Rivera 3,180 176 8,866 6.2 Subtotal 16,877 16,704 82,081 52.1 Southwest Bendol S,450 2,813 6,28B 84.0 Lagos 5,038 1,267 6,296 20.0 Ogun 3,694 1,396 6,090 27.4 Ondo 4,888 2,621 7,409 $4.0 Oyo 8,911 2,230 8,141 86.8 Subtotal 22,981 10,217 38,190 80.8 Northwest F.C.T., AbuJo 271 0 271 Kaduna d/ 6,676 766 6,442 11.9 Kwars 8,242 1,601 4,748 81.6 Niger 1,458 20 1,4?7 1.4 Sokoto 2,595 0 2,695 0.0 Subtotal 18,287 2,287 16,624 14.7 Northeost Bauchi 2,888 66 2,448 2.8 Borno 2,603 12 2,816 0.5 Gongola 2,003 s8 2,859 2.0 Keno 8,688 838 8,901 8.6 Platoeu 8,111 44 8,166 1.4 Subtotal 14,478 500 14,978 8.a TOTAL FOR 19 STATES 66,068 29,708 96,776 81.0 (Including FCT) Sourest FMON, Medical Statistice Division. a/ Fedoral, State and Locol Government beds. b/ Community, Mission, Joint, Corporation and Private beds. c/ Including Akwo-lbom. d/ Including Ketulne. c:teblo 1.4.vkl Qo) - 11 - 1.19 Immunization Services. Nigeria's Expanded Program on Immunization (EPI) against the six childhood diseases (measles, diphtheria, pertussis, tetanus. polio, and tuberculosis) is an essential component of MCH care. Immunization efforts got off to a slow start and were hampered by administrative weaknesses and technical factors, in particular the difficulty of maintaining reliable and uninterrupted refrigeration of vaccines to protect their efficacy. The number of children fully protected against the six diseases was therefore very low. The FMOH estimated that in 1984 only 2.5 percent of infants were adequately immunized against all six EPI diseases. However, by 1987 this had risen to 23.3 percent, and nationwide, complete immunization coverage of children was estimated at 24.1 percent with DPT, 23.7 percent with polio, and 36.6 percent with measles vaccines. (Details of the rapid expansion of the EPI program are shown in Table 1.5.) However EPI, although making successful progress, is a vertical program not yet integrated into MCH services. This integration will be necessary to sustain the program as donor support levels are adjusted. 1.20 Family Planning Services. The Government is aware of the economic implications of continued high population growth and includes provision of family planning (FP) services as a major component of its National Policy on Population launched in April 1989 (para. 2.39). At present, however, the number of family planning acceptors remains low and mostly limited to urban areas. Only about 6 percent of married women of reproductive age use any form of contraception, modern or traditional, including abstinence. 17/ Low contraceptive rates may be explained by the low educational status of women, high infant mortality, inadequate MCH/FP services, and cultural and religious attitudes. Again, avoiding dual delivery, training and equipment by integrating family planning into MCH is essential to sustain the program. Table 1.6 VACCINATION ACTIVM RETIS FOR EPI TARGIET DISEASES AMON CHILDREN UNDER TWO AND TETANUS IN PRENANCY (Percentag. of Target Populations Fully Inrunized against Specific Disease) a/ Al I EPI Tetanus Pollo- Target In TuberculoIes DPT myolitie Measles Dimea Pregnancy 1984 11.6 6.2 2.6 16.9 2.4 5.4 198i 18.8 9.7 9.1 18.1 8.8 7.0 198 85.6 17.7 18.4 26.7 17.7 11.0 1987 44.6 24.1 28.7 86.6 28.8 16.2 Sourco: FMOH. Target populations: - zmaternal targat population = 6% of total population; - Immunixations = 80 of total population. 17/ Preliminary data from Nigerias Demographic and Health Survey, 1990. - 12 - Conclusion 1.21 Nigeria has a commendably comprehensive National Health Policy which recognizes that "Primary Health Care is the key to attaining the goal of health for all the people." However, Nigeria's public health care services have serious shortcomings. Ironically, demand is decreasing while the number of potential users is rising. Staff do not adequately serve rural areas; facilities suffer from poor maintenance; and most public sector health care providers have few drugs for their patients. The report examines these problems in the context of public sector spending while noting that private health expenditure is also significant: individuals spend money on pharmaceuticals and on traditional healers, and in addition, the private sector is active at primary, secondary and tertiary levels. After reviewing the context in which the national policies on health and on population are being implemented, the comparative efficiency of public and private facilities is examined in depth, using data from Ogun State. Against this background, the household demand for health care in Ogun State, from both private and public sources, is analyzed. Opportunities for policy adjustment, including cost recovery, are then developed. The report concludes that although the basic components may be in place for an effective health care program, in Ogun State specifically and in Nigeria in general, adjusting these components could provide a more balanced and responsive program. To achieve its goal of "Health for All by the Year 2000" the Government will need to recognize the complementary role of the private sector and insure that the limited public sector health resources are used as efficiently as possible. - 13 - NIGERIA: HEALTH CARE COST, FINANCING AND UTILIZATION II. INSTITUTIONAL FRAMEWORK AND POLICY DEVELOPMENT IN THE NIGERIAN PUBLIC HEALTH SECTOR Introduction 2.01 Following the peak in the early 19809 federal, state and local government spending declined sharply in real terms. As spending slowed, capital investment fell faster than recurrent speuding, reflecting the stability of personmel costs. Health has received a variable share ranging between 2.7 and 1.0 percent of the actual federal budget during 1981-90; by 1990 that share was worth only 40 percent of the 1981 figure in real terms. (Details are in Chapter III, Tables 3.2 and 3.3.) The Government has made a commitment to providing 'Health for All by the Year 2000.1 The strong primary health care focus of health policies is aimed at meeting this commitment, but the scale of demand exceeds what the public sector can realistically provide. This chapter examines the institutional context of the government commitment, the operation and organization of the public sector, and introduces the complementary role of the private sector in meeting the health needs of the population. A. Background on National Health Policy 2.02 After Nigeria gained independence in 1960, Nigerian government health policy emphasized maintaining the health system inherited from the British and developing the capacity for national planning. Since then, health personnel development and training, delivery of preventive heal.th services, and health system expansion, particularly in rural areas, have had priority. 2.03 The First National Development Plan (1962-68) introduced innovative pilot programs to deliver maternal and child health and health education services utilizing paramedical personnel. These programs were successful and well received by the population, but their replication was delayed by the outbreak of civil war in 1967. The Second Plan (1970-74) covered the immediate postwar period and emphasized reconstructing health infrastructure in the war zones. During the Third Plan (1975-79), the Basic Health Services Scheme (BHSS) was formulated to deliver a full range of health services to the majority of the population. Py 1975, Nigeria's exparding health manpower base was helping to pressure the Government to improve the quality of care, particularly hospital care. The result was a widening gap between rhetoric and reality: while national policy objectives continued to focus efforts and resources on basic and preventive health services, actual expenditure patterns increasingly favored capital-intensive programs oriented to curative care. 2.04 The Fourth Development Plan (1981-85) cited the following weaknesses in the health delivery system: (a) Health systems were biased toward curative, hospital-based programs; - 14 - (b) Urban areas dominated investment in hospital facilities and equipment; (c) An estimated 35 percent of the population had no access to any form of modern health care services; (d) The health sector was unable to attract and retain capable health personnel; and (e) Communicable diseases continued to be the major cause of morbidity and mortality in the country. Identified constraints included (i) manpower shortages; (ii) maldistribution of, inadequate coverage of, and limited access to health services; (iii) inefficient utilization patterns and poor management of health facilities; and (iv) a continuing imbalance in federal and state health allocations between curative and preventive health services. The Fourth Plan aimed to decentralize planning responsibility and to expand basic services through the BHSS. One of its major failings was that the planning of the Scheme was done primarily by the Federal Government, but as its troubled implementation got under way, the facilities were then turned over to the states.18/ 2.05 The goal of the National Health Policy (October 1988) is to enable all Nigerians to achieve a level of health that will enable them to lead socially and economically productive lives. Primary health care is the basis of the comprehensive national health system and encompasses health education, nutrition, sanitation, safe water supply, and family planning. The states are expected to extend essential primary health care to the local populatior., utilizing community and private sector involvement. While the Fourth Plan emphasized extension of coverage through the BHSS, which involved building and equipping new facilities, the draft Fifth Plan (1988-92) focuses on rehabilitation and increased utilization of existing facilities. 2.06 The National Health Policy recognizes the integrated nature of health services in Nigeria: Federal, State and Local Government shall support in t coordinated manner a three-tier system of health care. Essential features of the system shall be its comprehensive nature, multisectoral inputs, community involvement and collaboration with non-governmental providers of health care.19/ The implementation of such a coordinated system will take much concerted effort and good will. The initial step is to recognize the need for it, and this has been firmly taken. 18/ Subsequently the BHSS was implemented on a limited basis and evolved almost as a vertical program, isolated from the regular state health operations. Some buildings, never comlissioned or fully operational, are now being turned over to the LGAs to operate. _9/ The National Health Policy and Strategy to Achieve Health for All Nigerians, FMOH, October 1988, p. 11. - 15 - B. Organization of Public Sector Health Care 2.07 Federal, state and local governments share responsibility for public health (see Annex A). With overall responsibility for the health sector, the Federal Ministry - Health (FMOH) sets national health policies and program guidelines; it m.. oars state and local government health programs for compliance. In addition, the FMOH (a) trains medical doctors; (b) sets uniform standards for all health workers; (c) operates institutions and provides services of a national character, such as teaching, psychiatric, and orthopedic hospitals; (d) controls communicable diseases; and (e) supplies drugs and vaccines. The FMOH has recently been reorganized in line with the new civil service reforms. The new structure for each ministry has up to eight departments, three of which are mandatory (planning, personnel and finance). This reform by the Government focuses on improved planning, rational management and the cost-effectiveness of each ministry, and provides a timely opportunity to lay the foundation for a more efficient health care system.20/ 2.08 The National Council on Health, Nigeria's highest advisory body on health matters, advises the Federal Government on health priorities in implementation of the National Health Policy and the organization of health servihes. Composed of the Minister of Health and the State Health Commissioners, the Council gets advice from a Technical Committee that includes the federal and state Directors General and their key professional heads. 2.09 Each of Nigeria's 21 states has a State Ministry of Health (SMOH) that (a) plans health system development; (b) oversees the operation and maintenance of secondary and nonspecialized tertiary health facilities; (c) implements public health programs; (d) trains nursing, midwifery and auxiliary health personnel; and (e) assists LGAs to manage their network of primary health facilities. 2.10 Until recently, each state also had a State Health Management Board (SHMB) (or State Health Council) which administered its health care facilities, including general hospitals, rural health centers, and urban clinics. Its main function was personnel administration and logistical support management (for drugs, supplies, small equipment, maintenance, etc.). However, some states have dissolved such Boards or Councils and the SMOH is directly involved in administering hospital-based care.21/ Within each of the 454 local government areas, the Local Government Authority (LGA) is now responsible for operating the primery health system, including community health activities such as immunization and health education; hygiene and sanitation services; and provision of basic outpatient services at its maternities, clinics and dispensaries (See para. 3.08). LGAs are assisted by 201 The chart in Annex A shows the organizational structure of the FMOH and identifies the various departments. 21/ This has highlighted the fact that skills in management are now needed more than ever, leading the FMOH to intensify its efforts to strengthen health planning, management and financial capability at all levels of health care delivery. (Some assistance is being provided by the Bank through the Sokoto project (2503-UNI).) - 16 - the State Ministry of Health because, in practice, many LGAs lack the institutional and financial capacity to complete their mandate. However, the increasing autonomy of LGAs, with financial independence under the new revenue allocation formula, will likely make states less inclined to provide support such as personnel, training and supervision. This has serious implications for the health care referral network, which depends on a supportive linkage between the primary and secondary services provided by LGAs and states respectively. C. The Budget Process 2.11 The nature of the federal system of government is such that unless the three tiers of government (federal, state and local) cooperate they will be unable to discharge their respective political and social responsibilities. Because higher levels of government have wider and stronger tax bases and revenue-generating financial resources than lower levels, there is usually a transfer of resources from higher to lower. Equalization transfers are arranged by the Federal or State Government on the basis of need, so that each state and/or LGA attains a minimum given level of services and does not impose an unbearable tax burden in the poorer sections in its efforts to raise the necessary revenue from internal sources. Both balancing and equalization transfers are generally regarded as unconditional intergovernmental grants-in- aid, whose function is ideally suited to offsetting or balancing any general fiscal deficiencies. Stimulation or balancing transfers are usually made with specific directions as to their disbursement. One example is the federal assistance to states and LGAs to enable them to undertake the Expanded Program on T munization (EPI). 2.12 The preparation of the state budget is triggered by the Federal Government, which periodically gives guidelines to states on the fiscul and moaetary policy for the country which may cover interest rates, external borrowing, grants-in-aid and debenture stocks. These guidelines from the Federal Government form the basis for formulation of the Call Circular, issued by the State Governor, which initiates the budget process. This Circular is sent to the Commissioners, Directors General, Chief Executives and Heads of Departments of Parastatals, and requests information on the amount spent in the current year, money balances, deviations from approved expenditures, and estimates of financial needs in the coming years (see Chart 2.1). 2.13 When the budget proposal triggered by the Call Circular has passed through tLe operating units, it is vetted by the State Executive Council. Following processing by the state hierarchy, the state budget is then forwarded to the Armed Forces Ruling Council for final approval. 2.14 The main forces which propel the machinery of transfer of revenue resources are the doctrine of fiscal responsibility and the maximization of state and national welfa.e. Both are essential for balanced and equitable national health policies. - 17 - Chart 2.l1 LAQOS STATE Structure of the Budgeting Apparatue STATE EXECUTIVE COUNCIL l OFFICE OF TE STATE OVNR SeeroUry to the Stats Governor PLANS PROOrAS, AND BtWO%EI DEPARTMENT Director Gneral Off.e oteStateovernI MINISTRIES AND PARASTATALS Dlreetor nOwrnl Chief Executives DPRMnTnS AN DIVISI Diretors and Mreto of DGler on r OPERATING UNITS/COST CENTERS Genoral hospitals, primary health centers D. Health Service Facilities 2.15 Nigeria's health care delivery system is an extensive network of public and private facilities. In 1987, primary care was available from about 4,080 health clinics and dispensaries. Secondary care facilities included 740 health centers and 3,090 maternity centers and homes with 26,650 beds. Tertiary facilities consisted of 14 university teaching hospitals with about 7,100 beds, 16 military hospitals with about 3,000 beds, 27 specialized hospitals with about 4,600 beds, and 760 general hospitals with about 50,100 beds. The remaining 4,300 beds were in specialist facilities.22/ Many government health establishments, particularly for primary care, are underequipped, inadequately staffed, and in poor physical condition. 2.16 The government health sector accounts for about 80 percent of health facilities and about 70 percent of bed capacity. The balance is in the private sector, whose facilities are located disproportionately in the southern states.2S/ For example, the private sector accounts for more than half the beds in the southeastern states of Imo and Cross River, but 22/ This is about 1:1222 people per hospital bed. 23/ In 1987 there were an estimated 1,700 private and church-related health care facilities with over 29,700 beds, mostly concentrated in the southern states. (See Table 1.4.) - 18 - contributes less than 5 percent of the beds in the northern states of Niger, Sokoto, Bauchi, Borno, Gongola, and Plateau states (see Table 1.4). One reason for the greater quantity and proportion of nongovernment beds in the southern zones is the comr.tratively large number of mission hospitals, the Christian religion being widespread in the South, while the North is predominantly Muslim. 2.17 Health Manpower. Nigeria has made considerable progress in training and deploying health manpower. During the period 1975-87, the number of doctors in Nigeria quadrupled from 4,248 to 16,145, of whom around one fifth were foreign (see Table 2.1 below). Over the same period, the number of nurses also increased sharply, tripling from 17,904 to 56,120, and the number of midwives increased from 13,101 to 45,852. The number of pharmacists tripled to 4,446. The most dramatic gains were seen in dentists and all categories of technicians, whose L-mbers increased sixfold and ninefold to 1,000 and 8,186, respectively.24/ The number of "community health workers," a group of several types of personnel introduced in 1979 as key manpower in Nigeria's strategy to provide primary health care services, rose from 1,000 in 1979 to 21,742 in 1987. In comparison with the gains in health manpower, hospital beds almost doubled between 1975 to 1987. The outlook is, however, unstable. With the economic decline of the 1980s and with devaluation which has reduced the value of remittances home, expatriates have been leaving Nigeria. This has especially affected the North, which, in addition, has difficulty recruiting personnel from the South. To complicate the situation, many Nigerian health personnel have gone overseas to more attractive employment. Finally, the private sector has absorbed trained personnel.25/ 2.18 In 1986, Nigeria had 14 teaching hospitals graduating about 800 doctors; 54 general schools of nursing graduating about 5,200 nurses; 46 specialized schools of nursing; and 63 schools of midwifery. In addition, "schools of health technology" (one in each state) trained auxiliary personnel for the primary health care pro3ram. In 1984, these schools graduated 147 community health officers, 232 community health supervisors, 1,679 community health assistants, and 1,801 community health aides. Nearly twice as many doctors and nurses graduated as did all categories of PHC workers -- evidence that only limited efforts were being made to link national priorities in preventive and primary health care with the production of health manpower. With adequate funding, Nigeria's health manpower training capacity, already well developed, is capable of producing sufficient health manpower over the near term. Three major issues remain unresolved: the poor quality of health 24/ Here technicians include Public Health Superintendents, Public Health Inspectors, Medical Laboratory Technologists, Medical Records Officers, Radiographers, Audiologists/Speech Pathologists, Dieticians, Physiotherapists, Dental Technologists, and Dental Therapists. 'Dentists' are also listed in a separate column of the FMOH table and may or may not overlap with the dental categories included here. 25/ In Imo State the number of private maternity homes grew from 336 in 1984 to 378 in 1986, and private clinics/hospitals from 188 to 259 over the same period. - 19 - training, the biased distribution of health personnel toward urban areas, and the skill mix. Table 2.1 HEALTH MANPOWER 1960-1987 1960 1965 1970 1975 1980 1985 1987 Total doctors 1,970 1,981 2,683 4,248 8,037 14,757 16,145 of which Nigerians 349 726 1,701 3,039 6,192 11,908 13,332 Total dentists n.a. :.a. n.a. 168 285 899 999 of which Nigerians n.a. n.a. n.a. 109 203 694 787 Registered nurses n.a. n.a. n.a. 17,904 27,241 45,976 56,120 Registered midwives n.a. n.a. n.a. 13,101 21,411 39,137 45,852 Hospital beds n.a. n.a. n.a. 54,174 75,722 87,835 95,776 Source: FMOH, Department of National Health Planning, Research and Statistics. 2.19 Establishment of the Primary Health Care Coordinating Unit (PHCCU). The focus on PHC has been evolving slowly. In 1978, a Special Commission was created to review the BHSS's performance in providing PHC. The Commission concluded that while its objectives and policy framework were appropriate, BHSS's limited attempts at implementation had been a costly failure. The Commission recommended (a) greater involvement of state and local governments and conummities in developing health programs; (b) full integration of existing facilities and health manpower into the desired health system; (c) development of less costly, more appropriate facility designs and equipment lists; and (d) adoption of explicit evaluation criteria. 2.20 After these recommendations were accepted in late 1981, the PHCCU was assigned the tasks of (a) training auxiliaries and professionals working in PHC; (b) supplying medical equipment and water and sanitation facilities to federally supported PHC centers; and (c) managing special programs, such as a women's health and development project, model MCH/FP programs, an overhaul of - 20 - PHC statistics and information gathering systems, and national workshops on essential drugs and their financing. Current plans include strengthening the Health Education Unit to better assist the states in family planning and developing the fledgling Nigerian AIDS Control and Prevention Program. In the reorganization of 1988, the PHCCU was upgraded to become the Department of Primary Health Care. E. Recent Developments in Health Planning 2.21 During the past several years, the technical shortcomings of the states in planning and evaluation and the need for consistency in implementing health programs has spurred the FMOH to support state and LGA planning efforts. Since 1982, the Department of National Health Planning, Research and Statistics (DHPRS) has taken the lead in developing health planning capacity at all levels of government nationwide. Related activities have been financed with assistance from the World Bank through the Sokoto Health Project (2503- UNI). 2.22 One goal of the DHPRS has been to translate health policies into operational programs. Assisted by international consultants, the Department prepared a manual and guidelines on health planning and a model health plan for Ogun State in 1987 (its Table of Contents is presented here as Annex B) and held national workshops to educate state and LGA officials on health planning. These materials explained how systematically to obtain, analyze, and put to use relevant information to help overcome identified constraints. Participating state officials were then each asked to submit a five-year state health plan to help prepare the national health plan. 2.23 The workshops have been well attended by senior LGA and state officials. Their unique combination of theory and practice has generated enthusiasm for systematic planning and has sparked deeper interest in health finance because identifying financial constraints and evaluating costs and cost-effective use of resources are planning components. The preparation of these plans has encouraged comunity participation and recognition of the need for cost sharing. This heightened interest in health finance could have a wide impact because health budgets often do not consider the actual or comparative costs of existing health programs. The outcome has been a series of high-quality comprehensive state health plans which were ti.nsformed into more manageable three-year rolling plans and became the basis for the draft Three Year Development Plan for Health 1989-91. The Plan's centerpiece is a long-term primary health care investment strategy which emphasizes innovative investment in maintenance and limited expenditure on new construction. F. The Nigeria Essential Drugs Program 2.24 Building on the National Health Policy's commitment to ensure that "essential drugs ... are available at all levels of the health care system," in June 1988 the FMOH and all the states endorsed a national policy on essential drugs as a basis for development of the proposed National Essential Drugs Program. This policy defines essential drugs (EDs) as safe, effective, but inexpensive drugs which satisfy the health care needs of the majority of - 21 - the population and hence are available at all times throughout the health care system.261 2.25 Many sub-Saharan African countries have demonstrated that effective national essential drug programs (EDPs) can save money.27t The potential economic benefits of the Nigerian EDP are high. Drugs are typically the second largest expenditure item (after personnel) in Nigeria's ever-shrinking health budgets (see Chapter III). It has been estimated that Nigeria could save 75 percent of current drug expenditures by using more cost-effective practices. Thus, current expenditure levels for drugs could serve four times as many illnesses. 2.26 A shortage of appropriate drugs continues to plague public sector health facilities in Nigeria. The drug shortage undermines the effectiveness of recommended treatments and reduces public confidence. One predictable result is low utilization rates at public facilities and possibly low employee morale. Small budgets and inadequate foreign exchange contribute to the shortage of imported and domestically produced drugs. 2.27 Prevailing drug selection, procurement, distribution, and utilization practices have also undermined access to priority drugs. World Health Organization (WHO) technical reviews28/ revealed that (a) ineffectual and even dangerous drugs are frequently procured;29/ (b) as a rule, branded rather than less expensive generic drugs are procured; (c) cost-effective procurement procedures are often not used. Drugs are often purchased locally in small quantities instead of in bulk through open international tender. As a result, some local governments pay up to six times international prices; 26/ The term "essential drugs" has come to be used for that sub-set of drugs that will effectively treat, at a reasonable cost, the large majority of ailments (say 952) experienced in a given geographic area. WHO has prepared a model list of about 250 essential drugs, selected from the many thousands marketed worldwide, which can be adapted for individual countries. 27/ Thanos Catsambas and Susan Foster, "Spending Money Sensibly: The Case of Essential Drugs," Finance and Development, vol. 23, no. 4, (December 1986), pp. 29-32. 28/ "Nigeria: Sokoto State Health Project (LN 2503-UNI), Essential Drugs Programme and Drugs Revolving Fund -- Report of a WHO Mission, 14 April-2 May 1986," (Geneva: WHO, June 1986); and "Nigeria: Imo State Health and Population Project, Essential Drugs Program and Drug Revolving Fund -- Report of a WHO Mission, September 8-21 1986," (Geneva: WHO, 30 September 1986). 291 For instance, 252 of Sokoto's 1986 drug expenditures were for several branded products no longer used in many countries. - 22 - (d) many drugs are damaged after purchase by faulty storage practices or disappear due to inadequate stock control procedures; and (e) because of inadequate diagnostic capabilities, health staff prescribe an excessive number of drugs in an attempt to treat a number of possible diseases simultaneously. 2.28 The FMOH believes that charging for drugs is an appropriate way of recovering some of the costs of health services. Sharing this view, the Bank's recent policy study on health financing cites three reasons why drug charges are particularly suitable for cost recovery: (a) consumers already widely accept the idea of paying for drugs; (b) charging for drugs can give the public sector a useful management incentive for improving efficiency; and (c) accountability of managers and distributors can be readily built into a revolving fund system for drug purchases.30/ Ensuring drug quality must be part of any cost recovery program,31/ or else the fees may drive away more clients and reduce usage of government health services further. 2.29 Aside from recovering costs, charging fees for drugs offers a method of manipulating consumer demand. Price differences can be used to shift preferences to less expensive substitutes. Patients who pay for their own drugs will probably be more careful about using them properly. Information on drug sales indicates demand and can help guide procurement. 2.30 As part of an EDP, a Drug Revolving Fund (DRF) can recycle the proceeds from charges on drugs to sustain their availability.32/ A DRF helps drug purchases to pay for themselves by using the revenue from drug charges to replenish supplies or to maintain a constant cash asset base. Depending on how subsidies are used in the fee structure, additional funds from annual health budgets may still be required to maintain a DRF's asset base fully. 2.31 With WHO and World Bank support, the FMOH is implementing a Nigerian Essential Drugs Program (NEDP) to help states (i) select, procuref store and distribute good quality essential drugs in sufficient quantities at the lowest possible cost; (ii) improve the diagnostic, prescribing, dispensing, and cost sensitivity akills of health care workers; and (iii) make the supply of drugs self-financing through cost recovery and drug revolving 30/ John Akin et al., "Financing Health Services in Developing Countries: An Agenda for Reform: A World Bank Policy Study" (Washington, D.C.: World Bank, April 1987). 31/ Drug quality is an issue of increasing public and professional concern because the reliability and quality of drugs available in the markets have declined. Reportedly, drugs have been diluted and/or mixed vith other, sometimes harmful, substances, most likely because of higher foreign exchange costs resulting from devaluation of the naira. Ensuring the quality of drugs at government health facilities through an effective EDP could give public facilities a competitive edge, particularly over the purchase of drugs at chemist shops for self-medication. 