102428 African Programme for Onchocerciasis Control 1995– 2015: Model-Estimated Health Impact and Cost ´ 2, J. Lennert Veerman1,3, Koffi B. Agblewonu2, Luc E. Coffeng1*, Wilma A. Stolk1, Honorat G. M. Zoure 4 2 2 Michele E. Murdoch , Mounkaila Noma , Grace Fobi , Jan Hendrik Richardus1, Donald A. P. Bundy5, Dik Habbema1, Sake J. de Vlas1, Uche V. Amazigo2,6 1 Department of Public Health, Erasmus MC, University Medical Center Rotterdam, Rotterdam, The Netherlands, 2 African Programme for Onchocerciasis Control, Ouagadougou, Burkina Faso, 3 School of Population Health, The University of Queensland, Brisbane, Australia, 4 Department of Dermatology, Watford General Hospital, Watford, United Kingdom, 5 Human Development Network, The World Bank, Washington, D.C., United States of America, 6 Independent Consultant, Enugu, Nigeria Abstract Background: Onchocerciasis causes a considerable disease burden in Africa, mainly through skin and eye disease. Since 1995, the African Programme for Onchocerciasis Control (APOC) has coordinated annual mass treatment with ivermectin in 16 countries. In this study, we estimate the health impact of APOC and the associated costs from a program perspective up to 2010 and provide expected trends up to 2015. Methods and Findings: With data on pre-control prevalence of infection and population coverage of mass treatment, we simulated trends in infection, blindness, visual impairment, and severe itch using the micro-simulation model ONCHOSIM, and estimated disability-adjusted life years (DALYs) lost due to onchocerciasis. We assessed financial costs for APOC, beneficiary governments, and non-governmental development organizations, excluding cost of donated drugs. We estimated that between 1995 and 2010, mass treatment with ivermectin averted 8.2 million DALYs due to onchocerciasis in APOC areas, at a nominal cost of about US$257 million. We expect that APOC will avert another 9.2 million DALYs between 2011 and 2015, at a nominal cost of US$221 million. Conclusions: Our simulations suggest that APOC has had a remarkable impact on population health in Africa between 1995 and 2010. This health impact is predicted to double during the subsequent five years of the program, through to 2015. APOC is a highly cost-effective public health program. Given the anticipated elimination of onchocerciasis from some APOC areas, we expect even more health gains and a more favorable cost-effectiveness of mass treatment with ivermectin in the near future. ´ HGM, Veerman JL, Agblewonu KB, et al. (2013) African Programme for Onchocerciasis Control 1995–2015: Model-Estimated Citation: Coffeng LE, Stolk WA, Zoure Health Impact and Cost. PLoS Negl Trop Dis 7(1): e2032. doi:10.1371/journal.pntd.0002032 ´n ´a-Gloria Basa Editor: Marı ˜ ez, Imperial College London, United Kingdom Received March 12, 2012; Accepted December 11, 2012; Published January 31, 2013 Copyright: ß 2013 Coffeng et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Funding: This study was funded by the World Health Organization/African Programme for Onchocerciasis Control (APOC/CEV/322/07). H.G.M. Zoure ´, K.B. Agblewonu, M. Noma, G. Fobi, U.V. Amazigo are or have been employees of the African Programme for Onchocerciasis Control (APOC), World Health Organization. The authors state that their employment has not caused any conflict of interest in any of the following: study design, data collection, data analysis, interpretation, decision to publish. Competing Interests: The authors have read the journal’s policy and report the following matters, which could be perceived as possible sources for conflict of interests: The International Bank for Reconstruction and Development/World Bank serves as Fiscal Agent to WHO APOC through its fiduciary oversight of the APOC Trust Fund and contributed to this paper through its work with the APOC secretariat in the maintenance of accurate financial records. The findings, interpretations and conclusions expressed in this paper do not necessarily reflect the views of the World Bank. Furthermore, H.G.M. Zoure ´, K.B. Agblewonu, M. Noma, G. Fobi, U.V. Amazigo are or have been employees of the African Programme for Onchocerciasis Control (APOC), World Health Organization. The authors state that their employment has not caused any conflict of interest in any of the following: study design, data collection, data analysis, interpretation, decision to publish. * E-mail: l.coffeng@erasmusmc.nl Introduction exposure to microfilariae. Most of the global burden of oncho- cerciasis (.99%) is found in sub-Saharan Africa. In the West Onchocerciasis is caused by Onchocerca volvulus, a filarial African savanna, where onchocerciasis is of a severely blinding nematode restricted to human hosts. The adult female worms form (savanna type), fear of blindness previously led to abandon- reside in subcutaneous nodules where they produce millions of ment of fertile river basins. However, by now, onchocerciasis has microfilariae during their on-average ten-year life span [1]. The been largely eliminated from West Africa by the Onchocerciasis microfilariae are found predominantly migrating through the skin Control Programme (1974–2002), which relied on intense vector and eyes and are transmitted by biting flies of the genus Simulium control and mass treatment with the drug ivermectin [2]. (the vector), an obligatory part of the parasite’s life cycle. In the more central and eastern parts of Africa, where Onchocerciasis is responsible for a considerable burden of disease, onchocerciasis is usually of the less blinding form (forest type), mainly because of visual impairment, blindness, disfiguring skin there was no control or control only at a limited scale until the lesions, and severe itching, which are the results of continuous inception of the African Programme for Onchocerciasis Control PLOS Neglected Tropical Diseases | www.plosntds.org 1 January 2013 | Volume 7 | Issue 1 | e2032 APOC 1995-2015: Health Impact and Cost Author Summary calculations, we used ONCHOSIM, an established micro- simulation model for transmission and control of onchocerciasis In 1995, the World Health Organization launched the [6,7]. African Programme for Onchocerciasis Control (APOC) with the aim to control morbidity due to the parasitic infectious Methods disease onchocerciasis (river blindness). APOC aims to set up sustainable national control programs against oncho- Project-population by endemicity category and project- cerciasis in 