32/ With the naira floating freely, prices should reflect the real cost. - 23 - funds. The project will help improve FMOH's technical capacity and establish small essential drug programs at key federal health care institutions.33/ A number of Nigerian states have already begun charging for drugs, and several have requested help in managing effective revolving funds. Better implementation of these measures, coupled with more rational selection, procurement, distribution, and prescription of drugs, could make drugs cheaper and more available and aid Nigeria's efforts to provide effective PHC. 2.32 Domestic Production of Pharmaceuticals. The local, mostly private, pharmaceutical industry imports or manufactures about 3,000 different final products and ingredients worth N300 million retail in 1985 (N150 million wholesale).34/ Operating at about one quarter of capacity, domestic manufacturers are extremely inefficient. They make tablets, capsules, powders, and liquids (but no injectables). There being no pharmaceutical fine chemical industry, all therapeutic ingredients are imported. A few components such as starch, glucose, glycerol, oxygen, etc., are manufactured locally, but their economic importance in production is low. Drug production, therefore, consists basically of preparation and packaging of imported raw materials. There are no important technical constraints to the expansion of local capacity to meet the government's requirements for essential drugs. The assured local demand under the proposed National Essential Drugs Program, with set, published lists of drug requirements, would also provide a more dependable market for local producers than in the past. However, in view of the limited local value added and the present inefficiencies associated with lccal industries, preliminary analyses indicate that with regard to only a very few drugs can Nigeria be competitive with foreign producers. However, wiAh fiscal incentives (e.g., tax holidays) this might change with regard to bome drugs. In any case, the private sector with its presently badly underutilized capacity clearly has a comparative advantage over the Government in the manufacturing of drugs, including essential drugs. 2.33 Vaccine production is a separate issue. Vaccine manufacturing is a risky activity for which consistently high quality standards have to be established and scrupulously maintained.35/ One faulty batch, resulting in one publicized fatality, can wipe out the public confidence in the immunization program which has been built up so painstakingly. Nigeria has a large population which may make production cost-effective, but high-quality vaccines can readily be purchased at very low prices from multiple international sources so that during this period of financial constraints it does not appear appropriate that Nigeria make the initial investment in what would become a major production activity. 33/ The National Essential Drugs Project (Loan No. 3125-UNI) became effective on September 10, 1990. 34/ About 100 pharmaceutical manufacturers and importers are active in Nigeria. The three publicly-owned manufacturers represent a very small fraction of the subsector's turnover. 35/ Some of the necessary equipment includes air conditioners, humidifiers, dust extractors, and separate manufacturing facilities for viral vaccines, bacterial vaccines, and human blood derivatives. - 24 - 2.34 In reaching decisions on production of both essential drugs and vaccines, the capability of the private sector should be thoroughly reviewed before substituting public facilities for private capacity. Meanwhile, adulterated drugs are flooding the market, making it vital that expenditure on quality control facilities to test food and drugs remains a high priority. G. Government Population Policies and Programs 2.35 Population growth rates in many Afr-ican countries remained low until about 1950. Many early African leaders considered their populations below optimum size because of low population densities, past losses due to slave trading, and high mortality. At the time of its 1963 census, Nigeria's population was 55.7 million and growing at 2 percent annually. (For 1985-90 that figure is estimated at an annual 3.2 percent.) But even Nigeria's 1962- 68 First National Development Plan recognized the interrelationship between population growth and resources: ...the population is increasing at a faster rate than the provision of new facilities, health services, housing, schools, transport.... Despite major investments, the situation is liable to deteriorate faster if the present increase in the rate of population growth continues. While the Government supported family planning services to help families to choose and to space births, it took little initiative to stimulate awareness, leaving the limited activities in this area to NGOs. 2.36 Until very recently, little was done to reduce population growth. The Second National Development Plan proposed but did not implement a national family planning program. The Third Development Plan (1975-79) coincided with the oil boom, when Nigeria's large and rapidly growing population was seen as no obstacle to economic progress. The increase in oil revenues financed an ambitious expansion of infrastructure and social services which did not then seem burdensome. Nonetheless, the high rate of rural migration to the cities was seen as worsening employment and urban development problems. Recognizing the potential implications of continued high population growth, the Third Plan called for (a) maintaining support for family planning; (b) strengthening the National Population Bureau and the Planned Parenthood Federation of Nigeria; and (c) establishing a reliable registration system to develop an ongoing demographic database. Yet no commitment to a population policy explicitly aimed at reducing the growth rate was articulated at this time. 2.37 The overriding objective of population policy in the Fourth Development Plan (1981-85) was to use sample surveys to generate appropriate, accurate population data for planning purposes. To address the dearth of demographic data, the National Demographic Sample Survey of 1980 and the Nigeria Fertility Survey of 1981-82 (part of the World Fertility Survey) were conducted in cooperation with the National Population Bureau (NPB). These surveys confirmed a population growth rate of 3 to 3.5 percent per annum that would double the population in less than 25 years. When oil revenues plummeted in the early 19809, deep cuts in recurrent budgets and postponements in the completion of ambitious capital projects became necessary. The current environment of debt and fiscal austerity has exacerbated the impact of continued high population growth on basic social services, food availability and urbanization, and has encouraged the Government to adopt a comprehensive population policy aimed at reducing the population growth rate. - 25 - 2.38 In 1988, the National Population Cormission (NPC) was established, with the primary task of undertaking a national census in 1991. Its secondary role is monitoring the impact of the National Population Policy. To undertake this work, the NPC has incorporated the skills of the former NPB. 2.39 The National Policy on Population. In April 1989, the National Policy on Population was launched, with the ambitious target of reducing the annual population growth rate to 2 percent by the year 2000.36/ Aside from reducing the population growth rate, the policy focuses on improving maternal and child welfare through encouraging smaller families by means of (a) reducing marriages under age 18; (b) decreasing pregnancies in women below 18 and over 35; (c) increasing child spacing to at least two years; (d) reducing the average total fertility rate from more than six to four children; and (e) promoting a continued decline in death rates, particularly for infants and mothers. Implementation of the policy will require commitment at all levels of government, with a coordinated effort across all involved ministries. In particular, the signals from the Government should consistently support the policy. This will be particularly difficult where the allocation of federal revenue to the states is concerned. Currently, forty percent of this amount is based on the size of the estimated population, providing a strong disincentive to curb the growth rate. Conclusion 2.40 The Federal Ministry of Health has recently unveiled two major policy document0s the National Health Policy and the National Policy on Population. Both are comnendable in their synthesis of needs and the strategies outlined to meet them. However, implementation will require much effort, in particular an adjustment in approach. The challenge confronting Nigeria's health system is not building buildings and training personnel, but 'doing more with less" -- developing the system with limited resources. Although this is a familiar prescription for countries facing budget constraints, it is particularly appropriate for Nigeria, where low utilization rates are common. The resources available to the public sector are the subject of the next chapter. 36/ The policy was drafted by a comittee composed of representatives of the Federal Ministry of Health, Federal Ministry of National Planning, National Population Bureau, Nigerian Institute of Social and Economic Research, University of Lagos, the Planned Parenthood Federation of Nigeria, and PAI Associates International. It was reviewed, modified and approved by the Interministerial Consultative Group on Population and Development. Technical consultation was provided by the World Bank, UNFPA, ILO and other national and international ezperts. The Bank is supporting implementation of the policy with the National Population Project, expected to become 4" - 26 - NIGERIA: HEALTH CARE COST, FINANCING AND UTILIZATION III. PUBLIC SECTOR HEALTH RESOURCES 3.01 To calculate annual total public health resources, data must be gathered from federal, state and local levels. In 1986, such an exercise was undertaken by the Federal Ministry of Health in conjunction with the World Bank on a sample basis.37/ This chapter shows how the decreases in government revenues have had a large impact on the public health sector at all levels. The first part of this chapter describes the flow pattern between federal disbursements and the state- and LGA-run health facilities and programs that spend the money. The chapter then reviews total government revenues and expenditures. Finally, it examines disaggregated budgetary allocations to the health sector at different levels of government.38/ 3.02 The highlights of this examination are that at all three levels of government during the first half of the 19809 (a) the shares of the budgets allocated for health were relatively stable; (b) recurrent expenditures became increasingly dominated by personnel costs; (c) expenditures generally rose in nominal value but fell sharply in real terms; and (d) as a result, the estimated total per capita expenditures were approximately halved in real terms. In the second half of the 1980e, the share of the federal budget for health declined and, based on more limited data, the picture appears to be similar at the state level. 3.03 Although the Federal Government provides most of the funds, it is estimated that about two-thirds of all public sector health expenditure occurs below the federal level.39/ The states are heavily dependent on federal funding to meet recurrent liabilities in health services, but the federal budget depends on oil revenues. Thus, even with major efforts to increase internal revenues, state budgets cannot avoid fiscal volatility. Such volatility is clear from the data in this chapter. 37/ A Survey of the Financing of Primary Health Care, 1980-86, Primary Health Care Department, Federal Ministry of Health. Prepared by Shirley and Francis Ltd., Lagos, Nigeria. 38/ In this chapter, the analysis is based on actual figures for 1981-85 rather than on the approved budget estimates, which are prepared annually by the Federal Government and by each state and LGA. The available data are limited, but they indicate that there can be substantial differences between actual revenues and expenditures and the budget estimates. The sources and reliability of the data at each level of government are described in Annexes C and D. In addition, Annex E includes an estimate of total government health expenditures in 1985 derived in part from the relationship between budget estimates and actual figures for a larger number of states in that year. 39/ The estimation is based on calculations in Annex E. - 27- A. The Flow of Health Sector Funds 3.04 Chart 3.1, which is divided into four quadrants, shows the flow of government health funds. Revenue sources are shown on the left of the chart. Resources flow to the right, directly to the facilities and programs or indirectly via the various governmental decision units. The horizontal division of the chart represents the distinction between the government institutions that oversee and support health care (above the horizontal line), and the facilities and programs actually providing health services (below tne line). 3.05 Revenues. As in other sectors, public health sector funds originate in revenue generated internally by domestic taxes or in revenue from external sources. The main source of internally generated revenue is the oil sector. Oil taxes feed the Federation Account, of which 50 percent is allocated to the federal level, 32 percent to the state level, 15 percent to the local government level and 3 percent to the Special Funds.40/ In recent years, the Federation Account has been the most important revenue source at all levels of government. 3.06 Internal revenues derived from other taxes, licenses, fees, etc., make up a variable proportion of state and LGA funding sources. LGAs have limited powers of taxation; state governments are required to distribute 10 percent of their internally generated revenue to the LGAs.41/ User charges at health facilities generate a small amount of internal revenue that is usually returned to the governmental unit administering the facility, although the health facility or program that provides the service may retain some amount. Communities may make in-kind contributions, typically for constructing community facilities or for buying a specific piece of equipment or stock of drugs for them. 40/ Until 1984, states received and often retained an undetermined portion of the 102 of the Federation Account allocatior. (changed to 15 percent in 1990) which should have been passed on to the local governments within their boundaries. Now LGAs receive their allocation directly from the Federal Government through a newly created LGA unit in the State Governor's office. The balance of the Federation Account is allocated to the Special Fund for national programs such as the Amelioration of Ecological Problems and the Development of Mineral Producing Areas. 41/ In recent years, the states often have not paid out to the LGAs any of the 10X of their internal revenue which was called for by the original law (Federal Allocation Bill of 1981). However, states have sometimes used all or part of this retained revenue to pay for LGA- employed teachers. CHART 3.1 Recurrent Expenditure R1PI£NU£S EXPENDIURES EXTERN INTERNL DOWc toIintnsc and Iant Insce federal on FeaA al fN-ou F1ME% Rvw Revenue FMFED | t~~~FMO federation Account FMOH 2 j~~~~3% 32% 0E | Speekg State S tat _ L_ __ _ m _ Externalxtinc IG A . I GAX ConmmfV \w Contbbutiont) leaching SchoolS6t Genercal P1-C and HoFitas FacdtXes Specilty and Hospitas P1Ogrms +~~~~~~~~_ t ,NaWl_. .k1g&y hsi ) CMOSog ciid h_cl .mp3w . - 29 - 3.07 Expenditure allocations begin with partitioning resources between the recurrent and capital budgets. The Federal Ministry of Finance and Economic Development (FMFED) makes this division at the federal level based on approved, annually prepared budget estimates. The State Ministry of Finance42/ and the State Governor perform this function at the state level, while Local Government Councils perform it locally. The FMFED allocates recurrent expenditures of the Federal Ministry of Health (FMOH) on the basis of FMOH proposals, between FMOH personnel expenditures, FMOH nonpersonnel expenditures, ard "Grants and Subventions." These grants and subventions support a variety of facilities and programs, such as the 14 teaching hospitals, other federal specialized hospitals, and federal primary health care programs, among them the Expanded Program on Ivmunization. 3.08 State health resources are divided between the State Ministry of Health (SMOH) and the State Health Management Board(SHMB). The state h2s the responsibility for all secondary level care. Generally, the SMOH used to handle planning and epidemiological surveillance and also supervise the state's nursing, midwifery and health .echnology training qchools. The operation of facilities (such as health centers, clinics and hospitals) was usually left to SHMB.43/ Until 1989 this division of services left the state providing some primary curative services through the operation of state-owned health centers, as well as operation of secondary level care provided by the general hospitals. However, the LGA health budget as provided from the federal allocation has now been increased, passing to these bodies the statutory responsibility to deliver all primary health care, including the selection, employment and paying of staff to operate such facilities. The LGA is now responsible for deciding on the allocation to sectors, including health, and preparation of a budget, to be reviewed and approved by the Governor's office. These LGA health allocations will now include salaries and operating expenses for all primary health care facilities (including clinics, dispensaries and maternities), as well as for certain public health activities such as immunizations, health education, health inspections and environmental sanitation; LGA activities, however, remain under the supervision of the SMOH. 3.09 Allocation of Capital Expenditures. At the federal level, the FMOH proposes a health sector capital spending program to the FMFED each year under the guidance of the Directorate of National Health Planning (DNHP) and the appropriate five-year plan. The FMFED then allocates the 42/ In some states the functions of finance and Ilanning are joined, as at the federal level, into a single ministry ca:led the State Ministry of Finance and Economic Development. 43/ To allow comparability between states in the following analysis, SMOH and SHMB expenditures have been separated from or combined with each other where appropriate, regardless of whether the SHMB had originally appeared within the SMOF or SMOH budget. - 30 - FMOH capital budget.44/ At the state level, analogous roles are played by the State Ministries of Health and Finance. The capital budget for health is usually small at the local level. Federal capital expenditures cover mostly curative care (such as the teaching and specialty hospitals) but also fund vertical programs in preventive care such as communicable disease and malaria control programs. 3.10 Each of FMOH's three involved Directorates (Hospital Services, Disease Control, and Primary Health Care) gets a portion of the FMOH capital budget, which it uses to support operations. The Ditectorate of Hospital Services and Training oversees the development of teaching and specialty hospitals, the same hospitals that receive most of the 'Grants and Subventions' component of the FHOH recurrent budget.45/ The Directorate of Disease Control and International Health spends about a third of its capital budget for federal disease control activities at Nigeria's ports of entry. The remaining two-thirds pay for the communicable disease and malaria control programs, which are implemented by state ar.d local governments. 3.11 Now formally a Directorate, the PHC Department receives a segment of the Fn0H capital budget, which it then allocates among nine expenditure categor!es. Some resources go directly to the SMOHs for PHC activities and for training LGA communit, health workers. Other funds buy equipment for the Basic Health Units and Nigeria's Expanded Program on Immunization. 3.12 At the state level, the capital health budget finances construction, improvements and maintenance of the state health manpower training schools and the health care facilities managed by the SHMB. The SMOH usually retains authority for the portion of the capital budget allocated to SHMB-managed facilities. Although SKMB capital budgets in some states are small, these do not generally fund construction. LGA capital budgets help operate local health facilities and paf for public sanitation, such as latrine construrtion and market cleanup. B. Total Public Expenditure 3.13 Total government health expenditure in 1985 was approximately N800 million or between N8 and V9 per capita. At 1985 exchange rates, this was about US$8 which was near the middle of the range of per capita expenditures for middle income countries. However, unlike all other Sub- Saharan African countries, it is estimated that Nigeria spent less than a 44/ Within the FMOH, the Directorates of National Health Planning, of Pharmaceutical Services, and of Food Drug and Laboratory Services each indirectly implements programs in support of health activities and is funded by a portion of the health capital budget. 45/ The Directorate also has responsibility for the 'Allied Health Ma.ipower Development" portion of the capital budget, which supports the federal schools of allied health manpower, but not the State Schools of Health Technology, which also train health auxiliaries. - 33 - third of this total centrally at the FMOH. Over half of expenditures occurred at the state level and about 15 percent to 23 percent at the local level. (See Annex E.) With expenditures so decentralized, all three levels of government must be considered in reviewing both the level and the volatility of Nigerian government health expenditure. 3.14 Reflecting national constraints, spending declined at each level of government during 1981-85, generally by more than the decline in revenues (Table 3.1). Federal spending fell 56 percent, exceeding the real revenue decline of 39 percent. State spending similarly declined 51 percent, compared with an average revenue decline of 31 percent. Limited data are available for LGAs, but it appears that they experienced similar declines. Table 3.1 TOTAL EXPENDITURES BY FEDERAL GOVERNMENT AND SELECTED STATES IN NIGERIA, 1981-85 (Actuals, in Millions of 1981 Naira) 1981 1982 1983 1984 1985 Change (Z) Federal 12,800 12,000 10,200 5,600 5,600 -56 States imo 409 328 258 210 n.a. -47 Lagos 477 465 396 230 264 -45 Ogun 221 149 131 72 n.a. -67 Ondo 258 232 186 117 n.a. -54 Oyo 323 315 269 209 252 -22 Kwara n.a. 135 131 102 76 -43 Sokoto 323 284 230 120 115 -64 Gongola 330 227 145 92 124 -62 CPI 100 108 133 185 195 Source: Annex F, Tables 1, 4 and 5 3.15 Investment in Health Care. While the nominal recurrent -allocations by the FMOH have shown a fairly steady increase, nominal capital allocations have been much less regular. The contrast between the behavior of these two aggregates is most apparent in the early 1980s, when oil revenue encouraged FMOH to implement a major hospital construction program (see Table 3.2). In real terms the total FMOH allocation in 1990 was worth about 40 percent of its value in 1981. - 32 - Table 8.2 FEDERAL ALLOCATIONS TO FUOH (Millione of Naire) Approved Actual Total In Year Recurrent Capit.l Total Recurrent Capital Total 1981 Noaie 1980 119.8 110.4 280.2 112.6 77.8 190.4 a/ 280.0 / 9t81 166.8 183.4 839.2 168.4 142.1 296.6 295.5 1982 143.6 186.0 329.6 165.2 120.6 278.7 264.1 1988 148.6 170.0 818.6 168.8 116.6 289.9 208.4 1984 180.1 61.8 190.8 168.1 42.8 210.9 118.8 1986 167.7 49.4 217.1 177.2 46.4 222.6 118.9 1986 289.2 81.2 820.4 246.8 181.6 377.a 188.2 1987 168.9 69.6 236.4 229.0 126.8 864.6 166.8 1988 269.9 183.2 443.1 819.6 117.3 438.9 189.8 1989 369.7 126.0 486.7 NA NA NA NA 1990 401.1 267.0 658.1 401.2 181.6 682.8 b/ 119.8 b/ a/ April - December. b/ Provislonal. Source: Federal Government of Nigeria, Approved Budgots; Federal Ministry of Health; and Staff Estimates. 3.16 As a share of the actual total federal budget, health care shows a distinct downward trend, from around 2.0 percent in the early eighties, to 1.0 percent in 1990 (see Table 3.3). Table 3.3 FMOH AS SHARE OF THE FEDERAL BUDGET ALLOCATIONS (Percentage) Approved Share of Share of Actual Share of Federal Federal Share of Total Recurrent Capital T-tal Federal Budget Budget Budget Federal Budget 1980 2.2 3.4 1.6 1.6 1981 3.5 4.4 2.9 2.3 1982 3.0 4.2 2.4 2.1 1983 2.5 2.6 2.4 2.0 1984 1.9 2.3 1.3 2.0 1985 1.9 3.1 0.9 2.0 1986 2.8 4.2 1.4 2.7 1987 1.3 1.5 1.0 1,3 1988 1.8 1.9 1.7 1.3 1989 1.6 1.7 1.4 n.a. 1990 1.7 1.5 2.0 1.0 Source: Federal Government of Nigeria, Approved Budgets; Federal Ministry of Health; and Staff Estimates. - 33 - C. Recurrent Expenditure 3.17 The share of recurrent expenditure allocated to health increases at the peripheral levels of government. Table 3.4 shows that during 1981- 85 health captured from 2 to 4 percent of the federal recurrent budget and around 10 to 11 percent in the five states for which actuals are currently available. Limited data suggest that health consumes approximately 31 to 40 percent of LGA recurrent expenditure. Table 3.4 PERCENT OF TOTAL RECURRENT EXPENDITURE ALLOCATED TO HEALTH BY THE FEDERAL GOVERNMENT AND SELECTED STATES IN NIGERIA, 1981-85 (Actuals) 1981 1982 1983 1984 1985 Average Federal 3.1 2.7 4.1 3.7 3.3 3.4 States Imo 9.0 10.7 9.2 9.4 n.a. 9.6 Ogun 11.0 12.3 12.4 14.6 n.a. 12.6 Ondo 11.9 11.1 12.5 12.3 n.a. 11.9 Oyo 10.5 9.1 9.1 8.2 8.2 9.0 Sokoto 9.1 9.7 n.a. n.a. n.a. 9.4 Source: Annex F, Tables 4, 5 and 7. 3.18 The larger percentage spent on health at the local and state levels than at the federal reflects the differing responsibilities of various levels of government, as well as the responsiveness of peripheral governments to the needs of their local constituency. Conversely, the federal level of government, with its larger international and national responsibilities, spends a smaller proportion of its government budget on health. This pattern is likely to appear in any federal system. 3.19 There is a current trend in Nigeria for the Federal Government to allneate an increasing share of its revenue for expenditure by state and local governments. 46/ The health expenditure patterns just identified have an interesting implication regarding the eventual effects of federal revenue sharing on health care expenditure in Nigeria. 3.20 Assuming the share of health expenditure remains constant at all levels of government, every M100 transferred from the federal to the sta-e levels would reduce health expenditures at the federal level by N2 to N3 while increasing it at the state level by N10 to Mil, for a net increase in health expenditure of 17 to N9. Similarly, every N100 transferred from 46/ The local government share of the Federation Account was recently increased from 10 to 15 percent. - 34 - the federal to the local levels would produce a net increase in health expenditure of N28 to N38.47/ 3.21 Thus one implication of fiscal decentralization and revenue sharing is that Nigeria's government health expenditure will, on balance increase. Furthermore the predicted increases are substantial and will occur at the local and, to a lesser degree, at the state level. The LGAs will need to plan for these increases, in order to assure their maximal benefit for the health of the local populations. 47/ N100 transferred from the states to the local governments would have a net positive effect on health expenditure on N20 to 130. The recent change in LGA resources, along with their increased health responsibilities, would reflect this effect. (These calculations assume that percentages at the margin are the same as the averages and are based on Annex E.) - 35 - Table 3.5 RECURRENT HEALTH EXPENDITURE AS SHARE OF TOTAl HEALTH EXPENDITURE FOR FEDERAL GOVERNMENT AND SELECTED STATES IN NIGERIA, 1981-85 (Actuals - Percentage) 1981 1982 1983 1984 1985 Average Federal 51.9 56.0 56.8 79.7 79.6 63.3 States Anambra n.a. n.a. 98.1 95.8 86.2 93.4 Cross River 84.6 93.6 98.3 98.6 92.5 93.5 Imo 94.3 82.0 96.4 99.7 99.1 94.3 Lagos n.a. n.a. 92.2 97.0 n.a. Ogun 81.9 93.5 96.3 97.1 n.a. 2.2 Ondo 93.9 96.2 91.2 98.8 n.a. 95.0 Oyo 83.4 90.1 90.1 77.4 92.7 86.7 Sokoto 64.6 75.5 n.a. n.a. n.a. 70.0 States, Avg. 83.8 88.5 94.6 94.8 92.6 (I in Avg) (6) (6) (7) (7) (4) Source: Annex F, Tables 7, 9 and 10 3.22 Two large shifts are clear from Tables 3.5 and 3.6. First, recurrent expenditures absorbed a growing share of health budgets (Table 3.5). Second, within recurrent expenditure, personnel costs were increasingly dominant (Table 3.6), especially at the state and local levels. Of the FMOH recurrent budget, 10 to 13 percent was spent within the central ministry for salaries of ministry personnel and overhead expenses, including drug supplies; Table 3.6 shows that within the central ministry, personnel spending stayed fairly constant during 1981-85 at approximately 70 percent of recurrent expenditures, except in 1984. The balance of the FMOH recurrent health budget passed through the Ministry in the form of grants and subventions to various entities, such as the federal hospitals located throughout the country, and to special programs, such as those of the Primary Health Care Coordinating Unit (PHCCU).48/ Most of the year-to-year variation in FMOH recurrent expenditure was caused by variation in its grants and subventions component, which grew in high expenditure years and shrank in years when the budget was tight, while the rest of the budget remained relatively unaffected by these changes (Chart 3.3). 48/ Support for the PHCCU, like support for the Directorate of Public Health Services, contributes indirectly to the delivery of health care by enabling the PHCCU to fulfill its mandate to reinforce PHC throughout the country. The PHCCU is now the Directorate of Primary Health Care. - 36 - Table 8.6 PERCENT OF RECURRENT HEALTH EXPENDITURES ALLOCATED TO PERSONNEL FOR FEDERAL, STATE AND LOCAL GOVERNMENTS IN NIGERIA, 1681-65 (Actual) 1981 1982 1988 1984 1985 Federal Mdinistry 71.8 69.1 89.9 58.0 68.0 of Health -/ states Anambra MGO n.e. n.e. 76.5 69.1 44.6 Iso MON n.o. 82.8 87.9 84.6 65.9 Ime HMB n.o. n.e. n.e. 72.0 79.0 Ogun MON 78.6 89.8 92.7 97.7 n.o. Ogun HMB 78.6 89.8 92.7 95.3 98.6 Oyo MOG 62.6 94.9 92.0 79.4 88.0 Oyo HUB 66.6 81.5 83.2 96.9 90.8 Avg. LOAs 84.2 86.8 96.4 96.1 93.1 (o In ova.) (3) (4) (6) (16) (14) Source: Annex F, Tables 7, 9, 10, 18, 17, 19 and 21. 0j In this table, the federal percentage is based on the central ministry only (I.o., the aix Directorates and PHCCU); grants and oubventions are excludod, although they Include expenditures for personnel at federally supportod institutions, ouch as federal hospitals. 3.23 State. As budgets were cut drastically, personnel costs absorbed a greater proportion of state health recurrent spending. And over the five-year period 1981-85, recurrent expenditures have often exceeded 90 percent of the state health budget (Table 3.5). While ambitious capital projects were started during the oil boom, these were cut back in favor of recurrent obligations, which already stood at an estimated 84 percent in 1981 at the state level. - 37 - Chart 3.3 ACTUAL FMOH EXPENDITURES, 1980-90 Total, Reaurrent & Capital Expendituresa Milins of 1981 Nairm 400 800 . ................I . . . . .... ... . . . . . . . . . .... . . . . . . . . . . . . . ...... . . . . 100 1980 1981 1982 1983 1984 1986 1986 1987 1988 1969 1990 Year - Total - Recurrent -- Caplta1 l959 data ore not evaUnbln 3.24 Local Government Authorities. Of the three levels of government, the LGAs allocated the highest proportion of their recurrent budgets to health, approximately 31 percent to 40 percent. But during 1981-85, health did not maintain its proportional share of the available resources, falling from 41 to 31 percent in the limited sample. Recurrent expenditures sometimes accounted for the entire amouat spent on health (that is, no capital expenditures were made). Within the recurrent budgets, the proportion allocated to personnel was very high, accounting for up to 99 percent of recurrent expenditure in some LGAs, with little remaining for other operating expenses. As federal revenue sharing brings more resources to the LGA health systems, these resources should be used to achieve the Ministry's target of 60 percent of recurrent expenditure (maximum) being spent on personnel. Only with the complementary resources financed by the nonpersonnel components of the budget can the local health personnel be productive. D. Capital Expenditure 3.25 Federal. Although nominal recurrent FMOH expenditures have shown a fairly steady increase, nominal capital expenditures have been much leas regular. The contrast between the behavior of these two aggregates is most apparent during 1981-83, when the FMOH implemented a major hospital construction program. _ 38 - Table 3.7 FEDERAL CAPITAL EXPENDITURE APPROPRIATION TO HEALTH (Actuals, Millions of Naira) Federal Program 1981 1982 1983 1984 1985 1986 Hospital Programs 82.0 101.6 91.9 25.3 31.1 61.5 Trairing 5.3 2.5 4.3 2.4 2.1 2.5 Mfg. Drugs & Vaccines 1.0 1.0 1.2 3.9 0.8 1.6 Laboratories 1.4 1.6 1.1 0.9 0.4 0.8 Primary Health Care 50.9 11.0 14.3 8.5 8.4 42.5 Communicable Disease 1.6 2.8 3.8 1.7 2.7 10.8 Other Specific Programs n.a. n.a. n.a. 0.1 0.1 11.7 TOTAL CAPITAL BUDGCT 142.1 120.5 116.6 42.8 45.5 131.5 Source: FMOH. Conclusion 3.26 This chapter has examined in detail public resources for health during the period 1981-85, and later data show a continuation of the trends then set in motion. The nominal recurrent budget for the FMOH has shown a fairly steady increase, more than doubling between 1981 and 1990, but in real terms it is only one half of its previous value. The capital budget has fallen more drastically, to only one quarter of its value, and shows more variability. But the federal budget is perhaps only one third of total public sector expenditures on health, and indications are that all three tiers of the sector have suffered from the current constrained economic situation. 3.27 With the recent sharp decline in resources, health policy has, of necessity, been more clearly defined. In accord with the National Health Policy of 1988, high priority has been placed on preventive health care rather than curative and on improving the efficiency of public sector services, including the provision of an adequate supply of drugs and better maintenance. 3.28 The current trend toward rechanneling federal revenue through state and local budgets can be expected to produce a slight reduction in federal health expenditures which could be more than offset by substantial increases in state and local expenditures on health. This change offers the promise of improved supplies and services at the theoretically more accountable local govermment health facilities. But careful planning will be necessary to assure that the insreased resources redress the balance between the nonpersonnel and personnel components of the state and LGA health budgets. - 39 - 3.29 The National Health Policy highlights the fact that the health care system is a comprehensive one, which supports community involvement and collaboration with nongovernmental providers of health care. A possible policy response to help cushion health care budgets against future shocks is to define, within the recurrent budget, both "core" and "noncore" expenditure categories. "Core" categories of activities, which might include essential public health services that generate significant positive externalities such as immunization and ante-natal care, would be fully funded regardless of budgetary eventualities. The "noncorel categories could include certain curative services for which funding would be reduced in times of need. To facilitate flexibility in the personnel budgets for noncore health care activities, the government could contract those services out each year to local providers.49/ 3.30 In subsequent chapters of this report, the health care system in one state, Ogun, is examined, and the strength of both the public and the private health care sectors assessed. Based on this assessment, the opportunities for future development of the provision of health care are identified, including possible action in areas such as personnel, drug supply and cost recovery. 49/ The size of these contracts would vary annually according to the governments budget that year. This policy option would be difficult to implement in states with few physicians in the private sector, but Ogun State, the focus of the remaining chapters of this report, might be a particularly apt place to begin such policy experiments. - 40 - NIGERIA: HEALTH CARE COST, FINANCING AND UTILIZATIOIJ IV. HEALTH CARE IN OGUN STATE A. Description of Ogun State 4.01 Introduction. This review of Ogun State's health situation is intended tc serve as a point of reference for the Nigerian health system as a whole. The discussion will strive to assess the extent to which the findings in Ogun State may be relevant for the rest of Nigeria. This chapter provides a detailed description of the State and its health care facilities. Chapter V then uses the survey results from Ogun State to provide detailed analyses of facility operating costs, household demand for health care, and cost-recovery opportunities. 4.02 Background. Created in 1976 as one of three states from the former Western State, Ogun State comprises the old provinces of Abeokuta and Ijebu and covers an area of 16,410 sq.km. Strategically situated to the north of the federal capital, Lagos, Ogun State borders the Republic of Benin to the east, Oyo State to the north, and Lagos State and the Atlantic Ocean to the south. The state capital is Abeokuta. (See Map IBRD 22161R) 4.03 Population. Ogun State's 2.85 million people (1986 estimate) are largely a rural (about 70 percent live in rural areas) ard Yoruba-speaking population.50/ The density of 174 persons per sq.km. is, however, significantly above the average for Nigeria (127 per sq.km.). At least 45 percent of the population is under 15 years old. Of the sample population of 20,600 people in 4,700 households, slightly more than half (53 percent) are Christians, while most of the remainder (46 percent) are Muslim. About 53 percent of household members are single, 46 percent are married, and 1 percent are separated or divorced. Employed persons constituted 52 percent of the sample population. English is widely spoken, and citizens from all parts of Nigeria and foreigners from various countries are represented. Seventy percent of respondents reported monthly income of less than N100 and only 6 percent above 1200, but underestimation of income is usual in such a survey.51/ 4.04 Ogun State ie divided into ten Local Governinert Authoritils (LGAs), listed here with their caritals in brackets: Abeokuta (Abeokuta] Egbado North (Aiyetoro] Egbado South [Ilaro] Ijebu-Ode [Ijebu-Odel Ijebu-Remo [Sagamu] Ijebu North [Ijebu-Igbo] Ijebu East (Ogbere] Odeda [Odeda] Ifo/Ota (Ota] Obafemi/Owode (Owodel 5O/ The main groups are the Egba, Egbado, Ijebu, Ijebu-Remo, Awari, and Egun. C£ 11 ' ' s ' ' ' ' t < thP ' WI 'ho - 41 - About two-thirds of the population live in four local government areas: Abeokuta, Egbado, Ijebu-Ode, and Ijebu-Remo. More than half the population of Abeokuta, Ijebu-Remo and Ijebu North live in cities or towns. On the other hand, Odeda, Obafemi/Owode, and Ijebu East are almost entirely rural. The 1986 Ogun State Gazette classified 15 towns as urban, 20 as semi-urban, and the remaining villages and hamlets as rural. 4.05 Agriculture and ,t-ustry. Farming is the principal occupation of Ogun's residents, who raise rice, maize, plantains, cassava, yams and bananas. Cash crops include cocoa, kolanut, rubber, palm oil, palm kernel and logging. Ogun State is the largest producer of kolanuts in Nigeria. Farming and Ogunls mineral resources (such as limestone, chalk, phosphates, clay and stone) have encouraged the growth of cement, food canning, foam rubber, and paint manufacturing industries. Relatively less developed than agriculture or industry, the service sector concentrates on commerce, hotels, restaurants and health services. 4.06 Infrastructure. Largely owing to its location, Ogun is relatively better developed than most Nigerian states, and its residents probably enjoy a higher standard of such social amenities as schools, roads and health services than the average. The state possesses a good basic infrastructure, although this has suffered through lack of maintenance in recent years. The Lagos-Ibadan expressway passes through Ogun, as do four other large roads linking Ogun with the northern and eastern Nigerian states. Yet many rural localities are all but inaccessible during the rainy season because the roads become impassable. The main railway line from Lagos to other points in Nigeria passes through Ogun, and the Lafenwa station in Abeokuta has become a major market center. Most urban towns have electricity, and most of those without should have it within a few years. Water for industry is already available in all the urban towns, and the State Water Corporation is currently designing and implementing an expansion of the piped water system. 4.07 Government. The highest policy-maki-g body in Ogun is the State Executive Council, which is chaired by the Mil tary Governor with assistance from 10 Civil Commissioners. The Civil Commissioners each head a state ministry, aided by a Director General. As part of the recent civil service reforms, the Commissioner has become the Chief Executive Officer accountable for the ministry finances. Each LGA had a Chairman and a Chief Executive assisted by four part-time staffers, but as part of the transition to civilian rule, LGAs now have a sole administrator until local elections are held in 1990. For health purposes, the state is divided into five zones: Abeokuta (Abeokuta, Odeda, Obafemi/Owode) Ifo/Ota Ilaro (Egbado North, Egbado South) Ijebu (Ijebu-Ode, Ijebu North, Ijebu East) Sagamu (Ijebu-Remo) - 42 - 4.08 Health care services are provided by the State Ministry of Health (SMOH), the State Health Management Board (SHMB), and the Local Government Authority (LGA). Chart 4.1 shows the organization of the health agencies in a typical state. The functions of the SMOH includes (1) General formulation and control of health policies; (2) Legislation relating to health matters: (3) Supervision of medical and health training and research institutions in the State including: - the Central Health Unit for training and research; - the School of Health Technology; - all Nursing and Midwifery schools. (4) The direct control of the statistical and epidemiological unit; (5) External aid for health services in the State; (6) Pharmaceutical inspection and licensing control; (7) Registration and inspection of private hospitals and maternity homes; (8) Monitoring the activities of the State Health Board and the Health Zones; (9) Representation at the federal level on all health matters and bodies including the Nigeria Medical Council, the Nursing and Midwifery Council of Nigeria, the Pharmaceutical Board of Nigeria, the Royal Soc'ety of Health, the West African Examinations Board and the National Council on Health; (10) Liaison with the Federal Ministry of Health and international agencies and organizations on health matters; (11) Procurement of drugs for and administration of the drug revolving fund scheme. The 3HMB manages and controls all state-level medical facilities and their staffs with the LGAs now responsible for providing the primary health care services which are facility-based (mainly primary health centers, dispensaries and maternities), including the appointment, posting and disciplining of the relevant staffs. The main functions of the SHMB ares 52/ (1) To manage and control all health institutions taking charge of activities and preventive services of the State Government; (2) To execute the general health policies formulated by the Ministry of Health and approved by the Government; (3) To provide and standardize all hospital equipment in the State; (4) To appoint, post and discipline health personnel from grade 07; (5) To promote health personnel from Grade Level 07 subject to the approval of the Military Governor; (6) To appoint, confirm, promote and discipline administrative employees Grade Level 01 - 06; (7) To consider and collect the Annual Estimates of the Zonal Health Boards; 52/ In 1990, responsibility passed to the LGAs for appointing and paying staff of primary health care centers now under their jurisdiction. - 43 - (8) To advise on revenue generation within and in keeping with Government Policies on Health Services; (9) To initiate and control building programs of hospitals and health irstitutions in accordance with government policy and to maintain buildings and equipment use in health institutions; (10) To purchase equipment and food to be used in health institutions. - 44 - CHART 4.1 Structure of Hoalth AEncies In a Typical State I/ I .. Co""l-ioner ~~Other Comil wioners/ Local Government Uniti Of He91th Ministries in Governor 's Offieo Ministry of Health or Hoepital of Health 1/ Monagement Soard State-operated State-opered training health Local Government Authorities (about 4) institutions (hospitals, health _ e~~~~~~enters, ae.) LOA Chairman Hoalth and Social Other Welfare Councillor Council,ors Snit aion Pr ;Xary Health Care and Helth Facilities A Progrea Insocton 1/ Consists of eight departments as a result of 1988 Civil Service Reform: ntate ministries follow pattern of the Federal Ministry of Health (See Annex A). - 45 - 4.09 Health Status of Ogun State. 53/ Ogun State's general health picture closely approximates that of the rest of qTigeria. The estimated total fertility rate of 6.2 is somewhat lower in Ogun State than in the rest of Nigeria (6.6),54/ and infant mortality rates are similar, estimated at 90 to 120 per 1,000 live births (compared with the estimate of 103 for Nigeria). At least 25 percent of children die before their fifth birthday. Death before age six is estimated to account for one half of all deaths. The main causes of morbidity and mortality are malaria, diarrhea, respiratory and infectious diseases, and accidents. Malaria alone accounted for more t'ran 75 percent of cases at comprehensive health centers in 1985, while dysentery, measles, pneumonia and whooping cough each represented 7 percent of cases. The vast majority of these deaths could be prevented by immunization, better hygiene, better nutrition, and early primary care intervention (such as oral rehydration therapy). Maternal mortality would decrease sharply with better prenatal and midwifery services. The principal group with preventable mortality in Ogun State -- childbearing women and infants and young children -- could be greatly aided by primary care methods already in existence. B. State Financial Resources for Health 55/ 4.10 State Government Revenue. The financial resources of Nigerian states come from a combination of internally generated funds and statutory allocations from the Federation Account. Until 1984, when revenue diversification began in earnest, the federal allocation was the primary revenue source for the Ogun State budget. It fell from over 90 rercent in 1981 to around two-thirds of total resources available in 1986. Internally generated revenue has come from various sources, a declining share being taxes (60 percent in 1986). The share of SMOH and SHMB reached 2.8 percent in 1985 as against 0.7 percent in 1982. The shift toward better management practices in the mid-1980s transformed state finances, mainly by achieving a better balance between revenue and expernditures, moving the State from deficit to surplus (see Table 4.1). 4.11 Revenue -- State Health Sector. The dramatic rise in the share of internally generated income in state resources was triggered by the decline in the amount of federal disbursements, which occurred because of falling oil revenues and rising federal indebtedness. Federal contributions to st.ate revenues declined steadily during 1981-84 before 53/ Ogun State Restructured Draft State Health Plan, 1987-1991, Ogun State, Ministry of Health, 1986. 54/ In 1988, the World Bank estimated a total fertility rate of 6.6. As both are estimates based on limited data, the difference is not considered significant. 55/ Household expenditures on health and private resources are not considered because of lack of data. - 46 - increasing in 1985. Not surprisingly, state revenues were adversely affected during this period, but they made a good recovery after 1984. 4.12 The bulk of Ogun State's health sector revenue comes from user fees, which accounted for 84 percent of such revenue in 1986. The most productive user fees are for drugs, prenatal care, and laboratory tests. Additional user expenses are bed fees, catering, ambulance fees, mortuary fees and laboratory fees. Other charges include fees for tenders, examinations, student boarding, patent medicine licenses and private facility registration, the last two being the most important. State health revenue in 1986 exceeded N1.1 million. This figure represented a 150 percent increase over 1984 state revenue. Table 4.1 OMUN STATE: REVENUE AND EXPENDITURE (ACTUAL) (Percentages) 1981 1982 1983 1984 1986 Revenue Internally generated 7.8 18.9 19.6 32.6 82.4 of which U HerIth 0.7 0.8 0.6 0.7 0.8 U Health Management Board 0.0 0.04 0.6 1.8 2.6 Federal Allocation 92.2 86.1 80.5 67.4 67.6 TOTAL 100.0 100.0 100.0 100.0 100.0 Expenditure Capital 86.6 22.6 19.6 14.0 of which U Health 1.9 1.4 0.9 1.5 X Health Management Board 2.8 1.6 1.1 1.1 Recurrent 68.2 77.4 80.4 88.0 of which X Health 2.8 2.8 2.6 1.9 X Health Management Board 6.7 9.6 9.9 12.1 TOTAL 100.0 100.0 100.0 100.0 Revenue as U of expenditure 68.0 81.2 78.9 118.9 148.1 Source: Ogun State Ministry of Health, Restructured Draft State Health Plan, 1987-91 4.13 The State Drug Revolving eund was introduced in 1984. Using a list of essential drugs and bulk purchasing for the supply of drugs to both primary health care units and hospitals, the Fund was established with the concept that all drugs prescribed (in both in-out patient settings and in hospitals) would be charged for. As of 1986, the Fund was reportedly yielding a profit and had increased the supply of drugs and services at all levels of government health facilities. This may be why a relatively high - 47 - proportion of the patients surveyed reported receiving drugs in public facilities (73 to 87 percent). 4.14 Revenue -- LGAs. Federal subventions from the Federation Account are the most important source of revenue for LGAs in Ogun State. These grants may reach 80 to 90 percent of a local povernment's actual yearly income, and in no case dip below 50 percent. Ijebu North LGA is the most dependent on federal funds. The only taxes that LGAs are permitted to levy under Nigeria's constitution are taxes on property called "tenement rates., Aside from tenement rates and federal grants, other sources of LGA income are from licenses, fees and rents. In addition, income from state grants, fees from health facilities, and investment income make up a minuscule proportion of LGA revenue in most cases. With the exception of Egbado South, which suffered a 50 percent decline, LGA income rose quite substantially during 1984-86, in most cases by more than 50 percent. In some cases, the rise was due to internally generated revenue. Most LGAs are financially prudent, spending somewhat less than their income. 4.15 State Expenditure Patterns. The expenditure pattern of Ogun State has shown wide variation in capital while recurrent levels have been sustained. Personnel costs especially have claimed an increasing proportion of recurrent expenditures in the public health sector, rising from 74 percent in 1981 to 98 percent in 1984. This is in contrast to the average expenditure on salaries for the State of 70 percent of the recurrent budget, rising to 82 percent.56/ Education, the largest sector, accounted for about 40 percent of actual recurrent expenditure during 1980-86 in Ogun State, followed by health at around 12 percent. 4.16 Health Expenditure -- LGAs. Like the State Government, the LGAs in Ogun State practiced prudent financial management during the early 198309,57/ running surpluses in 1984 and 1985. Personnel costs accounted for around 80 percent of LGA recurrent health expenditure during 1984-86. Health received the greatest share of recurrent expenditure from most LGAs during the same period, about one-third. This is not surprising because LGAs are responsible for operating the State's primary health care facilities, dispensaries and maternities. C. Modern Health Services in Ogun State 4.17 Types of Facilit.'es. Ogun State's 445 modern health care facilities are heavily weighted toward primary or secondary care. Primary care facilities typically provide the first point of contact between the patient and the health care delivery system. Secondary care institutions (mainly hospitals and comprehensive health centers), with their larger staffs and better equipment, accept referrals from primary care providers and emergency cases that the primary institutions cannot handle. Tertiary care institutions in Oguu State include the University Teaching Hospital at 56/ In 1986, 792 of public hospital recurrent expenditure went on personnel;'the figure was 82? in public maternities. 57/ The period for which data were collected. - 48 - Sagamu and Aro Hospital at Abeokuta, which provide various types of specialized care. By its own count, the state has 38 hospitals, 9 comprehensive health centers, 56 health clinics, 17 primary health clinics, 164 maternities, 128 dispensaries and 34 other special facilities. In addition, there are many providers of traditional care. This group of practitioners includes herbalists, traditional birth attendants, spiritual healers of various religious denominations, and other alternate sources of health care. Attempts are being made to identify and register herbalists and traditional birth attendants. The SMOH has a committee made up of traditional health practitioners who are charged with identifying, standardizing and coordinating this type of care with the modern health care system. 4.18 Ownership of Health Facilities. About 70 percent of health institutions are publicly owned; 30 percent are privately owned. Local governments own about 58 percent of all health facilities (mainly dispensaries and maternities); state governments own only about 10 percent. Private facilities exist at primary, secondary and tertiary levels, but within each level the market share of private care providers varies greatly. For example, two-thirds of all hospitals are public but two- thirds of comprehensive health centers are private. Most maternities (68 percent) are public, while basic health centers are almost evenly divided (57 percent public-43 percent private). Despite the dominance of public over private hospitals, both in number and size, only 57 percent of available beds are in the public sector. Among primary health facilities (dispensaries, maternities and basic health centers), 72 percent are public. Of health personnel, about 30 percent are under private control; 50 percent of physicians are private sector employees. 4.19 In the private sector, health facilities can be grouped under three main headings: for-profit, nonprofit and missions. Of the three, private for-profit institutions account for 91 percent of private facilities. Mission facilities account for 2 percent of private facilities, while private nonprofit institutions comprise the remaining 7 percent. Public facilities are owned by the federal, state or local government. Some 84 percent of all public facilities are owned by LGAs, the great majority being maternities and dispensaries. The federal government owns three comprehensive health centers in Ogun State, or 1 percent of public facilities, while the State Government owns 15 percent of the public facilities. 4.20 The public sector also operates various program- or disease- specific health institutions, such as dental centers, chest clinics, leprosy control units, family health centers, and maternal and child health units. 4.21 Facility Services. The services offered vary according to the type of facility. With respect to inpatient care, both public and private hospitals are heavily oriented toward maternity cases. Comprehensive health centers and health clinics admit a smaller proportion of patients for maternity care and instead concentrate on surgery, infectious diseases and general care. Primary health care centers and dispensaries offer - 49 - almost no inpatient caret if they do, it tends to be for maternity cases (see Table 4.2). Table 4.2 TYPES OF INPATIENT ADMISSIONS BY FACILITY TYPE Infectious Total Type of facility Deliveries Surgery Disease Other Admissions Hospitals Public 46 32 7 15 3318 Private 49 18 13 20 1791 Comp. Health Centers Public 29 21 14 37 836 Private 11 2 6 82 325 Maternities Public 100 - - - 305 Private 89 1 2 7 203 Primary Health Care Public 50 - 2 48 278 Dispensaries Public 100 - - - 9 Health Clinics Private 2 23 37 38 220 Sources AHCS. 4.22 Most care offered by outpatient services (except at hospitals and dispensaries) is preventive. This is especially true of health clinics and maternities, which classify 90 percent of patient visits as preventive (see Table 4.3). - 50 - Table 4.3 TYPES OF OUTPATIENT VISITS BY FACILITY TYPE -- PUBLIC VS. PRIVATE Type of facility a/ Preventive Curative Total visits visits visits (Z) (Z) (I) Hospitals Public 41 59 22,760 Private 48 52 24,706 Hatermities Public 97 3 6,228 Private 92 8 2,856 Source: AHCS. a/ Data for comparisons between public and private facilities are available only for hospitals and maternities. 4.23 Distribution of Health Facilities: Public and Private. In the urban areas of Ogun State private health facilities'outnumber their public counterparts. In fact, 82 percent of private health facilities are located in urban areas (presumably because of profit considerations), and 57 percent of urban health facilities are privately owned. Government care providers tend, though not exclusively, to locate in rural areas where private care providers are fewest. Almost 75 percent of public care institutions are rural. Around 60 percent of private facilities are located in Abeokuta, Ifo-Ota, and Ijebu North; on the other hand, 56 percent of public facilities are found in Egbado South, Abeokuta, and Ijebu-Remo. Obafemi/Owode has no private health institutions at all. Odeda, Obafemi/Owode and Egbado North have no secondary or tertiary health care providers. Although Ogun State's urban/rural population ratio is 30:70, about 44 percent of all facilities are in urban areas while 56 percent are rural. This implies that urban residents enjoy a bias in their favor in the distribution of health facilities in Ogun, a situation typical of Nigeria and o her developing countries. Compounding this inequity are the poor state of rural roads and the lack of reliable motor transport to rural areas. Rural residents vho require more than primary care thus have to pay more money to travel longer distances in less comfort and safety than urban residents. 4.24 Bed Distribution. The AHCS survey reported that Ogun State has about 4,000 beds in health facilities of all types. About half of these are devoted to maternity; about 25 percent are pediatric beds; about 22 percent are used by surgical/medical patients; and the remaining 3 percent are allocated to patients suffering from infectious diseases and other _ 51 - ailments. Sixty-three percent of all facilities are located in the Abeokuta, Ifo-Ota, Egbado South, Ijebu-Ode, and Ijebu-Remo LCAs. These areas are home to about 60 percent of Ogun's population but account for about 75 percent of the available hospital and clinic beds. At the other extreme, Odeda, Obafemi/Owode, and Egbado North have the lowest number of beds. These three LGAs have about 25 percent of the population and 20 percent of the facilities but only 11 percent of the beds. While some concentration of beds in urban areas reflects a deliberate policy based on the referral system, relative access must also be considered. 4.25 Distribution of Health Personnel -- Geographic. Ogun State has around 400 practicing doctors (a 1:7,100 ratio) and almost 200 pharmacists (a 1:15,000 ratio). Nurses and midwives make up 25 pe.cent of health professionals (a 1:5,000 ratio). Since the onset of economic austerity in 1984, including a hiring freeze, employment in the public health sector has declined. For example, during the first six months of 1986, 287 of 1,488 government nurses and midwives left their jobs. Many government health workers who lost their jobs found work in the private sector, so the general availability of health manpower was not greatly affected. Although statistics are not available, it appears that far fewer nonprofessional workers have left their public sector jobs. This may help to account for the growing proportion of salary outlays going to service, maintenance and support personnel who are not actually engaged in providing health care. 4.26 The urban-biased pattern which prevails in the distribution of health facilities also obtains in the distribution of health personnel. However, facilities appear to be more evenly distributed than health personnel, which is perhaps a better measure of health service availability. Abeokuta, Ifo-Ota, Ijebu-Ode, and Ijebu-Remo LGAs, with 47 percent of the population, have almost 75 percent of the doctors, 82 percent of the nurses and midwives, and 69 percent of all health workers. On the other hand, Odeda, Obafemi/Owode, Egbado North, and Egbado South LGAs, with 39 percent of the population, have 11 percent of the doctors, 12 percent of nurses and midwives, and 18 percent of all health workers. Egbado South, with 14 percent of the population, has many lightly staffed dispensaries but no hospitals, perhaps explaining why it has relatively many facilities (18 percent of the total) but relatively few personnel (8 percent of the total). These results should be interpreted carefully in light of two other points. Population estimates currently available are questionable because they are projected from the 1963 census. In addition, rapidly growing urban centers may include a growing proportion of the population in the future. The distribution of facilities and beds compared with population is shown in Table 4.4. 4.27 Distribution of Health Personnel -- Public and Private Sectors. Considering that the private sector owns only 30 percent of all health facilities, it has had great success at attracting doctors. Roughly half of all doctors work in private sector jobs. One explanation for this disparity may be double counting, because it appears that many publicly - 52 - employed doctors also have private practices, especially in urban settings.58/ The ratio of public to private sector employees in other job categories is more in line with the 70:30 ownership ratio: 66:34 among nurses and midwives and 61:39 among support staff. The ratios get skewed again for community and public health workers (95:5) and pharmacists (83:17). The State Government, which owns only 10 percent of the health facilities in the state, employs 46 percent of the health personnel. The private sector owns about 30 percent of facilities and employs roughly the same proportion of health workers (31 percent). The LGAs own 50 percent of all of Ogun's health facilities but employ only 20 percent of its health workers.59/ This is not surprising because LGA-owned dispensaries are very lightly staffed. Finally, the Federal Government employs only 1 percent of Ogun's health workers. Table 4.4 OGUN STATE: DISTRIBUTION OF FACILITIES AND BEDS COMPARED TO POPULATION (Percentage) Population Facilities Beds Abeokuta 14.8 13.5 23.5 Ifo/Ota 9.5 12.2 10.9 Odeda 7.5 7.0 2.5 Obafemi/Owode 8.8 5.9 3.3 Egbado North 8.4 6.5 5.0 Egbado South 13.9 17.3 13.9 Ijebu-Ode 12.2 14.9 13.3 Ijebu North 7.4 6.8 7.8 Ijebu East 7.5 5.0 6.9 Ijebu-Remo 10.0 11.3 13.0 TOTAL 100.0 100.0 100.0 (2,847,511) (445) (3,969) Source: AHCS. 58/ Although Ogun State in 1986 had 185 registered doctors (87 in the public sector and 98 in the private) the survey results reported 382 doctors (191 each in the public and private sectors). 591 This will shift under the 1990 changes in LGA responsibilities. - 53 - 4.28 Staffing mix -- ,tiblic and Private Sector. Staffing mix at health facilities is measured by analyzing the number of staff per admission (or visit) and the relative numbers of doctors, senior health workers, and non-health workers. With two exceptions, private facilities have higher staff per admission and per visit ratios than public facilities. This is true with respect to senior medical profescionals and to health workers in general. The exceptions are physicians per admission and health worker per visit in small hospitals where the public facilities have higher ratios. (While suggesting that private facilities offer a higher level of care than public facilities, the data may actually reflect underutilization.) No clear pattern emerges in comparing ratios of health with non-health workers at public and private facilities. Of workers at public health facilities, 59 percent are doctors, nurses, pharmacists or other professionals; the remaining 41 percent are support staff. By comparison, the proportions are 52 percent and 48 percent respectively in the private sector. Except in small hospitals, there are more physicians per health worker in private than public facilities. The ratio of senior health workers to health workers indicates that private facilities depend more on senior health workers in all facility types (see Table 4.5). 4.29 Availability of Equipment - Public and Private Sector. Inventories for all kinds cf equipment -- X-ray machines, operating rooms, laboratories, and vehicles -- are higher in private than public facilities. In addition, private facilities tend to be newer than their public counterparts. For instance, while public hospitals are on average 28 years old, private ones average only 20 years old; public comprehensive health centers were built on average 32 years ago, but private ones were built only 7 years ago; public maternities tend to be about 25 years old, but private maternities average 18 years old. Equipment at private and public institutions is not always usable, however, due to maintenance problems and lack of spare parts: at any given time, only 63 percent of the X-ray machines, around 83 percent of the laboratories and operating rooms, and 79 percent of the two-wheel-drive vehicles are in usable condition. - 54 - Table 4.5 STAFFING MIX FOR PUBLIC AND PRIVATE FACILITIES: RATIOS (Standard Deviation in Parentheses) Senior a/ Health Health Facilities Physician/ Worker/ Worker/ Operational Health Health b/ Non-Health Classification Worker Worker Worker Small Hospital Public 0.14(0.02) 0.97(0.02) 1.80 (0.52) Private 0.13(0.01) 0.99(0.1) 0.75 (0.67) Comprehensive Health Center Public 0.17(0.15) 0.82(0.18) 1.87 (0.66) Private 0.26(0.10) 0.96(0.06) 1.29 (0.79) Basic Health Clinic Public 0.23(0.25) 0.60(0.24) 1.44 (1.50) Private 0.26(0.18) 0.96(0.09) 2.03 (1.81) Maternity Public 0.09(0.17) 0.68(0.37) 0.91 (0.84) Private 0.22(0.19) 0.94(0.11) 1.63 (1.30) Dispensary Public o(o) 0.05(0.19) 2.05(11.95) Private 0(0) 1.00 (0) 1.00 (1.41) Source: Calculations based on data in Ogun State Health Financing Study, AHCS. a/ Senior health worker = doctors + nurses + midwives. b/ Health workers = senior health workers + community health workers + pharmacists + public health workers. 4.30 Fees and Pricing. As Table 4.6 shows, private facilities clearly charge more for every form of health service than public facilities. Moreover, in general, they do a better job than their public counterparts in recovering their operating costs, including drug costs, from patients. Fees in public facilities are far lower than fees in private facilities. Lower charges, however, mean lower revenues. This implies that public institutions, squeezed as they are by an ever larger proportion of expenditure for personnel, have even less money to buy drugs and other supplies. Patients are likely to avoid institutions where they know they will not be able to get treatment, even if the price is low. 4.31 The 1986 survey results showed that in Ogun State 92 percent of private patients were supplied with drugs, while about 80 percent (73 percent to 87 percent) of patients at public facilities had access to drugs. This is high relative to public facilities in other states and - 55 - reflects the fact that Ogun State had initiated a Drug kevolving Fund in 1984. Table 4.6 MINIMUM FEES AND CHARGES FOR SELECTED HEALTH SERVICES (In Naira) Type of Service Private Facilities Public Facilities (lowest, highest) (lowest, highest) Registration -- maternity (2, 25) (1, 20+) Registration - outpatient (1, 15) (0.50, 2) Registration -- inpatient (1, 50) (1, 10) Immunization (1. 5) Free Paracetamol (.1, 1.2) (.02, .9) Contraceptive pills (3, 15) Free Condoms (.30, 3) Free Laboratory -- CBC (1.5, 30) (2, 25) Laboratory -- Urinalysis (1.5, 30) (2, 5) Chest X-ray (10, 25) (10, 10) Prenatal care (per visit) (10, 100) (20+, 2) Delivery, basic (2, 80) (20, 2)) Delivery, sections (60, 650) (40, 40) Appendectomy (60, 500) (20, 40) Room/board (per day) (3, 12) (.50, 2) Dressings (1, 25) (.50, 10) Malaria (full treatment) (0.50, 40) (0.50, 10) Setting broken arm (2, 100) (5, 20) Sourcet AHCSe - 56 - Conclusion 4.32 Within Ogun State, both the State Government and LGAs experienced good financial management during the mid-198Os. The State increased revenues by 150 percent during 1984-86, and the LGAs operated at a surplus. However, the high proportion of expenditure on personnel squeezed the funds available for drugs, supplies and maintenance, although the introduction of a drug revolving fund has facilitated public sector drug supplies. The availability of drugs is repeatedly cited as the factor of primary importance in attracting patients to public sector facilities. 4.33 Clearly the private health sector plays an important role in Ogun State. Around a third of Ogun's facilities and personnel are privately controlled, and this percentage may reach 50 percent for doctors and senior health workers. Patients at private facilities may get better care, mainly because inventories of all types of equipment and supplies (including drugs) are higher than in the public sector. In addition, public facilities are older than private ones. The private sector controls a large part of secondary care facilities, such as comprehensive health centers. The public health sector dominates primary care facilities, such as dispensaries, and maternities. Private facilities predominate in the towns, but they are practically nonexistent in some parts of the countryside. 4.34 Ogun State's health care system has the expected urban bias, which is more evident in personnel than in facilities. City dwellers have much easier and wider access to health services of all types than people in rural areas. In some places the nearest clinic is over 20 km. away. another possible problem area in the public health sector is staffing. Most private facilities have higher staff per admission and per visit ratios than public facilities and depend more on senior health workers. In addition, as discussed in Chapter V (paragraphs 5.20 and 5.21), utilization levels appear to be low in both the public and the private sector. The analysis of facility efficiency used as a minimum acceptable level 600 visits per health worker per year -- a very low level, which, if evenly spread would amount to an average of around two visits per day.60/ 4.35 Ogun State shares these concerns with other parts of Nigeria. In the following chapter, the problems are examined in detail with a view to identifying appropriate policy adjustments to address ther 60/ Initially, a moderate level of efficiency of 800 visits per health worker per year was the norm selected, but over half of the 46 facilities in the statistical analysis failed to achieve this and the level was lowered to 600, this being considered the minimum acceptable level of efficiency. - 57 - NIGERIA: HEALTH CAKE COST, FINANCING AND UTILIZATION V. HEALTH CARE COSTS IY; OGUN STATE 611 Introduction 5.01 The health care sector in Nigeria faces the problem of how to provide Improved services in a period of budget constraints. As improvements are costly, investments must be carefully selected in order to provide the most service for the least funding. Using a comprehensive survey of health care in Ogun State, opportunities for facilities to provide better care that more closely responds to patients' needs have been identified. The analysis is presented in three sections. In Part A, health facilities in both the public and private sector are analyzed with respect to technical efficiency and costs. In Part B, the household demand for health care is examined, to understand better where prospective patients look for services and how much they are prepared to pay for them. In Part C, the information on facilities' supply of health care and the household demand for it are combined in a series of cost recovery simulations. These simulations are aimed at providing policymakers with a better understanding of cost effective ways to improve health care services. A. Facility Analysis 5.02 After 1970, Nigeria considerably expanded its public and private health care system and improved the ratio of health facilities and personnel to population. However, since the early 19808, a marked deterioration in the quality of the public health facilities and their utilizatior has occurred. This deterioration stems from limits on public spending due to economic constraints which have sharply reduced expenditure on maintenance, supplies and equipment. 5.03 Part A of this chapter examines issues surrounding the cost of health care, using survey data collected in Ogun State: (a) the financial status of both public and private providers; (b) the cost structure of health care service provision, in both the public and private sectors; (c) efficiency in the public health sector; (d) whether nonpublic providers of health care (such as private for-profit and not-for-profit providers and missionaries) are more efficient than public providers. 61/ The analysis in this chapter is based on the survey work carried out in 1987 by African Health Consultancy Services (ARCS), Lagos, sumaarized in Ozun State Health Financing Study, submitted to the Federal Ministry of Health and the Ogun State Ministry of Health, March 1989. - 58 - 5.04 The facility analysis is in two sections. The first section is a descriptive analysis of recurrent costs, revenues, gross recurrent profit, and cost recovery for public and private facilities. The analysis focuses on small hospitals and maternities, mainly because there are enough of these types in both the public and the private sector to permit comparisons. Financial trends during the period 1985-86 are examined in as much detail as the data allow. The second section discusses the conceptual framework and empirical results of a detailed statistical analysis of efficiency in public and privace facilities. (Further background on the data set, methodology and estimation teLlhniques is presented in Annex G.)62/ 5.05 A Note on Nomenclature. Nomenclature can be problematic because it tends to vary between public and private institutions and because some institutions have not followed their original mandate. In Chapter IV, the discussion of the health sector in Ogun State used the definitions employed in the State. For the statistical analysis in this chapter, the raw data were reviewed to ensure accurate classification.63/ Operational definitiors, applicable to both public and private facilities, were developed to identify six major types of facilities depending on actual services performed. These are listed in Chart 5.1 below. 5.06 Technical definitions: (a) Technical efficiency uses minimum quantities of physical inputs to produce a given amount of services; (b) Cost-efficient production uses technically efficient combinations of physical inputs that are also the least costly combinations of inputs.64/ 62/ Technical efficiency is examined by estimating a frontier production function (the relationship between personnel and nonpersonnel inputs and the amount of various services produced in a health care facility). Allocative efficiency (cost efficiency) is analyzed by determining whether staffing patterns within a given facility yield a given amount of services at minimum cost. Deviations from cost minimization or cost efficiency are measured by comparing the marginal product of health workers and non-health workers with the relative wages of health workers and non-health workers. The statistical analysis of efficiency concludes by estimating a cost function: the relationship of total recurrent costs to utilization levels, personnel prices and capital stock. The cost equation is used to determine the costs associated with allocative inefficiency, the average and marginal costs for inpatient and outpatient services, and to examine economies of scale and scope. 63/ For example, although a facility may be designated as a comprehensive health clinic with 60 beds, if no doctor is available the facility should more appropriately be called a basic health clinic. Me ,. I - 59 - Chart 5.1 CONTRASTING NOMENCLATURE OF OGUN STATE HEALTH UNITS Analysis in Chapter V Ogun State Ministry of Health Dispensary No physicians; maximum of 1 bed. No physicians; no beds. Staffed by dispensary assistant. Staffed by dispensary Minor ailments are seen; assistant. Minor ailments simple treatments are available. are seen; simple treat- ments undertaken. Basic Health Clinic About 1 physician and 10 beds. Staffed by general prac- Minor curative care and tioners. Up to 14 beds. deliveries can be performed. Minor curative care and operations can be performed. Comprehensive Health Center 2-3 physicians; about 20 beds. Up to 30 beds. Rural Minor routine operations can based. Staffed by at be performed. Some are rural least one public health/ based. community health physician. Maternity No change in nomenclature. No resident physicians; about 7 beds. Staffed by midwives and ward maids. Small Hospital District Hospital 4-6 physicians; 60-120 beds. Serves 50,000 people; 40-60 beds; more than one physician; sophisticated equipment. Large Hospital General and State Hospitals Services of generalist and Serves 100-250,000 people in semi- specialist doctors, nurses, urban areas; 50-200 beds. More than 4 and health workers; physicians. sophisticated equipment. More than 150 beds. - 60 - 5.07 Sample. A detailed health facility survey instrument was used on a stratified random sample of 86 health facilities out of 445 facilities identified and surveyed in less detail by a complete census. The collected data detailed staffing, equipment, cost, drugs, prices charged, etc. (See Annex H). The types of facilities selected and number of survey responses (totaling 68) retained are presented in Table 5.1. Table 5.1 NUMBER OF SELECTED HEALTH FACILITIES (BY TYPE/OWNERSHIP) AND SURVEY RESPONSE RATES Number Number that Type/Ownership Contacted Responded Private maternities 11 9 Private hospitals a/ 13 8 Private clinics 11 6 Local government dispensaries 15 13 Local government maternities 16 15 Public comprehensive health center 2 2 Private comprehensive health center 1 1 Basic health center 6 4 Public hospitals b/ 10 9 Teaching hospitals 1 1 TOTAL 86 68 Source: Ogun State Health Financing Study. a/ Includes 2 maternity hospitals b/ Includes 1 maternity hospital 1. Financial Analysis of Public and Private Facilities, 1984-1986 5.08 A primary concern of policymakers is the financial sustainability of health care services. Government budget constraints have jeopardized availability and quality of service in public facilities but little is known about the financial status of private institutions. This section presents a descriptive analysis of recurrent costs, revenues, gross recurrent profit and cost recovery for both the public and private sector. Small hospitals and maternities are the focus of discussion because there are a sufficient number of these types in the two sectors for comparison. 5.09 Cost Analysis of Public and Private Facilities. Personnel costs account for the greatest proportion of recurrent costs in all facilities in Ogun State. This is especially true for public facilities. In 1986, personnel costs were 79 percent and 82 percent of public hospital and maternity recurrent costs respectively. In private hospitals and maternities, personnel accounted for 57 percent and 60 percent of recurrent costs. The breakdown of recurrent costs - 61 - between personnel and nonpersonnel components for comparable facilities is presented in Table 5.2. 65/ 5..'0 Drugs accounted for the largest share of nonpersonnel recurrent costs. In 1986, private hospitals and maternities spent more of their recurrent costs on drugs (26 percent and 21 percent) than public hospitals or maternities (16 percent and 10 percent). Once personnel and drug costs are subtracted, little is left in budgets for other nonpersonnel items such as maintenance, transport and supplies. Private hospitals and maternities spent 17 to 18 percent of their budgets on nonpersonnel, nondrug iteme. On average, public facilities spent 5 to 8 percent of their budgets on these items. 5.11 This spending pattern is consistent with observations of health facilities in many other developing countries where inputs complementary to personnel are lacking and drug supplies are inadequate. The pattern suggests that public health care services in Ogun State are run inefficiently and unable to maintain quality. The pattern of expenditures was similar in 1985 and 1986, except for public hospitals, which seemed to increase expenditure on drugs, presumably reflecting the introduction of the drug revolving fund. In 1985, public hospitals paid 11 percent of recurrent expenses for drugs; by 1986, this figure had risen to 16 percent. To balance this increase, the share of public hospital personnel spending decreased. 65/ The survey provides some data on capital expenditures but contains no information on what activities were supported by these costs. - 62 - Table 5.2 DISTRIBUTION OF PERSONNEL AND NONPERSONNEL RECURRENT EXPENDITURE IN PUBLIC AND PRIVATE HEALTH FACILITIES, 1985-1986 a/ (Operational Classification) Small Hospital Maternity d/ Recurrent Cost Public Private Public Private Category 2 z 2 z #Observations 2 2 16 8 Personnel 1985 84 63 79 62 1986 79 57 82 60 Nonpersonnel b/ (exc. drugs) 1985 6 13 10 18 1986 5 17 8 18 Drugs 1985 11 24 11 20 1986 16 26 10 21 TOTAL PERCENT 100 100 100 100 TOTAL NAIRA c/ 1985 1,366,702 360,444 31,679 35,882 (maximum) (1,884,031) (441,839) (8,054,173) (92,280) 1986 1,488,049 414,433 26,202 32,087 (maximum) (2,120,880) (458,965) (7,973,952) (92,247) Sources Based on data in Ogun State Health Financing Study (AHCS). a/ Based on mean values for each facility type. b/ Other supplies, transport, maintenance and repair. c/ Maximum value in parenthesis. d/ Salaries 1986 (maternities) in naira: Puh4tc Private Average health worker 3, >3 2,454 Average non-health worker 1,301 662 - 63 - 5.12 Revenue Analysis of Public and Private Facilities. Revenues come from essentially five main sources: (a) government, (b) grants, (c) loans, (d) cost recovery, and (e) miscellaneous. Government s3urces include the federal, state and local governments. Grants consist of affiliation and private grants. Cost- recovery revenues can be generated through mechanisms such as user fees, drug revolving funds and private insurance. Table 5.3 shows how these various sources of revenues supported public and private hospitals and maternities during 1985- 86.66/ Table 5.3 SOURCE OF REVENUE FOR PUBLIC AND PRIVATE MATERNITIES Public Private I Revenue Source 1985 1986 1985 1986 Government 79.0 74.0 0 0 Grants 0 2.3 11.0 9.7 Loans 0 0 0.85 0.43 Cost recovery 18.4 19.8 79.0 82.0 Other 2.6 3.5 8.6 7.4 TOTAL 100 100 ;00 100 TOTAL NAIRA 47,702 52,223 45,204 42,283 (std. dev.) (39,478) (40,625) (34,955) (31,584) Source: Calculations based ori data in Ogun State Health Financing Study (ABCS). 5.13 In sunmmary, public maternities get three-quarters of their funding from the government and generate only about 20 percent of their revenues through user fees.67/ This pattern of revenue sources remained stable during 1985-86. Private maternities depend much more heavily on user fees for their revenues (82 percent) and receive no support from the government. They obtain the remainder of their revenues from grants (10 percent) and miscellaneous sources (7 percent). This pattern of revenue sources has also been stable over the two-year period. However, in some instances private providers have experienced a 12 to 15 percent drop in the proportion of revenues generated from user fees. The data suggest that the reduction has been compensated for, in part, by an increase in loans. 66/ Insufficient data were available to examine revenues for earlier years or for other facilities. 67/ Most facility administrators did not know whether government funds originated at the federal, state or local levels. The State Ministry of Health or the State Health Management Board is often responsible for channeling monies. - 64 - 5.14 Whtle public facilities rely very heavily on government funds for financial support, private facilities depend on user fees for most of their revenues and on grants and other contributions. Facilities in both sectors lack a diverse portfolio of revenues. Public facilities are at the mercy of government health allocations; private facilities depend on the continuing ability and willingness to pay of potential health care users. As Nigeria's current economic situation reduces government budgets and household incomes, both public and private facilities could experience financial hardship. 5.15 Gross Recurrent Profit in Public and Private Maternities. 68/ Overall profitability varies tremendously in maternities, both public and private. However, fo. the public sector, actual budget deficits may be hidden since the government finances any loss. On average, the profit percentages were: 1985 1986 (std. dev.) (std. dev.) Public maternity +2.26 +40.27 (40.11) (61.18) Private maternity +4.88 -13.50 (36.15) (40.16) A key finding in this profit analysis is that, in 1985, private maternities earned on average small profits (only 5 percent). In addition, private maternities were under increasing financial strain, incurring in 1986 average losses of 13.5 percent. In light of the economic situation, this result is not surprising. Devaluation of the naira increased the cost of many inputs. High unemployment and general recession reduced the income of many households. These constraints have intensified in recent years, suggesting that the financial strain on the private sector could be deepening. 5.16 Cost Recovery in Public and Private Facilities. 69/ Private facilities charge more than public facilities for almost all services.70/ Table 5.4 shows average patient fees for selected services at various facilities. It is notable that lower-level facilities, public and private, do not necessarily have lower fees. This may hurt the referral system because fee structures evidently do not discourage patients who want to bypass the primary care facilities.71/ 68/ This revenue analysis only pertains to public and private maternities because the revenue data were inadequate to reach general conclusions about other types of facility. 69/ The facility survey provides information on patient charges for selected services, revenues, and expenditures for drug revolving funds, and on revenues from user fees in general. 70/ The only exceptions are the clinic price for paracetamol and the hospital price for chest X-rays. 71/ Although vertical referral systems are the predominant organizational system in developing countries, the regular bypassing of the structure by patients results in inefficiencies and inequities - 65 - Table 5.4 AVERAGE a/ CHARGE FOR VARIOUS SERVICES RENDERED IN PUBLIC AND PRIVATE FACILITIES (Naira) Small Basic Health Hospital Maternity Clinic Public Private Public Private Public Private Setting broken arm 12.50 20.00 0.00 0.00 11.25 54.50 Malaria treatment 7.08 23.00 2.12 15.50 2.57 15.00 Dressings 1.75 3.88 3.13 30.25 3.10 5.70 Room and board 2.50 20.00 1.30 15.83 1.17 5.00 Appendectomy 23.33 187.50 0.00 50.00 32.50 0.00 Basic delivery 7.50 26.25 6.04 31.21 8.33 23.37 Chest X-ray 20.00 18.75 0.00 10.00 0.00 16.25 Urinalysis 1.50 17.50 0.00 5.00 1.00 12.50 Blood count 3.50 5.25 2.00 5.00 1.00 10.00 Paracetamol 1.00 9.00 0.37 2.97 0.58 0.07 Immunization 0.00 2.50 0.00 2.38 0.00 4.00 Source: Ogun State Health Financing Study (AHCS). a/ Average service charges for those facilities which have a non-zero fee for the service. 5.17 Drug revolving funds are a popular cost-recovery mechanism because many facilities make profits from drug sales. The financial status of these funds is shown in Table 5.5. Private facilities especially were (Mwabu, 1989). Reforms which could encourage the effectiveness of referral networks include clearly differentiated prices, with lower prices for primary care and for patients in secondary and tertiary level facilities who arrive through referral. - 66 - more likely to make money.72/ Drug availability offers a possible explanation. In 1986, private facilities supplied at least 92 percent of their patients with drugs compared with 73 co 87 percent for public facilities. While large public hospitals, maternities and dispensaries in general did not make profits from drug sales, the introduction in 1984 of a drug revolving fund policy in Ogun State seemed to boost net reserves. Table 5.5 AVERAGE DRUG PROFIT FOR HEALTH CARE FACILITIES 1985 1986 Z std. dev. 2 std. dev. Large hospital Public -10 (0) -28 (0) Small hospital Public -20 (76) 67 (132) Private 22 (11) 22 (14) Comprehensive health center Public 26 (0) 4 (0) Private 22 (25) 31 (11) Basic health center Public -25 (49) 125 (276) Private 268 (344) 116 (188) Maternities Public -15 (50) -4 (86) Private 27 (97) 10 (80) Dispensary Public -63 (45) -51 (48) Source: Calculations based on data in Ogun State Health Financing Study, AHCS. 5.18 The mixed success of cost-recovery schemes has implications for policies which promote such schemes for the public sector and recommend expanding private sector activities as they currently exist. Many private owners note that patients are delaying or avoiding treatment altogether due to inability to pay, and many private facilities are not reaching moderate 72/ This was true especially in 1985. These results are ambiguous since there was a large degree of variability. Drug revolving funds may have a strong profit earning potential but their success is not guaranteed. If private facilities are more successful at generatiLg profits from drug revolving funds, the reasons for their relative success should be researched. - 67 - utilization levels of 800 visits per health worker per year.731 Moreover, profitability of private maternities has been declining even to the point of incurring losses. Field observations indicate that although the private sector is growing, many facilities are beginning to experience financial strain due to the ongoing economic crisis. Thus the continued rapid expansion of the private sector is uncertain. Before designing new policies, decisionmakers should ascertain the most recent financial status of private owners. If losses persist, they will need to identify reasons for these losses, and to determine the short- and long-run effects of these losses on private sector growth. These issues are examined in depth in Part C. 5.19 Efficiency. Inefficiency may result from poor engineering (that is, improper input combinations or underuse of resources) or noneconomical use of inputs (such as overuse of expensive inputs). The statistical analysis of efficiency in Part A.2 of this chapter (page 69) examines both aspects. At this juncture, a general description of the potential inefficiency can be provided. However, the data are only suggestive because a crude proxy for output is used, namely, visits per health worker, which does not distinguish between the technical and economic causes of inefficiency. Nonetheless, although the data presented in Table 5.6 cannot pinpoint inefficiency sources at the facility level, they do suggest that several provider groups have efficiency problems. Strikingly, a majority of facilities fall below a moderate level of efficiency (800 visits per health worker per year). 5.20 Facilities that fall below the efficiency norm are usually very lacge and have large staffs whose utilization levels are relatively low. Of all facilities, public and private, only dispensaries are most likely to meet minimum acceptable efficiency levels. These results are surprising since private facilities are assumed to be more cost-conscious. With respect to inpatient services, occupancy rates for small public and private hospitals of 40 percent and 57 percent, respectively, also suggest substantial underutilization. 73/ This was the norm selected for a moderate level of acceptable efficiency, based on a qualitative assessment of observed facility activity levels. The statistical analyses used a minimum acceptable efficiency level of 600 visits per health worker per year. - f8 - Table 5.6 EFFICIENCY AS PROXIED BY VISITS PER HEALTH WORKER (HW) AND FACILITY CHARACTERISTICS Low Medium High 0-799 800-2000 2001 + Visits!HW Visits/HW Visits/HW Number of Facilities 27 14 5 Types of facilities Small hospital 2 1 0 Large hospital 1 0 0 Comprehensive center 0 1 0 Maternity 13 4 2 Basic health center 8 3 1 Dispensary 3 5 2 Public 19 12 3 Private 8 2 2 Average number of visits 6,399 15,101 8,918 (Maximum) (11,917) (25,513) (8,951) Average number of HW 21 11 2 (Maximum) (41) (15) (2) Source: Calculations based on data in Ogun State Health Financing Study (AHCS). 5.21 These conclusions must be interpreted with caution. First, descriptive statistics describe the average behavior of different provider types. Case study analyses of specific facilities may yield very different results. Unfortunately, the sample size was too small to conduct a detailed micro level analysis. Second, the quality of the data is often poor. Based on field experience, the data are likely to contain errors that are not readily apparent. Financial information systems in Nigeria are rudimentary, yet with modest modification, the current data collection system could provide the fundamental information needed by policymakers. (See Annex I.) However, they do provide the information for carrying out a basic comparative analysis of the operational status of public and private facilities, a task which has not before been possible.74/ 74t As Abel-Smith (1989) points out, "There is a danger in assuming, as the basis of policy, that private hospitals (almost by definition) are bound to be more efficient than public hospitals.a - 69 - 2. Statistical Analysis of Efficiency in Public and Private Facilities 5.22 This section will (a) assess the technical and economic efficiency of health care facilities in general; (b) compare the efficiency of public and private facilities;75/ and (c) analyze the cost structure of health care facilities with respect to marginal and average costs, the additional cost due to inefficiency, and economies of scale and scope. 5.23 Technical and economic efficiency are examined using a frontier production function to describe the maximum levels of health services possible with given combinations of personnel and nonpersonnel health services. Measures of the physical productivity of health and non-health workers are compared with their respective wages to determine whether cost- minimizing staffing patterns are in effect. Differences between public and private facilities are explored. The costs of inefficiency and the general cost structure of health care service provisions are determined by estimating a short-run cost function. From these results, marginal costs, average costs and economies of scope and scale are calculated. The model and the cost function estimates are detailed in Annex J. 5.24 The data set is drawn from a heterogeneous group of nonhospital facilities whose production and cost functions may differ dramatically.76/ Although these facilities offer a wide range of services, the estimations can only focus on two aggregate output measures: 75/ Since personnel cost is the largest component of total costs, the efficiency analysis focuses on staffing patterns, namely the mix of health and non-health workers. 76/ General information on facility location, ownership, type of facility, services offered, usable equipment and staffing patterns was collected for all 445 modern health care facilities in Ogun State. Another survey of 86 facilities was conducted to collect more detailed information on utilization (inpatient and outpatient), target populations, pe-sonnel and nonpersonnel costs for the period 1981-86, sources of revenue, selected service charges, measures of cost recovery and availability of drugs. These facilities were chosen through stratified random sampling. Of these 86 facilities, 68 were fully responsive. - 70 l inpatient admissions and outpatient visits.77/ Table 5.7 presents the composition of this small study group.78/ Table 5.7 CHARACTERISTICS OF PRIMARY AND SECONDARY FACILITIES: OWNERSHIP AND EFFICIENCY Operational Classification <600 visits/HW/yr >600 visits/HW/yr Total Comprehensive health center Public 0 1 1 Private 0 0 0 Basic health center Public 4 6 10 Private 1 1 2 Maternity Public 5 5 10 Private 7 2 9 Dispensary Public 1 9 10 Private 0 0 0 TOTAL 18 24 42 Public 10 21 31 Private 8 3 11 Source: Calculations based on data in Ogun State Health Financing Study (AHCS). 77/ This is the largest subgroup of facilities for which services can be reasonably lumped together into only two output measures. Comprehensive health centers, primary health care clinics, health clinics, maternities and dispensaries generally provide the same type of outpatient primary health care and the same type of inpatient care, consisting mostly of deliveries and minor surgery. The data set includes 61 observations for the nonhospital sector. After cleaning the data, 42 nonhospital observations remain. Several observations had to be dropped from the sample because of missing data. 78/ The group is small, especially for estimating the frontier production and cost functions which include only those facilities with more than the minimum acceptable level of 600 visits per health worker per year (24 observations). This means that there is the potential for omitted variable bias since the sample size is not large enough to fully incorporate all variables that affect production or expenditures. This selection process is discussed in more detail in Annex G. - 71 - Cost Analysis Results 5.25 Table 5.8 and Figures 5.1 and 5.2 reveal relevant features of primary and secondary health care facility service costs.79/ As expected, inpatient services are more costly to provide (by 100 percent) than outpatient visits in terms of marginal and average costs. Inpatient average costs are more than double outpatient average costs. As shown in Figures 5.1 and 5.2, marginal costs are less than average costs, indicating that facilities are generally operating on the downward portion of their average cost curves. This is especially true for admissions. Although also downward sloping, the marginal and average costs curves for visits are somewhat flatter. T:.ese results are consistent with the product-specific scale measures which indicate increasing returns to scale for admissions and nearly constant returns to scale for visits. Table 5.8 ESTIMATES OF COST ANALYSIS MEASURES a/ (Naira) Admissions Visits Marginal cost 6.85 4.82 Average incremental cost 18.05 3.73 aI Details of calculations are shown in Annex J. 79/ These results are averages; facilities show wide variability in results. Because only three private facilities met the production frontier criteria, the following comments pertain primarily to public facilities. MARGINAL COST (MC), NAIRA AVERAGE COST (AC), NAIRA 0 -- IC o~~~~~~~~Z 40 50 I. to~ ~~0 0 *070~090 001010 - 74 - 5.26 Because inpatient service facilities are apparently operating under capacity, high fixed costs are distributed over relatively few admissions. Low bed occupancy rates are consistent with these trends in average costs. A likely explanation for low utilization levels is lack of complementary inputs. Increased availability of drugs and other supplies would shift average and marginal cost curves upwards, but utilization would also increase such that the minimum portion of the average cost curve could be reached. 5.27 Since outpatient services appear to be less affected by undercapacity, one might consider the extent to which some of the fixed inputs such as personnel can be cransferred from inpatient to outpatient care. The fungibilS,. of various types of health workers should be fully explored. Allocative tfficiency analysis indicates that public facilities are not using cost-mixiimizing combinations of health workers and support staff; the public sector appears to employ too many support staff relative to health workers. A surprising result in the private sector was that most private providers, like the public sector, do not achieve the minimum acceptable technical efficiency of 600 visits per year per health worker. 5.28 Finally, the economies of scope measure shows that under existing conditions there are no sizable advantages to combining inpatient and outpatient services. By offering these services together, overall fJxpenditures remain clearly constant or increase slightly (10 percent). Combining an underutilized inpatient department with an active outpatient department might raise overall costs. 5.29 These results must be qualified, beceuse the data upon which the analyses were performed were often incomplete and suspect and the data set used for the statistical analysis was extremely small. An important indirect recommendation from this cost analysis is that financial information systems at the facility level must be vastly improved; financial data need to be coordinated with appropriate and adequate measures of health care quality, facility utilization, and health status. Without such systems, the ability to examine, recommend, implement, and evaluate strategies to improve resource allocation within the health sector is limited. (See Annex I for a summary of basic data requirements of a management information system.) B. Household Demand for Outpatient Health Services 5.30 People's need and desire for health services differ across geographic areas, as do their willingness and ability to pay. Demand analysis answers questions about the causes of demand for public sector health services relative to the demand for other types of providers (e.g., private, mission, traditional).80/ Health planners can use information 80/ Too many Lhings that affect health care usage are always changing in any population for the effect of any observed price changes to be directly identified correctly, even when they can be observed. Realizing this, social scientists and policymakers have looked to well-designed demand analyses to help predict how various changes, such as raising fees at public facilities. will affect health care usage by specific income groups. A more technical discussion of demand estimation is presented in Annez H, and the economic and statistical models and estimation procedure used for this analysis - 75 - on local health care preferences to design fee schedules and to target price changes, exemptions and subsidies so as to protect the poor from higher prices. Demand studies also allow policymakers to predict the likely demand for insurance (risk-sharing) coverage and the need for government involvement. This part of Chapter V uses household survey data to analyze the relationships in health care usage and its determinants, and to estimate actual usage patterns. 1. Surveys and Samples Selected 5.31 This study uses household survey data to analyze the relationships among health care usage and its determinants and to estimate actual usage patterns (see Form C at Annex K). One hundred patients were randomly selected from each facility. Interviews were carried out at their homes within the next two weeks. To obtain a more random sample of other households near the facilities, interviews were also conducted with neighbors of two of each five interviewed patient households. Data for 1,763 non-pregnant adults who had recently been ill are used in the demand analysis.81/ are presented in Annex N. 81/ Some 6,800 households were theoretically available from 68 facilities, but some of the rural clinics could not supply 100 names during the observation period and some addresses turned out to be nonexistent. In total, 3,380 household interviews of patient households were finally obtained. The neighbors sample added 1,360 Ononpatient" households to the total sample. Thus, a total of about 4,898 household interviews with 22,074 individuals were obtained. The demand analysis on the adult outpatient sample was not performed on individuals less than 16 years old and those who did not report an illness in the last 4 weeks, resulting in a sample of size 2,264. Also deleted were all individuals whose reported illness could not reasonably be considered an outpatient illness. The specific reported illnesses for those individuals included in the sample were: fever, cough, stomach trouble, diarrhea, weakness, headache, burns, bites and stings, poison and falls. The number of individuals reporting one of these illnesses was 1,807. Finally, deletion of observations due to missing values for independent variables resulted in a final analysis sample of 1,763 individuals. - 76 - 2. The Socioeconomic Situation in Ogun State 82/ 5.32 The samples analyzed were mostly urban (57 percent). Nearly half (46 percent) of the population was aged 10 or younger; only 9 percent was above age 40.83/ Around half (49 percent) of the sample population had no formal education, while less than one quarter (23 percent) had secondary education. More than half (56 percent) were too old or too young to work (they formed the dependent population, in other words). The composition of the employed sample was: 32 percent traders, 16 percent farmers, 14 percent professionals, 11 percent office workers, 9 percent teachers, 7 percent artisans, 4 percent applicants, and 7 percent in other categozies. Christians outnumbered Muslims (53 percent vs. 46 percent). Around half (52 percent) of household members were employed. Earnings were low. Some 79 percent of those working earned less than 1200 per month, while 14 percent earned N401 to N600, and 3 percent earned M601 or more.84/ 5.33 Other socioeconomic indicators reveal that sanitary and public health conditions in Ogun State are somewhat deficient. Only about one quarter (27 percent) of the sample had piped water in their home; 43 percent used water from unpiped sources such as ponds, springs, tankers and open rivers. Around 16 percent traveled over 500 feet from home for water. About 24 percent of the sample used flush toilets; 63 percent used pit latrines. Almost three quarters (73 percent) of households had nons of the "comfort/leisure" items listed in the survey: only 19 percent had radAos; 2 percent had televisions; 1 percent had stereo sets; and 5 percent had one or more other items (video, refrigerator, or other). One-third (33 percent) of the households owned some land. 5.34 Health Status.85/ As shown in Table 5.9, the sample was divided into five household income (earnings) classes. Somewhat surprisingly, self-reported health status classifications differ very 82/ This section of the report is largely based on information provided - by African Health Consultancy Services (AHCS), in Report III of the Ogun State Health Financing Study, dated March 1989. The AHCS report was based on a sample of 4,700 households, including 3381 "patients and 1359 "nonpatient" households. (These differ slightly but not significantly from those we obtained in evaluating the data set). Some of the numbers do not come from that report but are based on direct results from the data set. 83/ The average age was 19.9 years; the median age was 18 years. 84/ At the time of the survey, 14 = US$l; per capita income in 1987 As estimated by the World Bank as US$370. But using earnings alone to measure income or wealth levels in developing countries can be misleading. Another indication of wealth is housing conditions and ownership patterns. The sample shows that 372 of individuals lived in a single room, 34? in a room plus parlor, and 372 in other housing configurations. About half of the sample of respondents lived in housing owned by other household members. 85/ Most of the information reported in this section comes from direct data analysis, rather than the AHCS report. - 77 - little by income group. The most obvious differences occur in the poorest group. Although only 1.7 percent of the poorest quintile said they were in poor health, this figure was still almost double the percentage for any other income group. The poorest group also reported a much smaller percentage in excellent health (1.4 percent, as compared to the 4.3 percent sample average and 5.5 percent reported by the N=51-100 monthly earnings group). 5.35 Surprisingly, many people (29 percent) in the highest income group said they were in 'fair' health (the sample average is 18 percent; 20 percent for the lowest earnings group). Much as expected, 80 percent of the highest income group (but only 66 percent of the lowest) who report illness sought help at a private or public facility. The other three income groups are all near the mean value of 71 percent seeking health care. The highest income group also reported a higher fraction of its members in fair or poor health (30 percent) than any other income group (the sample average is 19 percent; 21.9 percent for the lowest earnings class). Table 6.9 HEALTH STATUS OF INCOME GROUPS Household Health Status Reported Earnings , (noirs/month) Excellent Good Fair Poor N 5 N X N % N % 0-26 38 (1.4) 1,998 (75.6) 532 (20.2) 46 (1.7) 26-60 176 (8.6) 8,776 (76.6) 901 (18.8) 37 (0.8) 51-100 349 (5.6) 4,860 (76.7) 972 (16.2) 68 (1.1) 101-200 209 (6.1) a,008 (76.1) 690 (17.0) 84 (0.8) 201+ 71 (4.0) 1,121 (63.7) 618 (29.1) 17 (1.0) TOTAL 841 (4.8) 14,801 (74.7) a,608 (18.2) 201 (1.0) Source: Calculations based on data In tho Ogun State Health Financing Study (AHCS). 5.36 The most important indicator of differences in health care use by income groups is the extent to which those who are ill seek outside care. Table 5.10 shows care by income groups. - 78 - Table 5.10 USE OF HEALTH CARE BY INCOME GROUP Use Public Monthly or Private No Income Facilities Consultatlon Total Lowest income group (10-26) Number 401 208 609 Percentage (66) (84) ' (14) Second income group (M26-60) Number 790 298 1,088 Percentage (78) (27) (24) Third income group (X11-100) Number 1,020 426 1,446 Percentage (71) (29) (32) Fourth income group (1101-200) Number 667 269 986 Percentage (71) (29) (21) Highest income group (N201 or higher) Number 885 6B 421 Percentage (80) (20) (9) TOTAL 8,218 1,286 4,499 !/ (71) (29) (100) These number. are based on the total sample of Ill people, Including those under ago 16 and those who were pregnant. For this reason the sample size is larger than that for the profit analysis sample, and the percentage choosing each option very slightly (for example 29% chose Ono coneultationO versus 81X in the choice-basd sample). 5.37 The great majority of the lowest income groups seek help from public health facilities while the highest income group makes much greater use of private facilities. Table 5.11 shows the first source of care by income group. The most notable differences are between the poorest and richest groups: Where people go for medical help, by income quintile (percent) Highest Lowest Public hospitals 27 21 Private hospitals 24 5 Public health clinics 22 46 Private health clinics a 7 - 79 - This does not mean that the more wealthy groups avoid public facilities -- 53 percent of the highest income group who used a facility sought help at a ,ublic clinic, a public maternity, or a public hospital. Table 6.11 SOURCE OF HEALTH CARE BY INCOME GROW Comprehensive Tradi- Public Privete Government Income tional Spiri- Covernment Private Health tenIth Health Public Private Group Pharmacy Healer tualist Matornity Matornity Clinic Clinic Center Hospital Hospital Total Lowest -Number is 12 2 19 8 162 24 1S 72 19 Nd849 -Percent 5s 3% 1% 6% 2X 481 70% 40% 21X 65 (18X) Second Lowest -Number 26 28 4 63 31 214 41 se 162 56 N=673 -Percent 41 4% 1X 8% 6% 82x 65 95 241 81 (241) Third Lowest -Number 36 27 6 88 46 194 63 58 277 98 WuSIS -Percont 45% 355 1 10% 5% 22X 7% 6% 31X 101 (821) Fourth Lowest -Number 28 20 3 s8 21 114 48 11 170 80 N1=73 -Percent 65 31s 1% 10% 41 205 8X 2% 31X 17X (213) Highest -Number 22 10 4 12 4 63 24 1 78 67 N=20C -Percent 851 4% 1% 4% 1X 22% 8ex O 27% 24X (101) TOTAL -Number 128 97 19 230 109 747 198 141 766 881 N=2766 / -Percent (6%) (4%) (1%) (8e5) (41) (27) (75) (5S) (281) (12%) 100.00 Source: Calculations basod on Ogun Stato Health Financing Study. a/ The difference between the total sample using health caro (including children and pregnant women) of 8,218 (Csble 6.10) and 2,766 is mado up of 448 observations having tncomploto dote, in that detaile of cource of care were not provided. 5.38 If a person uses a different facility from the one he first chose, where will he go? Only 11 percent (317 of the 2,765) who use a source listed in Table 5.11 for a first consultation use a second source other than a pharmacy. The patterns for the choice of a second consultation for the poorest and richest groups are strikingly similar to the one presented in para 5.45. Of those going to sources other than a pharmacy, almost half of the poor but only around one-quarter of the richest go to a public hospital. The groups go about equally to private hospitals (around 20 percent), and to traditional healers (around 15 percent) (see Annex L). - 80 - (a) Demand Analysis 5.39 The information presented above on health status or usage by income groups is informative, but not sufficient for policy purposes. For example, policymakers need to know why the highest income group exhibits high formal care usage. Do the poorest and the richest groups have similar illness patterns? Do the richest groups live closer to hospitals than the poor? Do the rich simply have a greater taste for formal care? To answer these questions, policymakers need a full model that considers all the elements that determine health usage (and thus expenditures). An appropriately estimated demand model will include (a) health care prices; (b) alternative care prices; (c) the ill person's income; (d) age; and (e) education level; (f) the seriousness of the illness; and (g) the quality of care available at each facility accessible to the ill person. It will separate out the above variables so that changes in the variable can be evaluated from a policy standpoint.86/ 5.40 Given the current situation of underutilization of facilities, usage is an issue of great concern to Nigerian policymakers. For the actual demand estimates, the choice-based methodology was used to weight the choice-based (chose to use a facility) and community-based (neighborhood) samples to create a combined representative sample of potential users with access to a facility.87/ This procedure utilized the extensive information on sick people obtained from the facility- related, choice-based survey, but allowed the total sample to be weighted in a manner that makes the results representative of the population of potential facility users. Policymakers who are concerned about current and probable usage of existing facilities can use these conclusions for guidance in facility use and pricing decisions.88/ 5.41 The dependent variable consisted of individual's choices among three methods of treatment: (a) no consultation; (b) public facility; 86/ See footnote 80/. 87/ The original respondents were selected because a household member had obtained care at a sampled facility. Interviewers selected choice- based households by meeting patients at health facilities and making a follow-up interview visit. The selection of facilities and patients was random. Making an appointment for a follow-up visit allowed people to prepare better for the time and effort involved in providing detailed answers to the lengthy questionnaire, guaranteeing a large number of completed interviews for health service users. Neighbors were also interviewed. In taking illness histories, respondents were asked "What did you do first--, second--, third--, so that they had at least three opportunities to report using various facilities. No evidence suggests that sampling households near a facility (who may or may not have used that facility) does not provide a random sample of possible users. 88/ However, the conclusions cannot be extended with full confidence to health care usage in general by the total population, including those who live in areas relatively inaccessible to the facilities. - 81 - (c) private facility.891 The frequencies for the dependent variable in the analysis sample were as follows: Sample: Choice-based Community-based No consultation/self treatment .3153 .5155 Public facility .4943 .3422 Private facility .1904 .1423 5.42 The empirical model employed assumes that choice of health care is a function of: (a) price of the care (b) --uality of the care (c) sex of the patient (d) education of the patient (e) wealth (assets) of the household (f) income of the household (lowest 20 percent) (g) urban residence (h) symptoms of the illness (i) seriousness of the illness Price and quality depend on the health care chosen (conditional variables); other variables are specific to the household (or individual patient) and do not vary across health care choices (unconditional variables). A discussion of the actual specification of these variables is in Annex 0. (b) Estimation Results 5.43 Table 5.12 presents results for the estimations. The results strongly support the views that (a) price affects health care choice but has an impact that is small in magnitude; (b) quality as measured has a large impact on health care choices; and (c) higher prices can be offset by higher quality. 891 Individuals were categorized depending on their answers to a series of questions. The first question was "Have you been ill in the past four weeks?" Those whe k not sick were not asked follow-up questions. Those who were sick in the last four weeks were asked what they did first, second, and third to treat the illness. The data set contains answers to the "do first" and 'do second" questions. The dependent variable was defined as "no consultationo only if they answered self-treatment to both questions. The dependent variable was defined as "public facility' if they went to a public facility first or if they went to a public facility second after responding self-treatment to what they did first. The dependent variable was defined as "private facilityF in a similar manner. If responder.c answers indicated a combination of public and private facility use, what they did first was used to define the dependent variable. About 962 of the time this algorithm for defining the dependent variable resulted in using the answer to the "do first" question. Only 731 of the 3,186 who pursued a first source (which includes self-medication) answered that they pursued a second course of action for treatment of the illness other than a pharmacist. - 82 - Table 5.12 ESTKWATION RESULTS OF USAGE Public Private Public versus versus versus No Consultation No Consultation Private Variable Coefficient "t' Coefficient ate Coefficient ate Conditional Price -.029 -1.68 -.029 -1.68 -.029 -1.68 Quality: (a) Expenditure on care per person in population served .029 1.56 .029 1.56 .029 1.56 (b) Drug .vailability (percentage of time available) .015 3.95 .015 3.95 .015 3.95 (c) Physical condition of facility index 1 good 2 - fair -.2;7 -4.54 -.227 -4.54 -.227 -4.54 3 = poor Unconditional Education of ill person .022 0.83 .022 1.97 -.001 -0.03 Sex of ill person (male - 1; female = 2) -.001 -0.01 .064 1.70 -.064 -0.91 Own a vehicle .197 2.14 .112 2.67 .084 1.09 Lowest income quintile .163 1.47 -.004 -0.08 .166 1.76 UTrban residence .010 0.09 .183 2.90 -.174 -1.65 Days of reduced activity .028 3.33 .013 3.54 .0'5 1.94 Stomach or diarrhea symptoms .557 4.27 .243 4.34 .314 2.59 Constant -1.679 -4.17 -1.579 -4.38 -1.00 -0.51 Source: Hultinomial probit estimation results, based on data in Ogun State Health Financing Study (ABCS). 5.44 Raising prices with constant quality will reduce usage; but spending the increased funds to improve quality may even increase usage. Given equal prices and identical tastes, people prefer health facilities that have (a) higher per capita spending on care; (b) higher drug availability; and (c) a relatively good physical condition.90/ Efforts 90/ The price of the specific type of care is statistically significant at the 10? level (t=-1.68) and has the expected negative sign. As the price of a care option increases, if the quality of care statistically is held constant, the likelihood of that option being - 83 - to include a price of time in the model are detailed in Annex 0. The price of time variables were not found to be statistically significant. The strength of the results on the money price and quality variables is reassuring. In all estimates and with various combinations of variables other than quality and price included, the price and quality variables almost always had the expected signs and were highly significant statistically. In addition, the results indicate that the lowest income quintile is more likely to choose public care than other income groups but behaves much like the other income groups in choosing between private care and no use of formal care (self treatment).911 chosen will decrease. Thus, higher prices will reduce usage, as expected. But because the profit relationships are inherently nonlinear, the magnitude of the effect can only be simulated by estimating the predicted usage of each type of care for different values of price. Two of the quality of care variables are highly significant; the third is significant at a low level (expenditures per capita for the population served only has a t value of 1.56). All have the expected signs. The multinomial profit model effectively estimates parameters of the utility function for each outcome. The parameters on the conditional variables are the same for each outcome and those on the non-conditional variables differ. Because price and facility characteristics (total expenditure per person served and facility physical condition) obviously are not relevant variables for the 'no consultation' choice, we use zero for the value of each of these variables when 'no consultation' is chosen. This in effect causes these factors to be omitted ("zeroed out") in the utility equation for the "no consultation, choice (i.e., the amount of utility one receives from "self treatment' is not affected by the physical characteristics of facilities). This is the r1 effective way to handle the impacts of such nonexictent factors as the physical condition of the facility which provides the 'no consultation' choice. 911 When fees are introduced, even if health service usage is not greatly reduced, financial burdens on the poor will be increased. Practical methods to identify the poor rs to exempt them from fees or to provide them subsidies do exist. Unless such means tests are utilized and the poor effective±y protected, the social cost could be large. Possible approaches to identification of the poor used in developing countries include certification by citizen committees or local officials. Over time, as tax records and other information become more reliable, more precise methods of identification of the poor should become available. Another possibility which would not necessarily require such good information is a self-declaration system with specific income and wealth standards. Subsidization could include charging small amounts or nothing for 'no-frills' levels of room and board in hospitals but full cost for less crowded rooms, better food, etc.; or subsidizing care for specific population groups (for example, infants and pregnant women) or regions of the country. - 84 - 5.45 Tastes and preferences of individual clients are surprisingly strong statistically in their effect on choices. The wealthy (those who own a vehicle) seek the most medical care. Urban residents, females, and the more highly educated prefer private medical care. Those with (a) stomach or diarrhea problems, and (b) more severe ailments, tend to seek treatment at public facilities, assuming equal prices and quality. 5.46 These results suggest that if public facilities offered private sector levels of both drug availability and physical conditions (assuming constant public and private sector spending per patient), they could raise outpatient prices to the level of the private sector (an 87 percent increase) and still increase usage. Some current nonusers would be drawn to the public sector, and around half of present private care users would also switch. This interpretation of the quality/price trade-off obviously oversimplifies the problems that would be encountered in an attempt to actually make such changes. Yet the results strongly suggest that if revenues from higher prices at public facilities were used to increase care quality, usage of these facilities (and of modern formal care in general) might well increase. 5.47 While this part of the analysis has explored how change affects an individual, it does not provide the total picture. By combining household responses and facility costs to reach estimates of aggregate costs ani aggregate revenue, a comparison of alternative policy formulations can be made. These combinations of options for a program of cost recovery in the health sector are explored in depth below. C. Health Care Cost-Recovery Simulations 5.48 "Cost recovery" nas been applied to many sorts of public sector activities, from electricity and water supply to health care. At the most basic level, cost recovery represents a government effort to have the beneficiaries of a government good or service contribute toward its osts by taxation or by imposing prices for the good or service. 5.49 One of the main conclusions of the analysis in this report is that raising prices for public health care services could increase the number and frequency of users and also total revenue. That conclusion is shared and supported by the results outlined in this chapter. This apparently paradoxical conclusion (cost savings from higher prices) will be true only if price rises are eventually accompanied by higher quality service. Better service means that health care facilities must have more medical supplies, more drugs, and a well-trained and committed workforce. Another important conclusion is that management capability needs to be significantly strengthened before an extensive cost-recovery program is attempted. 5.50 Status of Cost Recovery. The Nigerian National Health Policy, laurched in October 1988, states: L.s a general policy, users shall pay for curative services, but preventive services shall be subsidized. Generally, public assistance shall be provided to the socially and economically disadvantaged segments of the population.92/ 92/ National Health Policy and Strategy to Achieve Health for All Nigerians, Federal Ministry of Health, October 1988, p. 49. - 85 - Despite these aims, cost recovery from users of curative government health care facilities is not yet widespread or systematic in Nigeria and is similar to levels observed in other African countries. The extent of cost recovery in the public sector in Ogun State is not widespread and is sporadic rather than systemic.93/ 5.51 A low level of cost recovery is a .:idespread phenomenon in Africa, where cost-recovery ratios (see Table 5.13) are typically below 8.0 percent.94/ In grappling with cost recovery policy, most African countries (including Nigeria) must take into account: (a) a multiplicity of providers and options for the consumer; (b) varying pricing practices; (c) some cross-subsidization between government and the church-mission sector; and (d) programs to provide free curative hea4th care for the indigent. Table 5.13 AFRICA: GOVERNMENT REVENUE FROM USER CHARGES AS A PERCENT OF RECURRENT GOVERNMENT EXPENDITURES ON HEALTH SERVICES Country Percent of Recurrent Expenditure Botswana, 1979 1.3 Burkina Faso, 1981 0.5 Burundi, 1982 4.0 Cote d'Ivoire, 1986 3.1 Ethiopia, 1982 12.0 Ghana, 1987 12.1 Kenya, 1984 2.0 Lesotho, 1984 5.7 Malawi, 1983 3.3 Mali, 1986 2.7 Mauritania, 1986 12.0 Mozambique, 1985 8.0 Rwanda, 1984 7.0 Senegal, 1986 4.7 Swaziland, 1984 2.1 Zimbabwe, 1986 2.2 Source: Vogel (1988) and (1989). 931 See Tables 5.3, 5.4, 5.5. 941 They range from 0.5Z for Burkina Faso in 1.31 to around 12X for Ghana, Ethiopia and Mauritania in 1987. - 86 - 5.52 First Principles for a Cost-Recovery Program Schedule. In setting a system of prices in the health care sector, a government should move carefully, according to a set of pricing principles that flow from its efficiency and equity objectives. One possible set of principles is contained in Annex R.951 These principles would establish a fee structure which encourages an effective referral system use of preventive health care and protection of the poor, with investment of revenue, and a monitoring and evaluation mechanism. 5.53 For example, fees cotld increase with the level of care, but outpatients referred to higher levels in the system would be charged a lower fee at the higher care levels. Nominal registration fees at the lowest level of care would deter urnecessary utilization, but be low enough for the majority of people to afford. In order to promote the best health outcomes, the fee structure could allow free follow-up visits and visits by children under five years of age who had an immunization card, with token fees being charged for preventive and family-planning visits. Amenity wards could bear higher prices and, eventually, possibly even prices sufficient to cover all of the costs of such wards. 5.54 Explicit procedures should be introduced to insure that the poor have access to care. For example, services could be without charge in poor areas (according to a definition of "poor" that would be considered fair). Registration fees could be set at very low levels at the dispensaries/health posts and at health ;enters. Free follow-ups and lower priced referrals to higher levels of care could ensure access to outpatient care.96/ Existing forn-al procedures to waive fees for the poor could be continued at each facility not classified as being in a poor area, or a system of reimbursement for the poor could be used because it would be administratively more transparent for facility staff. The system should be regularly audited to minimize fraud. At eacb. facility, a centralized 95/ Along these same lines, Annex S contains e short discussion of the apparently successful cost-recovery experience in Ghana, and also contains the Ghanaian price schedules (for 1983 and 1985) together with the Ghanaian "Modalities for Collecting New Hospital Fees., These documents are included so that the interested reader can see part of the evolution of the cost-recovery pricing schedule in Ghana. Of particular interest is the increase in prices between 1983 and 1985 which eventually made it possible for the cost-recovery ratio to reach 12.1 percent by 1987. It should also be noted that there has been some discussion about the kinds of utilization declines which took place in Ghana as the result of the higher prices in 1985 (see Waddington and Enyimayou, 1989). 96/ Given that the objective is to get people to go to primary level facilities first and then progress up, in line with their own preferences with respect to time and money, the price signals have to be very clear. This means that primary care must be significantly cheaper. However, if primary care is perceived as Obetter" at the tertiary level facility, the price differential may be insufficient to discourage overuse of the tertJary level facility. This can be overcome by opening a separate primary care department at the tertiary level facility, with the same differential pricing system with respect to use of primary care and referral service. - 87 - admissions location would allow the fee and monitoring information to be collected and a receipt issued. The patient would then see a health care provider. If a follow-up visit or referral were necessary, the provider would indicate this information on the receipt. 5.55 A significant portion (50 to 75 percent) of the fees collected at each facility could be made available to improve the quality of care at the facility, without any commensurate reduction in the government allocation that it received. Likewise, the MOH's funding from the Ministry of Finance should not be reduced as cost-recovery revenues began to accrue. The remaining 50 to 25 percent of fee revenues could be divided among health authorities to enable increased spending on preventive services and ir order to improve quality at facilities that might not initially introduce fees because they were in poor areas. In order to prioritize the expenditure of revenues generated from user fees, a system of guidelines could be introduced to enable local officials to determine quality- enhancing priorities at the facility level. A similar system could also be ased to determine priorities for expenditures on prevention and for subsidies to poorer areas. 5.56 Based upon the initial experience as indicated by the monitoring system, the fee structure and administrative system could be adjusted in the future. Expected changes could include (a) raising many of the fees, (b) lowering fees at some levels of care, and in some geographical areas, in order to protect the poor, and (c) adjusting relative fee levels, across levels of care, in order to improve incentives for the appropriate use of facilities. A community-based waiver system for the poor could be introduced, possibly administered under a program that would also apply to other government services, such as education and water. The possibility of extending the waiver system to NGO facilities, with the Government paying for the care of the poor at these facilities, could be explored. Specific service fees (as in Ghana), e.g., for surgery, laboratory work, or X-rays, might be introduced at the secondary level. Amenity ward charges could be raised to more than cover costs. 5.57 The principles outlined in Annex R have been used as an example of how a consistent system of prices could be designed in order to achieve efficiency and equity objectives in the health care sector. They represent one model towards which the Government could move. At present, cost recovery in the public sector in Ogun State exists in a rudimentary state (see Table 5.3 in Section A of this chapter). However, by simulating various cost-recovery strategies within the context of the pricing system now in use, in both the public and private sectors, some insight i..to consumer response to price and quality changes can be developed, together with the cost consequences of various quality changes. From a policy perspective these insights become invaluable fir gradual reform of the present system. Using the cost and demand information provided by the research done for Parts A ana B of this chapter, we now turn to these cost- recovery simulations. - 88 - 5.58 Simulations. A number of simulations were carried out by *:hanging the valued of important independent variables to determine their hypothetical effect upon health care choices.97/ Because the prices used are registration fees -- the price of entry -- rather than the total price of service, and the cost figures are based on a small sample of facilities, the costs and revenue estimated are appropriate for identifying the relative magnitude of the effects of various policy changes, but not for identifying the actual total costs and revenues involved. 98/ 5.59 Simulations were carried out for le alternative scenarios. These are shawn in detail as Cases 2 to 19 in tho Appendix to this chap:;er (see page 97). To indicate the scope of these simulations, Table 5.14 presents one part of the results. It is drawn from Table 5.16 (see page 98) ih the Appendix. To read the example simulations in Table 5.14, note that the first row (Case 1) shows that an individual with average values for all variables (including illness symptoms), and facing mean prices and quality for each facility choice, will choose self treatment about 52 percent of the time, will seek public care 35 percent of the time, and private care 13 percent of the time. In Case 3, the cost of visiting a public facility is doubled, and so is the cost of a private sector visit. In this case, the probability of self treatment increases to 55?, and the likelihood of seeking private treatment falls from 132 to 10?. If drug availability falls to 50? in both the public and private sectors, visits to private sector facilities would be negligible and the probability of self treatment would rise from a mean of 52? to 69Z. In another simulation, Case 16, the public sector facility was in "poor" condition, while the private sector remained at "good". The estimation results showed a big decline in public sector visits, from 35? down to 25Z, and the probability of visiting the private sector increased for 13? to 202. In the full set of simulations, Cases 2 to 19, various changes are made in the key variables at the public sector facilities, namely the price charged and the quality of the facility (i.e dr-ug availability, investment and general condition) with various reactions from the private sector. In each Case, = the simulated outcome shows the likely number of visits that would result from each combination, to each type of facility, and from this the relative change in the public sector gross revenue can be estimated. The results are discussed in detail in the Appendix to this Chapter (see page 97). 97/ Simulations are needed due to the difficulty of directly interpreting coefficients in multinomial probit estimation results. 981 For the simulations, we use only the community-based sample, omitting those individuals who were chosen because a household member was identified at a facility. We believe this to be a random sample of households with access to facilities and the choices predicted for this group to be reasonable estimates for the total population of those with access to facilities. For the reasons explained above, however, we utilized the choice-based observations in the weighted sample for estimating the parameters used in the simulation exercises. - 89 - TAKLE S. 14 SAMPLE OF SIIILATIONS FOR THE PUBLIC SECTOR U/ PeobabI I Itlia Chageo In Privato Public Pri vte Policy Public Sector Sector Self Sector Sector Options Variable Variable Reaction TreateMnt Visits Visit Ceos" - - I.2 . .13 Coe a Price of Public Facility Vielt Double Double .8 .90 .10 (Set to Nt.40 (Set to N4.44 Inatead of N1.20) Instead of N2.20) Caeo 4 Price of Public lNelity Vielt Double No change . 94 . I (Set to N2.40 (Reosine et instoed of N'.20 mean) Cus* 10 Public Facility Drug Set at g0 Alo eest at .69 .81 .00 Aveilabillity S Cses 10 Cendition of Public Facility Sot et por, No change .54 .28 .20 (RomaIne ins a / Full det.ile are In Table 8.16, page 98. 5.60 Gross Revenue Effects Broadly speaking the results (as shown in the Appendix at Table 5.16, page 98) highlight two issues. The first is that the public sector increases revenue by raising its price with only a small change in number of visits, so long as the relative price of public sector health care is less than the private. For example, if the public sector raised its price but remained below the private sector price there was only a slight decline in demand. However, if the public sector maintained its mean price and the private sector doubled its price, public sector visits increased only marginally; self-treatment became more popular. The change in gross revenue is relatively greater in the former case. 5.61 Second, increases in investment (expenditure per person in the population served), in drug availability, and in quality (condition of the facilities) sharply increased public sector visits, with the magnitude affected by private sector reaction to only a limited extent. However, while the revenue effects on the public sector can be significant, such improvements are also costly. 5.62 Gross Cost Effects Table 5.17, in the Appendix of this Chapter (see page 102) shows the gross cost estimates for the quality improvements. For example, while Cases 3 and 4 yielded a relatively large surplus for the public sector of N3.6 million, and Case 16 a more modest surplus of N2.2 million, Case 10 showed a loss of N1.1 million. The general picture shows that price increases have a clearly positive effect on the surplus or loss calculations, but each quality improvement requires significant investment which in some cases may swamp the revenue effect of the additional visits that it generates. These calculations highlight the need for careful selection of improvements. - 90 - a. Choosing among the 19 Policy Options 99/ 5.63 Table 5.15 summarizes the revenue and cost policy options, and these are plotted against visit frequency in Figare 5.3 100/. In these simulations, it is emphasized that what is illustrated is relative magnitudes not actual revenues and costs. 5.64 Figure 5.3 helps to answer the question: Which is the "best' of the 19 cases for public policy to pursue? However, one must first define "best," i.e., what government wishes to maximize at a given level of cost. One possible approach to the maximizaticn problem would be to try to maximize the frequency of visits to modern health care facilities, on the grounds that it is presumably more technically efficient (there is a greattr probability of having a cure) if a person has formal health care, rather than self-care, when that person experiences an episode of illness. Therefore, two possible maximizing solutions would be to pursue the policies underlying either Case 18 or Case 19 (in the northwest quadrant of Figure 5.3), where the sum total of public and private visit-frequencies is the highest, at 0.53. However, Case 18 yields a net loss of N4.4 million, while the net loss in Case 19 is N4.8 million. 99/ These estimates of total gross revenues and total gross costs, and the resulting surplus or losses, are only as realistic as the assumptions upon which they are based. On the gross revenue side, N, the total number of episodes of illness in Ogun State, had to be estimated for 1986. This estimate is the key to Table 5.16, because the frequencies are applied to it, and, as a consequence, the number of public visits and the amount of public revenues for the various cases are derived from it secondarily. Likewise, on the cost side, many of the assumptions about costs, e.g., the value of what we have called the "volume" marginal cost, came from econometric work that had to make do with a sample of only 24 facilities, out of a total of 445. Nevertheless, the assumptions were based upon the combined previous health economics and field work of all of the authors. 100/ In Figure 5.3, the dots show the frequencies of public visits and surplus or loss, and the number to the right of the dots shows the Cases 1 through 19; the boxes show the sum of public and private frequencies (total formal health care frequencies) and public sector surplus or loss for the Cases 1 through 19. For example, in the northeast quadrant, Case 4 is shown with a dot at a public visit frequency of 0.34 and at a public sector surplus of M3.4 million; Case 4 is also shown with a box, higher up on the visit frequency scale, at a public plus private sector visit frequency of 0.47, and at a public sector surplus of 13.4 million. For reference, Case 1 is - ,- _ ._ - 91 - Table 5.15 SUMMARY OF REVENUE AND COST POLICY OPTIONS AND VISIT FREQUENCIES -----------------------------------------------------------__----- Surplus Public Public Policy or Loss Sector Plus Private Options (Index)a/ Frequencies Sector Freq'encies ----------------------------------------------------------__------ Case 1 n.a. .35 .48 2 - .35 .46 3 ++ .35 .45 4 ++ .34 .47 5 0 .36 .45 6 -- .37 .50 7 -- .37 .48 8 0 .34 .49 9 - .37 .47 10 - .31 .31 11 - .44 .51 12 - .44 .51 13 0 .35 .48 14 + .30 .31 15 0 .36 .49 16 + .25 .45 17 0 .38 .39 18 - .47 .53 19 - .46 .52 a/ Index for surplus!loss is as follows: + or - means within one standard deviation of average surplus or loss, respectively; ++ or -- is within two standard deviations of average surplus or loss; 0 is approximately breakeven. - 92 - FIGURE 5.3 Frequencies ot Vts$ and Surplus or Loss: The 19 Cases frequency of VW§s .54 018 019 12 .52 Be ~~~~1 1 50 06 0 ISo 8 1 .48 07 Oe8 o9 13 04 190- 02 .46 12 5 160 03 0 .4.4 11.d .42 .40 170 170 .38 7 - 09 5 * 150 a .36 6 02 131 . 03 8 .34 04 -.32 0 0 014 10 ' .30 014 * A,oe S?o Ptqnc .26 A f 016 -6 5 .4 .3 .2 -1 0 1 2 3 4 5 6 LOw (Million Naira) StpM Pw4700 - 93 - 5.65 If large losses, or indeed any losses, are judged impossible to sustain, then trade-offs have to be made between visit-frequencies and cost-recovery surplus or break-even. Case 4 maximizes the sum of public and private sector visit-frequencies, if the MOH desires a net surplus that could, for example, be spent on the poor or spent on pharmaceuticals in the next time period. Cases 3 and 16 also result in relatively high combined visit-frequencies, with a net surplus, but Case 16 only has a 0.25 public sector visit-frequency, which might not be acceptable to the MOH, and Case 16 results in a lower net surplus than either Case 3 or Case 4. b. The Administrative Costs of a Cost Recovery Program 5.66 The cost calculations in Table 5.17 do not contain administrative costs, for the simple reason that not much is yet known about the additional costs of administering an expatnded cost-recovery system.101/ On the one hand, it could be argued that, because there is already so much non-health worker slack in the health care system (see Part A of this chapter) the non-health workers could be more efficiently used by retraining them to become cost-recovery agents ir. the health care facilities. Apart from marginal training costs, this labor-shifting aspect of cost-recovery administration would cause little additional expense to the MOH because the salaries of the non-health workers would be paid, whether they were idle part of the time or whether they were to become more fully employed as cost-recovery agents. On the other hand, if, because of institutional labor rigidities, the MOH has to hire additional personnel for the administration of expanded cost recovery, then administration of the expanded program could become relatively expensive, particularly if higher priced personnel, such as accountants, had to be hired in large numbers. 5.67 In the absence of any firm data on the subject, Annex U attempts to simulate two sets of administrative costs for the 19 cases on the basis of two sets of assumptions, one making administrative costs solely a function of total gross revenue generated, and the other making administrative costs a function of the number of new personnel needed per change in 10,000 visits. c. Cost Analysis Results 5.68 The preceding theoretical and empirical work indicates that more intensified cost recovery in the public health care sector in Ogun State is both a desirable and an achievable objective that the government could pursue. However, the results show that the two "best' policy scenarios do not even include the changes in quality that both the theoretical analysis and anecdotal information deem to be so important for a successful cost- recovery policy. Within the framework of analysis in Part C of this paper, the quality changes appear too expensive to justify, relative to the changes in demand that they engender. However, this type of economic analysis has the inherent limitation that it is "comparative-static,' in the sense that its methodology forces it to move from state-of-the-world to state-of-the-world, without being able to take into account the inherent dynamics of the real world situation. Thus, Case 3 or Case 4 would yield a net surplus in this time period (N3.04 million). In the next time period 101/ Although Vogel (X988) presents an intensive study of cost recovery in four sub-Saharan countries, data on administrative costs were not available at the time of the study, so that the empirical administrative-cost aspect of cost recovery could not be researched. The topic deserves much more attention and further empirical study in - 94 - part of this surplus could be used to pay for more and better care for the poor and/or to increase drug availabilit.y at public sector 13cilities, and/or improve the physical condition of public facilities.102/ 5.69 In a third time period, we would expect the demand for public health care to shiTT outward even further, as a result of the quality changes (see the change in probabilities in Cases 11 and 15 in Table 5.16). Using the results of the comparative-static analysis within a dynamic-time context, they y4eld final empirical results that are consistent with both the theoretical economic framework and with intuition. 5.70 Finally, good administrative and managerial practices are essentials for succeasful cost recovery and, as such, cannot be neglected or ignored.103/ Before Ogun State embarks upon more extensive cost recovery than in the past, it must make ample provision for training the managerial cadre, the quality of which will almost surely determine the success or failure of any new intensified cost-recovery effort. Moreover, and yerhaps more importantly, success or failure of cost-recovery efforts cannot be measured solely in terms of the amounts of net revenue generated or lost. Equity for the poor is another extremely important objective of cost recovery in Nigeria.104/ There are two administrative aspects in achieving equity. The first is the establishment of procedures for defining and identifying those persons who are too poor to pay any fees when they seek treatment at public facilities; the administrative tasks involved in this equity aspect of cost recovery have not proven easy (Vogel, 1988). The second is creating and implementing allocation mechanisms whereby a trade-off is made between vhe portions of the net cost-recovery revenue that will be left at the collecting facilities in order to give them strong incentives to continue to collect fees, and the portions that will go to the MOH to be distributed to poorer lower level facilities, or to areas of the country (or state) where the poor live and seek health care. Much more analytical attention and empirical research needs to be given to these two administrative aspects of cost recovery. 102/ Case 3 or Case 4 would yield a net surplus of K3.039 million (here, even after having subtracted out the most expensive administrative- cost scenario) in this time period. In the next time period, part of this surplus could be used to pay for more and better care for the poor, and/or to move the drug availability in public sector facilities from 0.72 of the time to 0.9958 of the time, which would only cost about N1.770 million (see Case 19 in Table 5.17). This would leave a further amount of N1.269 million (3.039 minus 1.770) that could be used for another quality-enhancing policy action in the same subsequent time period, such as changing the physical condition of public facilities from the mean 1.2867 ("good" to "fair') to 1.0 (*good"), which would only cost about an additional NO.384 million (see Casee 15 and 18 in Table 5.17). 103/ Vogel (1988) and Vogel and Stephens (1989). Both of these studies indicate that major difficulties in health care facility cost recovery and the failure of most of the rerolving drug funds (RDF) in donor projects in Africa can be primarily attributed to weak administration and managerial neglect. 1041 In October 1988, the Nigerian National Health Policy was launched; it states that, 'As a general policy, users shall pay for curative services, but preventive services shall be subsidized. Generally, public assistance shall be provided to the socially and economically ..^I --- - &1^ --..10a4.4e, a - 95 - Conclusion 5.71 In undertaking this work, it has to be recognized that the analysis is not based on definitive data but on relative data, to indicate magnitudes and directions. For example, the data is not of total costs to the patient (being based on registration fees) but relative costs, and limited proxies are used for quality measures. The analysis yields striking results. 5.72 Facility Analysis It was found that for botb inpatient and outpatient services facilities are operating with marginal cost less than average cost, especially for outpatient services highlighting the opportunities for increasing returns to scale for admissions and nearly conctant returns to scale for visits. The fact that even private sector providers were not achieving the minimum acceptable technical efficiency of 600 visits per health worker was surprising. 5.73 The allocative analysis showed that public sector facilities are not using cost-minimizing combinations of health workers and support staff; the relative productivity of aon-health workers is less than their relative wage, suggesting health care providers should re-examine their use of such staff to ensure that they are fully and optimally employed. 5.74 Household Demand The majority of the lowest income group seek health from public facilities while the highest income group makes more use of private facilities. But the more wealthy groups do not avoid public facilities: over half of the highest income groups who used a facility used a public facility. As expected, of all income groups, self treatment is greatest in the lowest income groups. 5.75 Although price is significant in selecting health care, it is of small magnitude. Of more significance is the perceived quality of the facility. Measures of quality used were lir4'ted, but provide good indicators of basic issues: drug availability, physical condition of the facility, and investment on provision of care. The first two were highly significant, and the third significant at a low level. It appears that higher prices can be offset by higher quality: if revenues from higher prices at public facilities are used to increase care quality, especially with respect to availability of drugs and quality of the facilities, usage of these facilities (and of modern health care in general) might well increase. 5.76 Cost Recovery Simulations Using the cost estimates developed in Part A of the chapter, and the probabilities of selecting each type of care responding to different variables as estimated in Part B, simulations of policy change were made. These results show the estimated relative change in magnitude and direction, rather than the actual net 18ss or surplus. The results are illuminating for policymakers, showing that if public sector prices are raised, the probability of seeking public sector care changes only marginally depending on the response of the private sector, but the revenue impact is large. This increase in revenue, if invested in the facilities, could then result in an increase in public sector visits. However, price increases which are not subsequently or concurrently rewarded with quality improvements, are likely to result in a loss of demand that could overwhelm the revenue expectations. - 96 - 5.77 These results highlight the major role that perceived facility quality plays in the decisions which patients make. But qualitl change can have very large cost implications, which could swamp revenue effectc. Therefore the specific quality improvements need to be carefully selected, choosing those which are most cost-efficient, providing the greatest demand response for the least amount of investment. Making this selection efficiently will require adequate information systems at the facility level. - 97 - APPENDIX COST RECOVERY: SIMULATIONS OF GROSS RMNUE AND GROSS COST EFFECTS 1. Simulations A number of simulations were carried out by changing the values of important independent variables to determine their hypothetical effect upon health care choiceslO5/. Becaus, the prices used are registration fees -- the price of entry -- rather than the total price of service, and the cost figures are based on a small sample of facilities, the costs and revenues estimated are appropriate for identifying the relative magnitude of the effects of various policy changes, but not for identifying the actual total costs and revenues involved.106/ 2. The results shown in the first row of celumns 4,5 and 6 of Table 5.16 (denoted as Case 1) show that an individual with sample mean values for all variables (including illness 3ymptoms), and facing mean prices and quality for each facility choice, will choose self treatment about 52 percent of the time, will seek public care 35 percent of the time, and private care 13 percent of the time. Cases 2 through 19 in Table 5.14 represent various possible realistic policy scenarios, or 'cases" that government could pursue with respect to public sector price and quality changes. Column 1 shows the public sector variable; column 2 shows the magnitude of the change in the public sector variable; and column 3 shows possible private sector reactions to changes in public sector pricing and quality changes. Columns 4, 5 and 6 give the respective probabilities of self treatment, public visits and private visits generated by the multinomial probit model, as the independent variables change.107/ Thus, for example, in Case 3, if the public sector doubled its price from a mean value of N1.2 to 12.4, and the private sector responded by doubling its price from a mean value of 12.2 to U4.4, the probability of no consultation/self treatment would increase from 52 percent (at the mean of all of the independent variables in Case 1) to 55 percent; the probability of public visits would remain the same at 35 percent; and the probability of private visits would decline from 13 percent to 10 percent. 105/ Simulations are needed due to the difficulty of directlv interpreting coefficients in multinomlal profit estimation results. 106/ For the simulations, we use only the community-based sample, omitting those individuals who were chosen because a household member was identified at a facility. We believe this to be a random sample of households with access to facilities and the choices predicted for this group to be reasonable estimates for the total population of those with access to facilities. For the reasons explained, however, we utilized the choice-based observations in the weighted sample for estimating the parameters used in the simulation exercises. 107/ The income, urban residence and wealth (own a vehicle) independent variables in Table 5.12 are not considered operator variables here, because they are not policy instruments that come under the portfolio TAME .t16 COST R lVER, tVSS RENL£E SXILATIS FOR PUILIC SETR (1) (2) (3) (4) ( C6) (7) (8) (9) (10) (11) Policy PubIic Sector S etor Trseatoaot PublicPrrivat vote Totl I Revetue. Chane" in Total Option. Variabip Variable Raection Conulttrn Vie Vivito CTreteton Public Private Public Sector ea Revenue Vitate Vsit* Consultattion Visit. Visits (lo; re '000) PbiSc.r (Nire QQ) Case 1 All Variabloo ot mn - - .52 .35 .is 4.412,717 2,970,008 1.109,179 3.500 0 Pricing Variatiom Caeo 2 Public Facility Price Sot to 0 Set to 0 .49 .35 .1S 4,158 137 2.970,098 1.867.759 0 -3.600 Cam 3 a Set to 2.40 Set to 4.44 .6 .35 .10 4 667 297 2 970,098 848,699 7,100 +3.600 Case 4 8 B Set to, 2.40 A. mean .69 .94 .13 4 497 577 2,865,238 1.103,179 6 900 +3.400 Ca"s 5 a e At casn Set to 4.44 .55 .36 .09 4.667.297 3.054.98 763,739 3.700 + 100 (uilitv Variation A Cese 6 Expenditure Per Porson Servod Set to 6.00 Set to 2.72 .50 .37 .13 4,242,"97 8,139.818 1.103,179 3,600 + 200 in Public Facility Cwo 7 * Set tr 6.00 At mean .52 .37 .11 4 412 717 3,189.818 993.459 3,800 * 200 Cao. 8 ' e At mean Set to 2.72 .51 .34 .15 4.327.857 2e885.238 1,272.899 3, SO - 100 Price and Oumlnti Variation. A Coae 9 Public Facility Price Set to 2.40 Set to 4.44 .53 .37 .10 4.497,577 3,139.818 848.599 7,500 C4.ODO Expenditure Per Person Served Set to 6.08 Set to 2.72 in Public Facility Ouslitv Variations S Coca 10 Public Facility Drug Set to . s Set to .50 .69 .31 .00 5,85,S336 2.630,65e 0 3.100 - 400 Avai lab l ity Cas 11 ' Set to 1.0 Set to 1.0 .48 .44 .07 4,073,277 3.733,87 594,020 4,500 + 900 Cas 12 a a Set to 1.0 At aeon .48 .44 .07 4,073,277 3S733,837 594,020 4,500 * 900 Caos 13 a a At man set to 1.0 .52 .35 13 4,412,717 2.970,098 1,103,179 3,600 0 Ruality Variatin C Caw 14 Public Facility Condition Index Set to 3 Set to S .6e .90 .01 5,770,476 2,S46,798 84,860 3,OCO - 500 Case 15 ° Sat to 1 Set to 1 .SO .36 .13 4,242,997 3.O54,958 1,103,179 3,700 * 100 Ca" 16 a a Set to 3 At mean .54 .25 .20 4,52,437 2.121,498 1.697,199 2.500 -1,000 Case 17 e B At mean Set to 3 .61 .38 .01 5 176,486 3,224S678 64,860 3S900 + 300 Prico and Quality Variation. A. A. and C Caeo 18 Public Facility Prico Set to 2.40 Set to 4.44 .47 .47 .06 3,988,417 3.988,417 509,160 9,600 +6.000 Eae nditure Per Person Served Set to 6.00 Set to 2.72 in Public Facility Public Facility Orvg Set to 1.00 Sat to 1.00 Avai labi I ity Public Facility Condition Index Set to 1.00 Set to 1.00 Case 19 Public Facility Price Set to 2.22 Set to 2.22 .47 .46 .07 3.986.417 3.903,567 594,020 8,700 .5,100 Expenditure Per Person Served Set to 6.00 Set to 2.72 in Public Facility Public Facility Drug Sat to .95 set to .998 Avai labilIity Public Facility Condition index Set to 1.16 set to 1.16 N un. of Indeaandent Policy Variabl Public Facility Price Per Visit * 1.1962 (in Nairn, per visit) Private Facility Price Per Visit * 2.2230 Cin Nairs, per visit) Public Faciii Expenditure Per Person Served a 3.0416 in Naira, per pereon in population oerved) Private Facility Ependiture Per Periog Served 1.8571 in Naire, per peras in population served) Public Fbcility Drg Aalability a 0.728 (ercentag of time availabl Privets F eltc D il bItty O. ".9So eenwo tim ev *vatlebl Public Faeility aenral Condition * 1.2667 1 d; 5-far; Svpoor) Private Faciy General Condition * 1.1634 leed; 2fair; por) t ober of Ebiaoike a Frequency x M. where N is the total nwaber of seplf-erceived e-iaoe of Ilinr_. 1/ With reepect to Cas 1, shere oll independent variables are at thlir _mns. I-1t - 99 - 1. Gross Revenue Effects 3. These probabillities can also be interpreted as frequencies of episodes of illness, and the number of episodes of self treatment, public visits, and private visits in columns 7, 8 and 9 in Table 5.16 result from multiplying the total number of episodes of illness times the respective frequencies in columnb 4, 5 and 6. 108/ Using outpatient registration fees as the pri_es, column 10 shows the total revenue that the public sector would generate by each case strategy (with the subsequent private sector reaction); column 11 gives the change in total revenue in the public sector relative to Case 1, where all of the independent variables are at their mean values.109/ 4. Price Variations: Cases 2 through 5 show that total gross revenue from registration fees will change if public and private sector prices also change. For Cases 2 and 3, as public and private prices move in the same direction, the quantity of public visits demanded does not change. Case 2 does show that reducing the public price to zero produces a large deficit in total gross public revenue. Doubling the mean public price (Case 3) produces an equally large surplus. Case 4 also doubles the public sector price, but, because the private sector dies not react, and in fact leaves its prices still higher than those in the public sector, the quantitv of public visits demanded declines only slightly. The change in 108/ The total number of identifiable public and private health care facilities was 437. The weighted average annual number of visits for the public and private facilities in 1986 was 9,321. Therefore, the total number of visits was 4,073,22-. At the mean of the irdependent variables, the sum of the public and private probabilities of visits is .48. Then, N = 4,073,277/.48 - 8,485,994. With a total 1986 population in Ogun Stats of 2,847,521, this would amount to 2.98 episodes of illness per person, and 1.4 visits per person per year for modern health care. 109/ Some of the public sector policy actions in Table 5.16 will cause changes in the quantity demanded, will shift demand, and cause movements along the marginal cost curve and shifts in the marginal cost curve (see Part A). Therefore, a discussion of net revenue results is not possible until the gross cost changes in Table 5.16 are more carefully examined. Also, it must be emphasized that the poor population is taken into account in the discussion by means of reference to the category "self treatment' in columns 4 and 7 in Table 5.14. As the price and quality variables change in Cases 2-19, relative to those in Case 1, the frequency "self treatment/no coisultation" (column 4), and the number of episodes of illness with "self treatment" (column 7) vary respectively in range from 0.47 and 4.0 million (Cases 18 and 19) to 0.69 and 6.0 million (Case 10). Some of the " self treatments' are the people who are being priced out of the market, going from 4.0 million to 6.0 million "self treatments," and some of these people are those who choose to leave the public and/or private formal system for quality-of-care reasons. As will be discussed below, administrative mechanisms have to be implemented to protect the poor who might initially be priced out of ** .* . . .1 , * ., f l' - 100 - total gross public sector revenue is, therefore, strongly positive.110/ Case 5 assumes that the public sector maintains its mean price per visit I while the private sector doubles its per visit price. In this policy opLion, the demand for public sector visits increases only slightly, as does gross public sector revenue. This finding suggests that the cross- price elasticity of demand between the public and private sector is low, holding the other policy variables constant. 5. Quality Variations (Expenditure per Person in the Population Served): Cases 6 throuRh 8 assume changes in public and private sector quality of care provided with no unit price variation.lll/ In Cases 6 and 7, the public sector doubles its expenditure mean (COSTPU),112/ while the private sector reacts by setting its expenditure (COSTPR) higher than the mean in Case 6 and at the mean in Case 7. 113/ Both policy options produce increases in the frequency of public visits and in gross public sector revenue. Finally, Case 8 assumes that COSTPU is at its mean, while the private sector raises COSTPR above the mean. This scenario of increased private quality with no quality imp-ovement in the public sector produces slightly fewer public sector visits and slightly less gross public sector revenue. 6. Combined Price and Quality Variations: Case 9 simulates the combination of price and quality variations. In this policy option, both public and private sector facilities double their per visit prices. Likewise, public and private sector facilities spend twice as much per person in the population served, a doubling of this quality index. Total gross revenue in the public sector in Case 9 far exceeds that of Case 1. The frequency of public sector visits increases somewhat as does the quantity of public visits supplied and demanded.114/ Not only would more people be served by the public sect3r, but the care receive' v*'ould be of better quality. 7. Quality Variations: Cases 10 through 13 shotG results for quality varii relating to changes n the public sector quality index, 110/ Because the percentage negative change in the quantity of public visits demanded is less than the percentage positive change in the public sector price per visit charged, the demand for public visits is price inelastic within this price range, and total revenue increases. 111/ The changes in the frequencies and number of visits should now be interpreted as shifts in the demand curve as the result of quality changes, rather than as movements along the demand curve. 112/ Defined as "public facility expenditure per person in the population served." 113/ Defined as "private facility expenditure per person in the population served.' 114/ There are four economic reasons for these changes: (1) movements along the demand curve caused by the change in price per unit; (2) a demand shift caused by increased quality; (3) movement along the marginal cost function caused by the demand shift; and (4) a marginal cost function shift caused by increased quality. The gross revenue calculations in Table 5.14 only show (for the time being) the revenue effects; the combined revenue and cost effects are considered in the - 101 - drug availability, and private sector reactions, holding all the other independent variables constant at their mean values. Cases 14 through 17 assume a quality variation in the index, physical condition of public facilities, and private sector reactions. The revenue consequences of these drug availability and physical condition simulations may be analyzed in the same manner as we,e Cases 6 through 8, where expenditure per person in the population served was the quality variable. 8. Quality Variations (Drugs, Expenditure and Condition): Cases 18 and 19 show the largest estimated gains in total gross public revenue. In these policy options, Dublic sector price and all three measures of public 3ector quality vary as policy options. In the extreme case (18) where for the public facilities drugs are available 100 percent of the time, expenditure per person doubles, and all buildings are in "good" physical condition, both public and private sector prices can double and still result in a significant increase in public sector revenue. However, as noted below, these quality chaniges cause even larger estimated increases in the costs. In Case 19, the public sector sets its prices at the private level, doubles the expenditure per population served quality index, and sets its other two quality indices at the levels of the private sector. 2. Gross Cost Effects 115/ 9. Gross cost estimates for the 19 scerarios are not as straightforward as the gross revenue estimates.116/ This is mainly because the quality changes that were used as independent variables had no quantitative cost content.117/ Consequently, the frequency changes that result from changing indiependent variables from their means do not directly yield quantitative cost estimates.118/ However, quantitative cost estimates for the purpose of cost comparisons can be derived from the econometric work in paragraphs 5.43-5.47. 119/ 10. Table 5.17 contains the gross cost estimates for the four categories of costs, a total costs column (column 6), and a net surplus or loss column (column 7).120/ 115/ The calculations of revenues and costs are shown in detail in Annex T. 116/ The gross revenue estimates for the public sector in each scenario were simply a function of the price charged (in this case, the registration fee) and the number of public sector visits. 117/ That is, they were relatives (i.e., index-number type vaiiables). 118/ They do, however, reflect movements along demand and marginal cost curves and demand responses to shifts in these curves. 119/ Specifically, they may be derived from the calculations that use cost function and facility expenditure function information. 120/ As these demand quantities change, there will be movement along the marginal cost curve. All of these cost changes are "changes at the margin," i.e., with reference to the mean values of the independent variables. in order not to cause confusion in the mind of the reader, column 2 in Tables 5.16 and 5.17 has been labeled 'Marginal To.al Cost...,' and columns 3, 4 and 5 have been labeled 'Incremental Total Cost...' Using this terminology then, "Marginal" refers to movements the as ' 'in Part A, Table 5.17 CoSM REO CRY. CS CM ATIN AND SURPLIJS OR LOSS FOR PRUC SCoR 1/ (Thouand Noira) C1) (2) (3) (4) (6) (6) (7) (8) Marginal Totel Cost Change in Surplus or Leos Public Change in of Change Inerowental Incremntal Incromental Totel Costs. (CCo. I, Trb 6.16 Pol;cy Sector Public Sector in Quantity Total Coat Total Coat Total Cost Public Sector ainus Cal 7 Optieon Variable 2/ Variable Supplied Of COSTpU of ORPJ of COQU (Cola 3.4.5.6) Tabl i 6.17 Case I 0 0 0 0 0 0 Pricing Variatione CDaO 2 Public Fac;lity Price Per Visit set to 0 0 0 0 0 0 -600 Coa. 8 Public Facility Price Per Vioit Set to 2.40 0 0 0 0 0 +3600 Coas 4 Public Facility Price Per Viit Set to 2.40 -200 0 0 0 -200 *9600 Case 6 Public Facility PriceP er Visit Sat at sean e200 0 0 0 .200 -100 Quality Variations A Coa 6 Public Facility Expenditure Par Person Served Set to 6.08 *400 *6100 0 0 .6500 -6300 Case 7 Public Facility Expenditure Per Person Served Set to 6.06 *400 +6100 0 0 +6500 -600 Coao 6 Public Facility Expenditure Per Person Servod Set at mean -200 0 0 0 -200 *100 Prisce and Quality Variations A __________________-_-- _____---_------- - -- - - -- - - Coae 9 Public Facility Price Per Vitit Sot to 2.40 *400 4610D 0 0 *65CO -2500 Public Facility Expenditure Per Porson Served Set to 6.08 Came 10 Public Facility Drug Availability Set to 0.50 -700 0 -800 0 -1i00 -1100 Coma 11 Public Focility Drug Availability set to 1.0 ,1600 0 +1600 0 *3200 -2300 Coco 12 Public Facility Drug Availability Set to 1.0 +1600 0 +1600 0 +3200 -2300 Coma 18 Public Facility Drug Availability Set at mean 0 0 0 0 0 0 Quality Variatione C ----------------------------------------------------- Coa 14 Public Fciltity Qoneral Condition Set to a -O000 0 0 -l000 -2000 *1500 CasO 1S Public Facility Owteral Condition Set to 1 -200 0 0 .400 *600 -500 Case 16 Public Facility Gcenral Condition Set to 3 -2200 0 0 -1000 -3200 .2200 Cmea 17 Public Facility Goneral Condition Set at meen +600 0 0 0 *600 -300 Price and Quality Variations A, B. and C Cosa 18 Public Facility Price Per Visit Set to 2.40 o2100 .6100 .1900 .400 *10400 -4400 Public Facility Drug Availability Set to 1.00 Public Facility Expenditure Per Person Served Set to 1.00 Caoe 19 Public Facility Price Per Visit Saet to 2.22 +190 *6100 Q1800 422DO 9900 -4800 Public Facility Expenditure Per Freon Served Set to 6.08 Public Facility Drug Availability Set to O. "8s Public Facility General Condition set to 1.16 1/ bale for revenue and coat calculatione are provided in Annex T. 2/ eoane of Independent Pol;cy Voriables Public Facility Price Per Vialt (RPU) . 1.1962 (in Naira. per visit) Public Facility Expenditure Per Person Served (COSTPU) - 3.0416 (in Naira. per person in population served) Pub I t c Fac I it y Druo Ava t I ob t I ty (CARF) * .O727 (pecentae of tise available) Public Facility General Condition (COPU) - 1.2667 (lgood; 2efair. 83oor) No m won um - 103 - 11. Cases 2 through 5 are concerned only with price changes, holding everything else constant. Setting public and private prices to zero (Case 2), or doubling both (Case 3), does not change demand for public visits. Thus the marginal total cost of either one is zerol2l/. Case 2 is predicted to cause a loss of revenue, while Case 3 (where the price of a visit to either a public or a private facility is doubled) would lead to a large increasel22/. In Case 4, where the public sector doubles its prices while the private sector price remains constant, costs to the public sector decrease because fewer clients are served.123/ The pricing scenario in Case 5 (private sector doubles prices, public sector price stands pat), slightly increases the frequency and quantity of public visits and, as a result, small marginal total cost increases are realized.124/ 12. Cases 6 through 9 examine quality variations (changes in expenditure per person in the population served by the public and private sector). Cases 6 and 7 have large and identical net costs to the public sector because they both postulate a doubling of public expenditures per person in the population served. As shown in Table 5.16 even though additional public sector expenditure per person in the population served increases, public sector visits are stimulated, not enough revenue results 121/ That is, within the public sector, there is no movement along the demand curve or the marginal cost curve. 122/ Of course, any surplus could be used to purchase more drugs, and/or to increase public expenditures per person served, and/or to improve the physical condition of public facilities, in the next time period. 123/ In Part A, we estimated that marginal cost in the 24 "efficient facilities (of a total of 68 facilities that were asked to provide utilization data) which were used for the econonometric analysis was N4.82. However, the definition of 'efficiencyf is a function of the number of visits per health worker per year, and these 24 facilities were working at full capacity and with marginal cost above the average incremental cost of N3.73. The data indicate that the other 282 government facilities (305 minus 24) are not working at full capacity and with marginal cost is probably below average cost. Therefore, for this set of estimates, the marginal cost of 4.82 was simply halved to 2.4. Adding these savings to the N3.372 million increase in gross revenue due to the Case 4 pricing scenario (column 11, Table 5.16), produces a net aurplus for Case 4 of N3.575 million that, as with Case 3, could be used to purchase more drugs, etc., in the next time period. The fact that the net surplus for Cases 2 and 3 calculate out to be the same N3.575 million is simply an artifact of the same value of 2.4 being assumed for marginal cost and the doubling of the mean public price to 2.4. Different assumptions about the value of marginal cost and/or about the change in price would produce surplus or loss results for Cases 2 and 3 which would not be the same for both cases (as has happened here). 124/ Essentially, costs rise because more patients are treated, but fees are not increased so little additional revenue is collected from the added users of public facilities. This occurs because the marginal i - 104 - to cover the increased cost.125/ In Case 8, public sector expenditure per person in the population served remains at the mean, but because private sector expenditure per person in the population served increases, public sector visits decline, which produces public cost savings. This case also caused a reduction in revenues, but by a lesser amount than the savings. Case 9 assumes increases in public sector price along with quality improvements, leading to a large increase in the cost of service provision. Even though Case 9 generates increased revenue, the results is a significant net cost to the public sector (Table 5.17). The increase in public sector expenditure per person in the population served and in marginal total cost simply swamp the large increase in revenue. 13. Cases 10 through 13 deal with changes in public sector costs associated withchanges in drug availability in the public and private sectors. In Case 10, drugs from public and private health facilities are available only half the time; the frequency of public visits drops; and both total public sector revenues and costs fall significantly. In this case, government accrues a net revenue gain, but effective health care declines. Cases 11 and 12 (where drugs at public facilities are always available) have identical large visit increases, but the increare in usage drives up total public costs. In both cases the increased costs exceed the increased revenues estimated. In Case 13, drug availability stays at its mean value (while private drugs are always available) and the frequency of public sector visits does no, change; both total revenue and total costs remAin constant. 14. Cases 14 through 17 use an index of the physical condition of public facilitie as the operative policy variable.126/. Changes in physical condition imply incremental cost changes that are a function of the number of facilities and the amount of money needed to improve the facilities, and also, marginal cost changes that are a function of the number of public visits.127/ 15. In Case 14, the condition of public facilities declines to "poor'. The resultant drop in spending per facility on quality produces a total cost saving. This cost saving turns out to be larger than the cost 125/ Specifically, the cost increase is N6,466,853. This value was estimated in the following way. In Ogun State in 1986, there were an estimated 2.85 million people. They were served by 306 public facilities and 131 private facilities, for a total of 437 facilities. Therefore, government served 70? of the entire population (306/437), or 1,993,265 people. If COSTPU increases by 13.04, then 3.04 x 1,993,265 - +6,059,525 Naira. This 16,059,525, caused by doubling COSTPU, is then added to the marginal cost of changes in quantity supplied of 1407,328 (Table 5.16). 126/ 1-good, 2-fair, and 3=poor. 127/ The percent of public recurrent expenditures for "nonpersonnel' (other supplies, transport, maintenance and repair) was 8 percent in 1986. With mean recurrent expenditures par facility of 170,313 mean expenditures on physical condition would be N5,625. Mean public physical condition was 1.2867 (closer to 'good' than to "fair"). To go from a condition of 1.2867 to 1.0 (Case 16) would require a 0.222818 percentage change, or going from a mean expenditure of 15,626 to a mean expenditure of 16.878. Conditiona 9 (feir) and 3 - 105 - revenue from reduced visits: one way for the government to save money is to spend very little, provide poor quality care, and serve few patients -- clearly an unacceptable way to resolve budget problems in the health sector. In Case 15, both public and private sector condition improves to *good" and average facility condition increases from the mean. Incremental costs of this quality improvement are large. The frequency of public visits rises, and, as a resilt, marginal costs increase. In this case, total costs rise more rapidly than total revenue (see Table 5.17), leaving a net cost to the public sector. For this cost, however, more people are served and they receive improved care. 16. Case 16 is somewhat similar to Case 14, except that the condition of private facilities remains at the mean. As with Case 14, the physical condition of public facilities in Case 16 declines and the frequency of public sector visits declines sharply; the private sector takes up most of the slack. Marginal volume costs decline more quickly than total gross revenue, creating a net surplus. In Case 17, the condition of public facilities remains at the mean, so costs remain constant. Because conditions in the private sector drop to "poor," the frequency of visits in the public sector increases, creating additional visit-volume costs. Total gross revenue increases more slowly than total costs, leaving a deficit. 17. Cases 18 and 19 128/ explore the effect on public sector health costs and demand ofsiEarp increases in all of the independent variables (including price). Case 18 assumes a doubled public sector price, a doubling in expenditure per person in the population served, high drug availability, and good facility condition. Public sector visits increase sharply, increasing marginal visit-volume costs.129/ The increase in drug availability combined with the change in-THe number of visits also boosts costs. Facility condition cost is the same as in Case 15. Total public sector costs increase much faster than the increase in total gross revenue and create a large loss. Case 19 is a slightly less ambitious permutation of Case 18 for the public sector. Total costs increase slightly less than in Case 18, but total gross revenue increases at an even slower rate than in Case 18, with the result that the total net loss is 34.8 million, versus a total net loss of N4.4 million in Case 18. 128/ Cases 18 and 19 are included for the purpose of illustrating options. There may, however, be double counting which cannot be separated out and the data is crude, but the results do give an indication of the relative magnitudes. 129.1 The In in per aerved the as in - 106 - NIGERIA: HEALTH CARE COST, FINANCING AND UTILIZATION VI. SUMMARY AND CONCLUSIONS Introduction 6.01 Nigeria has a commendably comprehernive National Health Policy which recognizes that "Primary Health Care is the key to attaining the goal of health for all the people." However, Nigeria's public health care services have serious shortcomings. Ironically, demand is decreasing while the number of potential users is rising. Staff do not adequately serve ruratl areas; facilities suffer from poor maintenancj; and most public sector health care providers have few drugs for their patients. The report examines these problems in the context of public sector spending while noting that total health expenditure far exceeds this amount: individuals spend money on pharmaceuticals and on traditional healers, and in addition, the private sector is active at primary, secondary and tertiary levels. 6.02 After reviewing the context in which the national policies on health and on population are being implemented, the comparative efficiency of public and private facilities was examined in depth, using data from Ogun State. Against this backgro4nd, the household demand for health care in Ogun State, from both private and public sources, was analyzed. Opportunities for policy adjustment, including cost recovery, were then developed. The report concludes that although the basic components may be in place for an effecti.e health care program, in Ogun State specifically and in Nigeria in general, adjusting these components could provide a more balanced and responsive program. To achieve its goal of "Health for All by the Year 2000" the Government will need both to recognize the complementary role of the private sector and to insure that limited public sector health resources are used as efficiently as possible. A. Health System 6.03 Despite the evident shortcomings, Nigeria's health system has made great strides in improving access to health care, especially in training and deploying health manpower. The number of doctors quadrupled to 16,145 during 1975-87. The number of nurses tripled from 17, 904 to 56, 120, and of midwives more than tripled, from 13, 101 to 45,852. Hospital beds almost doubled. The resulting ratio of health professionals to population compares very favorably with ratios in countries with per capita national incomes similar to Nigeria. The comparative data (for 1984) on population per physician in Sub-Saharan Africa show Nigeria had about three times the number of physicians as the regional average and twice the number of nursing staff. 6.04 Many states have an adequate network of dispensaries, clinics, health centers, maternities and hospitals (both public and private), at least in the urban areas. Many of these facilities are relatively new and have a sufficient complement of equipment, although not necessarily in a good state of repair. But the health status of the population is not as high as it could be, relative to the high level of past investment in the sector. - 107 - B. Health Status 6.05 Nigeria is the dominant country in Sub-Saharan Africa, yet it has health status indicators that place it on a par with many of its smaller, poorer neighbors. Given the past commitment of significant resources, such a finding is disappointing and emphasizes the task ahead. Comparative statistics show: 130/ (a) Infant mortality in 1988 was higher in Nigeria (103 per 1,000 live births) than in Togo (92), and Ghana (88). (b) One quarter of Nigerian babies had low birth weights in 1985, a proportion notably higher than in a number of neighboring countries, among them Ghana (17 percent), Togo (20 percent), Benin (10 percent) and Niger (20 percent). (c) Life expectancy at birth for Nigerians of 51 years (in 1988) was average for Sub-Saharan Africa and shorter than life expectancy in Togo (53), Liberia (54), and Ghana (54). C. Health Problems 6.06 The generally low standard of health is related to the limited use of the health care services. The relevance of the emphasis on preventive care as stressed in the National Health Policy is borne out by the fact that an overwhelming number of deaths in Nigeria are due to readily preventable infectious and parasitic diseases, in particular measles, malaria, pneumonia and diarrhea. Childbearing women are especially at risk in Nigeria because of frequent pregnancies and closely spaced births beginning at a very young age and because of medically inadequate and hygienically inappropriate child delivery techniques. In addition, few mothers receive modern prenatal or postnatal care. Use of modern methods of family planning is very low. 6.07 Inadequate nutrition is a growing problem. In 1988, the average caloric supply available in Nigeria was only 94 percent of the minimum requirement, and it is thought to have worsened under the recent constrained economic conditions. 131/ 6.08 The size of Nigeria's rapidly growing population places an added strain on the health sector. Nigeria has one of the highest rates of population growth in the world, estimated at around 3.2 percent per annum in 1985-90. By the year 2030 it could have the third largest population in the world, behind only China and India. Continued rapid population growth 130/ World Development Report, 1990. 131/ World Development Report, 1990. In 1977, daily per capita calorie supply was estimated at 83Z of requirements (World Development Report, 1980), 91? in 1981 (World Development Report, 1984) and 1042 in 1982 (World Development Report, 1985). - 108 - will have negative consequences in a number of sectors, including agriculture, wood resources, employment, education and housing as well as health. The Government, in response to these concerns, launched the National Policy on Population in 1989. As part of the overall goal of improving the quality of life, the policy seeks to bring about continued decl,ne in the mortality of mothers and children and a reduction in fertility. The increased provision of family planning services is seen as vital to the achievement of these goals. The public health sector has a major role to play, together with other health care providers. D. National Health Policy 6.09 In the remarkably frank policy document entitled, "The National Health Policy and Strategy to Achieve Health for All Nigerians" (1988), the Federal Ministry of Health has identified what it sees as the major defects in the health services, including an urban bias, disproportionately high investment in curative services, poor coordination between the public sector and other health care providers, minimal involvement of the community, limited statistics, inadequate f¢rnancial resources, poor maintenance, and lack of medical supplies and drugs. This report goes one step further to analyze the net result: lack of patients. It then suggests ways to attract them back. E. Financing 6.10 One underlying reality that exacerbates Nigeria's health care problems is the decreasing availability of public funds at the federal, state and local levels. State and local spending in Nigeria is highly dependent on federal aid, but the Federal Government itself is highly dependent on oil revenue. The downturn in the oil market in the 1980s seriously reduced the resources available for public spending to the more modest levels of the 1970s. Allocation of funds to the public sector has fallen sharply: the federal health budget (recurrent and capital) in 1990 was only 40 percent in real value that of 1981. This squeeze is felt down through the system, so that states and local governments will likely have experienced a similar reduction in real resources for health care. By comparison, the total actual federal budget in 1990 was cver 90 percent of the real value of 1981, indicating how the health sector has been cut back relative to other parts of the Federal Government. External agencies have tended to overlook Nigeria, with the result that outside assistance for health and population is very small. Ir. 1986, it was estimated at less than $13 million, or $0.12 per capita (see Annex V.) 6.11 Examination of the state and local data for the early 1980s shows that expenditure on nonrecurrent costs (such as maintenance, medical supplies and drugs) has been cut back to the barest minimum, while personnel costs have increasingly dominated recurrent outlays. 6.12 Falling oil revenue helps to explain the fiscal crisis in Nigeria's public health sector at a macro level. At the micro level, the sector faces a woefully inadequate system of cost recovery. Lacking - 109 - resources, especially drugs, the sector provides a poor level of service. despite the availability of trained staff. 132/ 6.13 Low prices also send the wrong signals to health care consumers. Hany people seek treatment for minor health problems at hospital outpatient emergency rooms rather than at clinics, because facilities have a weak signaling system for referrals, and patients are drawn to where supplies are most likely to be found. 133/ F. Ogun State 6.14 This report concentrates on Ogun State and treats it as a paradigm for the Nigerian health system. This stems from the fact that the Federal Ministry of Health has given the Bank access to the results of an unusually thorough and comprehensive survey of health care facilities and consumers that took place in the State in 1987. 134/ While Ogun State residents possibly enjoy relatively more social amenities such as schools, roads and health care than residents of some Nigerian states because of their close proximity to Lagos, the State is not exempt from the fundamental problems noted. 6.15 In the urban areas of Ogun State, private health facilities outnumber public ones. In fact, 82 percent of private health facilities are located in urban areas, and 57 percent of urban health facilities are privately owned. Government care providers tend, but not exclusively so, to locate in rural areas where private care providers are fewest. Almost 75 percent of public care institutions are rural. Although Ogun State's urban/rural population ratio is 30:70, about 44 percent of all facilities are in urban areas while 56 percent are rural. This implies that urban residents enjoy a bias in their favor in the distribution of health facilities in Ogun, a situation typical of Nigeria and other developing countries. Compounding this inequity are the poor state of rural roads and the lack of reliable motor transport to rural areas. Rural residents who require more than primary care thus have to pay more money to travel longer distances in less comfort and safety than urban residents. 6.16 Considering that the private sector owns only 30 percent of all health facilities in Ogun State, it has had great success at attracting doctors. Roughly half of all doctors work in private sector jobs. One explanation for this disparity may be double counting, because it appears 132/ Although in urban areas drugs can be readily purchased in the market place, this still does not encourage use of facilities, but rather helps to account for the high percentage of "self treatment". 133/ In the Ogun State survey of households, for those who sought treatment at a facility, 28? went first to a government hospital and 27X to a primary health care center or dispensary. 134/ The study was undertaken by African Health Consultancy Services, Ltd., Lagos. - 110 - that many publicly employed doctors also have private practices, especially in urban settings. 135/ The ratio of public to private sector employees in other job categories is more in line with the 70:30 ownership rat O: among nurses and midwives (66:34) and support staff (61:39). The ratios get skewed again for community and public health workers (95s5) and pharmacists (83:17). The State Government, which owns only 10 percent of the health facilities in the state, employs 46 percent of the health personnel. The private sector owns about 30 percent of facilities and employs roughly the same proportion of health workers (31 percent). The LGAs manage 50 percent of all of Ogun's health facilities but employ only 20 percent of its health workers. 136/ This is not surprising, since LGA-owned dispensaries are very lightly staffed. Finally, the Federal Government employs only 1 percent of ugun's health workers. 6.17 Using the relatively abundant information available on Ogun State, the report analyzes the efficiency of the state's health care sector, both public and private, the structure of health care demand, and the current and hypothetical levels of statewide health sector cost recovery. The analysis explores ways to improve the sector's efficiency from the point of view of public health policymakers at both the state and federal levels. G. The Analysis 6.18 The financing of health care was examined from two aspects: the demand for care and the supply of it. First, consumers' demand for health care was examined, against their socioeconomic characteristics, and preference for private care, public care or self treatment. Second, from the perspective of the facilities, the efficiency of service provision was examined, comparing the public and the private, to see whether there were any significant cost differences, and if so, what were the key explanatory variables. Also, costs of services were estimated. Putting the two parts together, simulations were run to assess the effect on health care usage of various changes: in price, in condition of facility, in investment per patient served, and in availability of drugs. By combining the information on consumers' preferences for public and private care, and the cost of providing care in public and private facilities, the likely effects of various combinations of policy adjustment-- in both the public and the private sector -- can then be assessed. These are summarized in the discussion below. 135/ Although Ogun State in 1986 had 185 registered doctors (87 in the public sector and 98 in the private) the survey results reported 382 doctors (191 each in the public and private sectors.) 136/ This will shift under the 1990 changes in LGA responsibilities. - ill - 1. Households 137/ (a) Data 6.19 Using a sample of 4,900 household interviews with over 22,000 individuals, and 1,763 adults 138/ who reported an illness in the four weeks prior to the survey, an assessment was made of how people seek care. As the focus of the study was to assess the utilization of facilities, the basic sample was drawni from patients selected randomly at facilities (both public and private), with intprviews carried out at their homes; to obtain a random sample of other households with access to facilities, interviews were also conducted with the neighbors of two of each five interviewed patient households. For the actual demand estimates, the choice- based 139/ methodology was used to weight the choice-based and neighbor samples to create a combined representative sample of potential users, making the results reprecentative of the population of potential facility users. 6.20 The sample analyzed was relacively urbanized (57 percent), half (49 percent) had no formal education, and more than half (56 percent) were too old or too young to work, but around half (52 percent) of household members were employed, and earnings were generally very low. Christians outnumbered Muslims (53 percent vs. 46 percent). Sanitary and health conditions were somewhat deficient. One quarter of the sample (27 percent) had piped water in their home; 43 percent used water from unpiped sources such as ponds, springs, tankers and rivers. A quarter of the sample used flush toilets (24 percent); most used pit latrines (63 percent). 6.21 With the generally low level of earnings, the lowest three income quintiles reported monthly earnings below 1100. The highest income group (with monthly earnings of 1200 and above) accounted for 9 percent of the sample. Surprisingly, self-reported health status differed little by income group. The most obvious differences occurred in the poorest group. Although only 1.7 percent of the poorest quintile said they were in poor health, this figure was almost double the percentage of any other income group. Of the total sample, 4 percent reported they were in excellent health, and 19 percent in fair/poor health. 6.22 The most important indicator of differences in health care use by income group was the extent to which those who were ill sought outside care: one-third of the lowest income group sought no outside help, compared with only one-fifth of the highest income group. The majority of the lowest income groups seeking health care went to public health facilities, with higher income groups making greater use of private facilities. However, the more wealthy groups do not avoid public facilities: over half 137/ Chapter V, Part B. 138/ Children and pregnant women were excluded from the illness analysis. 139/ Chose to use a health facility. - 112 - of the highest income group who used a facility sought help at a public clinic/maternity/hospital. 6.23 The behavior of the 1,763 nonpregnant adults who reported an illness in the four weeks prior to the survey was analyzed. (Intrahouse- hold analyses was omitted and the survey focused only on adult behavior.) Of the sample reporting illness, for the first treatment 68 percent headed to a public facility (28 percent to a hospital, 27 percent to a clinic, 8 percent to a maternity, and 5 percent to a comprehensive health center) and 23 percent to private facilities (12 percent to a hospital, 7 percent to a clinic and 4 percent to a maternity), with the balance seeking first treatment at a pharmacy (5 percent), traditional healer or spiritualist (5 percent). The use of maternity facilities for primary health care by non- pregnant adults is notable. At least 57 percent of preventive/primary care offered at maternities covered immunizations and health education. Another 35 percent covered prenatal and postnatal care. (b) Methodology 6.24 A model was estimated assuming that the choice of health care is a function of the price and quality of the care; sex and education of the patient; wealth and income and location of the household; and type and seriousness of the illness. The price used was the outpatient registration fee, as this was considered to best reflect the price patients consider when deciding whether and where to go for treatment. Quality was derived from the facility operational costs per capita, together with observations on the physical condition of the facility and the availability of drugs. Income entered the model as a control for tastes and preferences, but income interactions with other variables -- such as price -- were not included in the final estimated model because they did not prove to be sta- tistically significant. Various asset measures were tested, and eventually vehicle ownership proved to be a good variable for classifying households by wealth status. The income variable differeatiates the income poor, who appear to behave differently from the other income groups in selecting health care, and the wealth variable appears to differentiate those with high incomes (that is, those who can afford vehicles). The illness symptoms that were eventually included were a combination of "stomach pains or diarrhea." These symptoms are ones for which primary health care would appear to be appropriate. The severity of illness was measured by number of sick days, and variables for urban or nonurban residence, education .level and sex of the ill person, were included. (c) Results 6.25 The results strongly suppurt the view that price and quality affect health care choices and that higher prices can be offset by higher quality. Given equal prices and identical tastes, not surprisingly people 0 prefer health facilitias which have a higher infestment in care and higher drug availability and are also in better condition. The wealthy (e.g., those who own a vehicle) seek the most medical care. Urban residents, females and the more highly educated prefer private medical care. - 113 - 6.26 Simulations were carried out by changing independent variables to determine their hypothetical effect on health care choices for the community with access to facilities. The r.-sults indicated that an individual with sample mean values for al. ariables (including illness symptoms) and facing mean prices and quality for each facility choice, will choose self treatment about half of the time (52 percent), public care one third of the time (35 percent), and private care for the balance (13 percent). 6.27 Price simulations showed that the impact of prices in the selection of health care choices was statistically significant. That the impact is not large is shown by the fact that even charging nothing for health care evidently would not drastically change care choices, nor would doubling health care prices change usage patterns appreciably. Increased investment in health care by both public and private sectors would appear to have a greater impact on public sector usage, with a large movement out of self-care (i.e. "no consultation"). 6.28 The key role played by drug supply in attracting patients was emphasized by the Eimulation results. Availability of drugs apparently affects private sector usage more than public. With the availability of drugs reduced to only half the time at both types of facilities, the simulations predicted a large outflow of users into self-care, particularly from private facilities. Similarly, if physical condition deteriorated for both, there would be a similar outflow to self-care. A final simulation suggested that reasonable price increases combined with reasonable quality improvements rould produce a large increase in public sector usage. 6.29 The simulations oversimplify the ease with which the quality/price trade-off can be implemented, but two basic conclusions can be drawn. First, if revenues from higher prices at public facilities were used to increase care quality, usage of those facilities (and of modern formal care in general) might well increase. Second, the quality levels needed are not unreasonably high; the levels used were a combination of continuing present public facility spending levels in general and increasing drug availability and the quality of facilities to the level now evident in the private sector. However, as quality improvements can be costly, there is a need for careful selection. 6.30 With respect to private facilities, in the simulated case where the public sector maintained its mean price per visit and the private sector doubled its price, no consultation/self treatment increased sharply and there was a small increase in public sector visits, indicating limited price competition between public and private facilities. T}lis suggests that there is only modest scope for significant price increases in the private sector, even given the limited competition from the public sector. 140/ 140/ Whether the private sector could increase profits by a strategy of higher prices and commensurately higher quality was not addressed in the study, but raises a question for future research. - 114 - 2. Facilities 141/ (a) Data 6.31 Turning to the facilities themselves, an inventory of the 445 modern health care facilities in Ogun State was undertaken, with a detailed survey of 86 of the facilities to collect information on utilization, costs, revenues and cost-recovery programs. This data set provided arL unusual opportunity to compare public and private facilities and included for-profit and not-for-profit facilities and facilities with and without inpatients. A surprising observation was that the majority of facilities failed to record a minimum acceptable efficiency level of 600 visits per year per health worker. 6.32 An examination was made of the financial status of public and private providers. The analysis was restricted to small hospitals and maternities lue to limitations on the data. As expected from the aggregate data in Chapter III, personnel costs dominated the budgets. In 1986, personnel costs were 70 percent and 82 percent of public hospital and maternity recurrent costs respectively. In private hospitals and maternities, personnel accounted for 57 percent and 60 percent of recurrent costs. Drugs contributed the largest share of nonpersonnel recurrent costs. In 1986, private hospitals and maternities spent more of their recurrent costs on drugs (26 percent and 21 percent) than similar public facilities (1.., percent and 10 percent). Private hospitals and maternities spent 17 to i8 percent of their recurrent budgets on nonpersonnel, nondrug items, compared with 5 to 8 percent for the public sector facilities. This spending pattern is consistent with observations of health facilities in many other developing countries where inputs complementary to personnel are lacking and drug supplies inadequate. 6.33 Revenue sources were predictably mostly government for the public sector, and mostly cost recovery for the private. For example in 1986 public maternities received 74 percent of revenue from government, and private maternities 82 percent from cost recovery. As Nigeria's current economic situation has squeezed both government budgets and household incomes, both publi: and private facilities are experiencing financial hardship. This was apparent from the profit figures. In 1985 private maternities on average earned only small profits (5 percent), which had turned to losses (minus 13.5 percent) in 1986. This weakening of the financial status of the private sector reinforces the need for coordination between the public and private sectors to avoid duplication of services. 6.34 A comparison of the fee structure of public and private facilities highlighted two issue3. The first is that fees at private sector facilities are higher than at public but the magnitude is not consistent. The second is that lower level facilities, both public and private, do not necessarily have lower fees, which will confuse efforts to establish effective referral systems. 141/ Chapter V, Part A. - 115 - (b) Analysis and Results 6.35 Analyzing the technical and economic efficiency of health care facilities in general, defined as a combination of staff (health workers and non-health workers), beds, and quality of care (measured by availability of drugs), the cost structure of health care service provision was determined. From these findings marginal costs, average costs, and economies of scope and scale were calculated. The results showed all inputs were positively related to the number of visits produced, with an elasticity of scale measure of 17.09 indicating that if, say, all inputs were increased by 1 percent, there would be a 17 percent increase in output. 6.36 Analysis of the cost function, which included wages, outputs, quality, and an inefficiency measure (estimated from the production function), showed that facilities are generally operating on the downward portion of their average cost curvee, especially for admissions. There are short-run increasing returns to scale for inpatient and nearly constant returns for outpatient services. For inpatient services, facilities are operating under capacity, with high fixed costs distributed over relatively few admissions. Low bed occupancy rates are consistent with these .rends in average costs. Increased availability of drugs and other supplies would shift average and marginal costs curves upwards, but utilization would also increase such that the minimum port!on of the average cost curve could be reached. 6.37 A comparison of the relative productivities of workers indicated that in public facilities too many non-health workers support too few health workers. Private facilities were more efficient in their staffing mix. This suggests a careful review is needed of the effect uf the suggested changes in staffing patterns on quality of care. 6.38 The economies of scope measure shows that there are no sizable advantages to combining inpatient and outpatient services by offering these services together. Overall expenditures, therefore, remain constant or increase slightly. Since outpatient services appear to be less affected by undercapacity, the fungibility of various types of health workers should be ex;lored, with the transfer of personnel from inpatient to outpatient care. 6.39 The ability to examine, recommend, implement and evaluate strategies to improve resource allocation within the health sector is limited by the data available. The type of cost data required for analyzing efficiency is, however, rather simple: basic inventories of personnel and accurate measures of health service utilization. A sample format for data collection is provided at Annex I. - 116 - 3. Strategies for Cost Recovery 142/ 6.40 The rerults of assessing wh-t people do when sick and how much it costs to provide health care facil.ties can be combined in simulations of possible cost recovery strategies. The Government, wishing to ensure adequate access to health care and to provide some segment of the population with subsidized health care, faces the dilemma of fostering efficient production and consumption of curative health care services within its own institutions, along with preserving its equity objectives as much as possible. A set of first principles for a cost recovery program has been laid out to assist policymakers in defining such a system of prices (Annex R). 6.41 Using the probabilities of seeking care (developed in Chapter V, Part B), and the cost of providing it (from Chapter V, Part A), various combinations of changes in prices and quality were simulated in an effort to identify the 'best" strategy for public policy. But first a definition of "best" has to be agreed upon. Is the government attempting to maximize the frequency of visits to modern health care facilities, on the grounds that it is presumably more technically efficient if a person has formal health care rather than self-care when that person experiences an episode of illness? If so, maximizing visits to any modern health care facility versus maximizing visits to public sector facilities provides a further alternative. Then the cost of achieving such levels of care has to be factored in. 6.42 The results from Ogun State show that more intensified cost recovery in the public health care sector is an objective both desirable and achievable. Simulations of price variations, with changes in the public and private prices, suggest that cross-price elasticity of demand between the public and private sector is low, holding other policy variables constant. Changes in public and private sector quality of care with no unit price variation produced increases in the frequency of visits: when the public sector doubled its mean, and the private sector held its quality level at the mean or increased it, public sector visits -- and gross public sector revenue -- increased. Further simulations were undertaken with variations in public sector drug availability and physical condition and different private sector reactiors. An extreme scenario of the effect on public sector health costs and demand of sharp increases in all of the independent variables (including price) showed an increase in visits, but costs increased faster than revenues. This highlights the need for carefully identifying the key variables and their relative costs and targeting improvements selectively. 6.43 Assuming that the government wishes to maximize total (public and private) visit-frequency or public-only visit-frequency but avoid sustaining losses, the surprising result is that the two simulated scenarios which maximize visits do not include the changes in quality which both the theoretical and anecdotal information deem to be so important for 142/ Chapter V, Part C. - 117 - a successful cost-recovery program. However, once these static results are reviewed over time, the level of demand for public health care shifts as a result of the quality changes financed by the increased investment. Thus the increased investment then sustains an increased demand. This outcome is consistent with the predicted results. 4. Recommendations (i) Improve cost recovery through higher fees consistently applied. The overwhelming message from the analysis of the Ogun State data is that higher fees at public sector facilities will only marginally discourage potential users. And, assuming that more funds are available to purchase supplies, to maintain equipment, and to build an adequate inventory of drugs, usage will likely increase despite the .Aigher fees. Increased funds would then be available for sustaining the improvements. Maintenance of a simple nationwide Management Information System is fundamental for assessing the impact of investment on health care usage. (ii) Protection of the very poor. Even if usage of needed health care by the poor were not greatly affected by fee increases, the financial burden on the poor of paying the fees would be important. Along with any move to increase charges for health care, therefore, Nigeria should put into effect measures designed to reduce the financial impact of such charges on the poorest people. The most obvious approach would be directly to identify those who are poor and either exempt them from charges or subsidize them. 143/ 143/ Possible approaches to identification of the poor actually in use in developing countries, each with its own practical problems, include certification by citizen committees or local officials; charging small amounts or nothing for "no-frills" levels of room and board in hospitals but full cost for less crowded rooms, better food, etc.; or subsidizing care for specific population groups (for example, infants and pregnant women) or regions of the country. Another possibility is a self-declaration system with specific income and wealth st.%ndards. Those who declared themselves qualified would be accepted as qualified for subsidies, and random (or other) audits, with relatively severe penalties for false declaration, could help to dissuade abuses. As income data from tax records and other information become better and more available, identification of the poor should become more precise and fee systems could be designed to vary by income levels. Meanwhile, in the absence of a means-testing approach, the only alternatives are to charge everyone equally or to charge no one (really the same as the first alternative except the premium is set at zero). Charging everyone the same positive price tends to be unfair to the poor, and the alternative of making health care free has the added problems of (a) failure to raise revenues to pay for the cost of the services used; (b) providing .Lo monetary disincentives to overuse of the health care system; and (c) being regressive, in that those who are able to kay (the relatively well off) have their health care paid for by taxpayers (the average taxpayer generally being relatively poor). - 118 - (iii) Encourage fee collection through retention at site. Given that the bulk of recurrent budgets at both state and local levels is now invested in personnel, little is available for the improvements in quality sought by the public. The Federal Government has recognized the need to adjust the balance in recurrent spending and has issued guidelines spelling out a goal of a maximum of 60 percent of recurrent expenditure on personnel. Achievement of such a goal would increase the availability of funds for maintenance and supplies. The investment of funds in the facilities themselves could be supported by permitting some retention of fees at the sites (for investment and/or incentive payments). (iv) Ensure an adequate supply of drugs. When addressing quality of facilities in Nigeria, the provision of drugs is most often mentioned as a key variable. In the data on Ogun State, which launched a Drug Revolving Fund program in 1984, a relatively high proportion of users were provided with drugs at public facilities (between 73 and 87 percent). However, this was still significantly below the proportion at the private facilities (92 percent). It is recommended that efforts to provide safe and affordable drugs through public health facilities be strengthened. The approval of the National Essential Drugs Policy and implementation of the National Essential Drugs Project have initiated activities to address this acute problem. For sustainability, drug revolving funds need to be based on a transparent payment system: a cash-and-carry system with reimbursement for the poor can greatly reduce administrative problems. (v) Strengthen the referral system through clear price signaling. For efficient service, a patient should generally work up through the system via referral. However, lower level facilities, in both the public and private sectors, do not necessarily have lower fees; for initial treatment in Ogun State, over one quarter of all respondents visited a public hospital for the first consultation. These findings highlight the weak price signaling of the referral system and contribute to the poor cost-effectiveness of the health care providers (See Annex R). (vi) Reduce the share of recurrent costs spent on personnel. The analysis of resource flows in Chapter III showed the high percentage of recurrent budgets at federal, state and local levels spent on personnel. Analysis of facility budgets in Ogun State (Table 5.2) shows a similar picture, which is especially true in the public sector: in 1986, personnel costs were 79 percent and 82 percent of public hospital and maternity recurrent costs respectively, compared to 57 percent and 60 percent for the same facilities in the private sector. Allocative efficiency analysis of the Ogun State data indicates that public facilities are not using cost- minimizing staffing combinations. The public sector appears to employ too many support staff relative to health workers and should consider employing more health workers and/or laying off support staff. (vii) Coordinate public and private health sector supply. The investigation of the health care sector in Ogun State suggests that neither the public nor the private providers are operating at efficient levels. - 119 - They are offering similar services and in some cases using the same staff. The result is duplication of services and higher costs than necessary. As noted in the National Health Policy, coordination with nongovernmental providers is recommended. Such coordination could include recognition of a useful division of activities, with the private sector concentrating on provision of curative care and the public sector on promotive and preventive care, which are already their self-selected directions. This would enable the public sector to focus its limited resources and provide improved services. Some form of support for private sector patients might be required. 144/ (viii) Establish effective management information systems In order to evaluate, implement and monitor adjustments in health policies, basic data needs to be regularly collec.ed and analyzed. The type of data required is simple: basic inventories of personnel and accurate measures of health service utilization (see Annex I). 6.44 This report has attempted to investigate issues and opportunities in the health care sector of Nigeria. The period analyzed is the mid-1980s. Recognizing the strength of the sector, the weaknesses in delivery, the diminution of resources and the growing needs of the people, continued adjustments in the operating approach and reallocation of resources are needed. The system currently is being reoriented toward more accessible primary health care, with increasing emphasis on preventive and promotive services. It is hoped that this report provides some insights into further opportunities for policymakers seeking to achieve the goal of "Health for All by the Year 2000." 144/ In addition, the strong role of traditional healers suggests collaboration with them as well, as noted in the National Health Policy. Their influence in promoting health programs such as nutrition, environmental sanitation, family planning and safe motherhood could be further developed to the general benefit of the public. Various options for collaboration between traditional and Western medicine exist, including licensing and training, supply of drugs and equipment, and sharing of information. Including traditional birth attendants in maternal and child health care programs can greatly increase service coverage. The range of options A. Ai . te _ - 120 - BIBLIOGRAPHY Abel-Smith, B. 1989. "Health Economics in Developing Countries." Journal of Tropical Medicine and Hygiene 92:229-41. 1-ton, J. P. 1973. Demand for Health Care When Time Prices Vary More than Money Prices. Santa Monica, CA: The Rand Corporation. Acton, J. P. 1975. Nonmonetary Factors in the Demand for Medical Services: Some Empirical Evidence. Journal of Political Economy. 83:595-614. Akin, J. S. 1986. "Equity for the Poor.' World Health: The Magazine of the World Health Organization (May):28-29. Akin, John C., Nancy Birdsall and David De Ferranti. 1987. Financing John Health Services in Developing Countries: An Agenda for Reform. Washington, D.C.: World Bank. Akin, J. S. and B. Schwartz. 1986. "Demand Analysis with Price Information Obtained Directly from Respondents." World Bank, Population, Health and Nutrition Technical Note 86-22 (June):1-24. Akin, J. S., and B. Schwartz, 1988. "The Effect of Economic Factors on Contraceptive Method Choice in Jamaica and Thailand: A comparison of Hixed Multinomial Logit Result." Economic Development and Cultural Change 36(3). Akin, J. S., C. C. Griffin, D. K. Guilkey, and B. M. Popkin. 1986. "The Demand for Primary Health Care in the Bicol Region of the Philippines." Economic Development and Cultural Change 34(4):755- 82. _ 1986b. "The Demand for Adult Outpatient Services in the Bicol Region of the Philippines." Social Science and Medicine 22(3):321-28. . 1985. The Demand for Primary Health Care in the Third World. Totowa, NJ: Littlefield, Adams. ___________. 1985b. "Determinants of Infant Feeding: A Household Production Approach. Economic Development and Cultural Change 34(1)s57-81. Akin, J. S., D. K. Guilkey, and B. M. Popkin. 1981. "The Demand for Child Health Services in the Philippines." Social Science and MedVcine 15:249-57. Anderson, D. L. 1980. "A Statistical Cost Function Study of Public General Hospitals in Kenya." Atkinson, S. E. and R. Halvorsen. 1984. "Parametric Efficiency Tests, Economies of Scale, and Input Demand in the U. S. Electric Power Generation." International Economic Review 25:647-62. - 121 - Bailey, E. E. and A. F. Friedlander. 1982. "Market Structure and Muitiproduct Industries." Journal of Economic Literature 20:1024-48. Birdsall, Nancy. 1987. Strategies for Analyzing Effects of User Charges in the Social Sectors. PHN Technical Note 87-1. Washington, D.C.: World Bank. Bitran-Dicowsky, R. and D. W. Dunlop. 1987. "The Determinants of Hospital Expenditures: An Analysis of Ethiopia." September. Mimeo. Blair, Roger D., P. B. Ginsburg and R. J. Vogel. 1975. "Blue Cross-Blue Shield Administrative Costs: A Study of Non-Profit Health Insurers." Economic Inquiry 1?:237-51. Brown, R. S., D. W. Caves and L. R. Christensen. 1982. "Estimating the Structure of Costs and Production for Multiproduct Firms." Southern Economic Journal 46:256-73. Caves, D. L., L. R. Christensen and M. Tretheway. 1980. "Flexible Cost Functions for Multiproduct Firms." Review of Economics and Statistics 62:477-81. Cowing, T. G. and A. G. Holtmann. 1983. "The Multiproduct Short-Run Hospital Cost Function: Empirical Evidence and Policy Implications from Cross Section Data." Southern Economic Journal 4:637-53. Cowing, T. G-, 2oltmann A. G., and Powers S. 1983. "Hospital Cost Analysis: A Survey and Evaluation of Recent Studies. In Advances in He,'th Economics Research. Ed. Scheffler and Rossiter. Connecticut, Vol. 4, pp. 257-303. Danzon, Patricia. 1982. "Hospital 'Profits': The Effects of Reimbursement Policies." Journal of Health Economics 1:29-52. De Ferranti, David. 1985. Paying for Health Services in Developing Countries: An Overview. World Bank Staff Working Paper, no. 721. Washington, D.C.: World Bank. iJeJong, J. 1989. 'Traditional Medicine in Sub-Saharan Africa: A Review of Its Importance and Potential Policy Options." Dor, A., P. Gertler, and J. van der Gaag. Forthcoming. 'Nonprice Rationing and the Choice of Medical Provider in Rural Cote d'Ivoire.' . , Journal of Health Economics. Dor, A. 1988. 'The Marginal Costs of Medicare Patients in Nursing Homes.' Working Paper 3586-04. Washington, D.C.: The Urban Institute. Eakin, B. K. and T. J. Kniesner. 1988. "Estimating a Non-Minimum Cost Function for Hospitals.' Southern Economic Journal. January. Feldstein, Paul J. 1988. Health Care Economics. 3d ed. New York, N.Y.: John Wiley and Sons. Fournier, G. M. and J. M. Mitchell. 1988. 'Hospital Costs and Competition 1, ^ ' : A of. r I - 122 - Frank, R. G. and C. A Taube. 1987. "Technical and Allocative Efficiency in Production of Outpatient Mental Health Clinic Services.' Social Science and Medicine 24:843-50. Gertler, P., L. Locuy, and W. Sanderson. 1987. "Are User Fees Regressive?: The Welfare Implications of Health Care Financing Proposals in Paris." Special issue of the Journal of Econometrics 36:67-88. Gertler, Paul and J. van der Gaag. 1988. Measuring the Willingness to Pay for Social Services in Developing Countries. LSMS Working Paper, no. 45. Washington, D.C.: World Bank. Goldman, F. and M. Grossman. 1978. "The Demand for Pediatric Care: An Hedonic Approach." Journal of Political Economics 86:259. Goldman, F. and M. Grossman. 1983. "The Production and Cost of Ambulatory Medical Care in Community Health Centers." In Advances in Health Economics and Health Services. Ed. Scheffler and Rossiter. Connecticut, Vol. 4, pp. 1-56. Granneman, T. W., R. S. Brown and M. V. Pauly. 1986. "Estimating Hospital Costs: A Multiproduct Output Analysis." Journal of Health Economics 6:107-28. Grossman, M. 1972. The Demand for Health: A Theoretical and Empirical Investigation. National Bureau of Economic Research Occasional Paper 119. New York: Columbia University Press. Heller, P. S. 1976. A Model of the Demand for Medical and Health Services in West Malaysia. Discussion Paper 62. Center for Research on Economic Development. Ann Arbor: University of Michigan. Holtman, A. G. 1972. "Prices, Time and Technology in the Medical Care Market. Journal of Human Resources 7:179. Holtman, A. G. and E. 0. Olsen. 1976. "The liemand for Dental Care: A Study of Consumption and Household Production. Journal of Human Resources 9:546. Jimenez, Emmanuel. 1987. Pricing Policy in the Social Sectors: Cost Recovery for Education and Health in Developing Countries. Baltimore, Md.: Johns Hopkins University Press. Joseph, H. 1971. "Empirical Research on the Demand for Health Care." Inquiry 8:61. Leibowitz, A. and B. Friedman. 1979. "Family Bequests and the Derived Demand for Health Inputs." Economic Inquiry 17:419. McFadden, D. 1973. "Conditional Logit Analysis of Qualitative Choice Behavior." in P. Zarembka (ed.), Frontiers in Econometrics. New York: Academic Press. Maddala, G. S. ldate. "Limited Dependent and Qualitative Variables in Econometrics." Econometric Society Magazine, Cambridge University. - 123 - Menzel, Paul J. 1983. Medical Costs, Moral Choices. New Haven, Conn.: Yale University Press. Mitchell, J. M. and G. M. Fournier. 1988. "Does Competitive Entry Lead to Higher Costs in the Hospital Industry?" March. Mimeo. Mwabu, G. M. 1989. "Referral Systems and Health Care Seeking Behavior of Patients: An Economic Analysis." World Development 17:85-92. Oppong, A.C.K. 1989. "Healers in Transition." Social Science Medicine 28:605-12. Palmer, Hans C. and Ronald J. Vogel. 1985. "Models of the Nursing Home." In Long Term Care: Perspectives From Research and Demonstrations. Ed. Ronald J. Vogel and Hans C. Palmer. Rockville, Md.: Aspen Press. Paul-, Mark V. and M. Redisch. 1973. "The Not-For-Profit Hospital as a Physician's Cooperative." American Economic Review 63:87-99. Phelps, C. E. and J. P. Newhouse. 1974. "Coinsurance, the Price of Time, and the Demand for Medical Services." Review of Economic Statistics 56:334. Reinhardt, Uwe. 1972. "A Production Function for Physician Services." Review of Economics and Statistics pp. 55-66. Samuelson, Paul A. 1954. "The Pure Theor7 of Public Expenditure." Review of Economics and Statistics 36:387-89. Schmidt, P. 1986. "Frontier Production Functions." Econometric Reviews 1985-1986 4:289-328. Schwartz, J. B., J. S. Akin, and B. M. Popkin. 1988. "Price and Income Elasticities of Demand for Modern Health Care: The Case of the Infant Delivery in the Philippines." World Bank Economic Review 2(1) :49-76. Terza, J. V. 1985. "Reduced-Form Trinomial Probit: A Quantal Response Model Without A Priori Restrictions." Journal of Business and Economic Statistics, 3(1):54-59. Thobani, Mateen. 1983. "Efficiency and Equity Implications of User Charges in Social Sector Services: The Financing of Education in Malawi." World Bank Staff Working Paper, no. 572. Washington, D.C: World Bank. Toda, Y. 1976. "Estimation of a Cost Function When the Cost is Not Minimum: The Case of Soviet Manufacturing Industries, 1958-1971." Review of Economics and Statistics 58:259-68. Vogel, Ronald J. 1977. "The Effects of Taxation on the Differential Efficiency of Non-Profit Health Insurance." Economic Inquiry 15:605-9. Vogel, Ronald J. 1988. "Cost Recovery in the Health Care Sector: Selected Country Studies in West Africa." World Bank Technical Paper, no. 82. Washington, D.C.: World Bank. - 124 - Vogel, Ronald J. 1989. "Trends in Health Expenditures and Revenue Sources in Sub-Saharan Africa." Monograph prepared in conjunction with the World Bank Sub-Saharan Africa Health Policy Study. March 1. Mimeo. Vogel, Ronald J. and B. Stephens. 1989. "Availability of Pharmaceuticals in Sub-Saharan Africa: Roles of the Public, Private and Church Mission Sectors." Social Science and Medicine 29:479-86. Waddington, Catriona J. and K. A. Enyimayou. 1989. "A Price to Pay: The Impact of User Charges in Ashanti-Akim District." International Journal of Health Planning and Management 4:17-47. -I Wong, Emelita, B. M. Popkin, D. K. Guilkey, and J. S. Akin. 1987. "Accessibility, Quality of Care and Prenatal Care Use in the Philippines." Social Science and Medicine 24:927-44. MAP SECTION I lAGOS -~~~~ KNAA) eAN I .~ /I 4~~~~~~~~~~~~~~~~~~~~~~X M N BfA0\<1 II ~~~~~~~~~KVI, tA A .I)AN 13T) WI OII)RN DO f lt \ LAGOS 131 NDI I f ( IJIt (I -NX -.e it, ttI GE RI ow2RRI z ~~~~~~~~~AKWA-IBOM