16 countries in sub-Saharan Africa, covering specific history of control over 100 million people who are at risk for infection. The The impact of APOC was estimated at project level (a project main control strategy is mass treatment with the drug being an implementation unit for mass treatment with ivermectin), ivermectin, which is donated by the pharmaceutical while taking account of the prevailing type of onchocerciasis (i.e., company Merck. Coordination of the mass treatment savanna versus forest or mixed forest/savanna, with different programs is made possible by financial contributions from donor and beneficiary countries. We estimated that patterns of morbidity) and the project-specific history of control. between 1995 and 2010, APOC has had a huge impact Project populations were further stratified by endemicity groups, on population health in sub-Saharan Africa, preventing 8.2 to take account of differences in the pre-control prevalence of million years worth of healthy life from being lost due to morbidity (which is non-linearly associated with infection) and the disease and mortality, at a cost of about US$257 million. potential impact of mass treatment (e.g., the impact is relatively We predicted that this health impact will double during lower in highly endemic areas due to more residual transmission the subsequent five years, at a cost of about US$221 between treatment rounds). We considered four endemicity levels: million. This makes APOC one of the most cost-efficient non-endemic (prevalence of onchocercal nodules in adult males large-scale public health programs in the world. We may ,1%), hypoendemic (nodule prevalence $1% and ,20%), expect even greater health gains in the future, given the mesoendemic (nodule prevalence $20% and ,40%), and anticipated extension of the APOC mandate with the aim hyperendemic (nodule prevalence $40%). to eliminate infection where possible. We estimated the size of the population at risk for infection in the 107 geographical project areas covered by APOC, for the years 1995–2010 (see File S1). These estimates were based on (APOC) in 1995. APOC is a morbidity control program scheduled records kept by community-appointed drug distributors, aggre- to be active until 2015, requiring that by that year, participating gated to the project level. From the same data, we took the countries support and coordinate control measures independently. reported number of individuals who were treated with ivermectin Since 1995, APOC has mapped infection with O. volvulus in 20 during mass treatment (File S1) and calculated the average countries [3] and has coordinated interventions in 16 countries therapeutic coverage of mass treatment in each project per where onchocerciasis is considered a public health problem calendar year (i.e., the fraction of the population at risk that was (Angola, Burundi, Cameroon, Central African Republic, Chad, treated). Based on data from extensive pre-control mapping Congo, Democratic Republic of Congo, Equatorial Guinea, studies, we estimated the fraction of the population in the different Ethiopia, Liberia, Malawi, Nigeria, South Sudan, Sudan, endemicity categories and the mean pre-control infection level in Tanzania, and Uganda), covering endemic areas inhabited by each endemicity category (File S1). about 71.5 million people in 1995. APOC’s main strategy is to For the years 2011–2015, we assumed that population size will implement annual mass treatment with ivermectin. increase according to the latest known national growth rate (as Ivermectin kills microfilariae and permanently reduces the reported by the United Nations World Population Prospects, production of microfilariae by adult female worms, slowing down published 11 May 2010, accessed 24 October 2011). If therapeutic transmission and preventing morbidity [4,5]. Annual mass coverage in 2010 was already at or above 75%, we assumed that treatment with ivermectin is implemented through a communi- coverage in the years 2011–2015 will remain equal to that in 2010. ty-directed treatment approach, empowering communities to take For those few project in which this was not yet the case, we responsibility for ivermectin delivery and to decide how, when, assumed that between 2011 and 2015, therapeutic coverage will and by whom ivermectin treatment is administered. Mass be scaled up by 10 percentage points per year (conservative treatment with ivermectin is enabled by donation of the drug by compared to reported coverage patterns in projects that started the pharmaceutical company Merck through the Mectizan mass treatment between 1995 and 2010), to a maximum of 75% Donation Program. Furthermore, coordination of the program is (conservative compared to the longest-running projects that funded by donor countries (through the World Bank) and national reported stable coverage levels around 80% in 2008–2010). onchocerciasis task forces (including beneficiary governments and non-governmental development organizations). To demonstrate Simulating trends in infection and morbidity APOC’s importance, validate the efforts of endemic communities For each unit of analysis (project, onchocerciasis type, and national task forces, and maintain commitment of all endemicity), we simulated trends in infection, morbidity, and stakeholders, it is essential to establish the health impact and cost mortality in the ONCHOSIM model [6–8], considering the of APOC. project-specific history of mass treatment (File S1). For each Here, we present the estimated impact of APOC on population endemicity stratum, ONCHOSIM was calibrated so that it could health and the costs involved up to 2010, with extrapolated trends reproduce the average pre-control level of infection (File S1). up to 2015. An impact assessment would ideally be based on Furthermore, ONCHOSIM was calibrated to reproduce the observed trends of infection and morbidity, but such longitudinal association between the prevalence of infection and morbidity data are of limited availability in APOC areas. We therefore (visual impairment, blindness, and itch) as estimated by analysis of estimated trends of infection and morbidity based on APOC data literature data (File S1). Based on previous studies with of pre-control levels of infection and history of mass treatment, and ONCHOSIM, we assumed that ivermectin instantly kills all literature-derived associations between infection and morbidity microfilariae and permanently reduces the capacity of adult female and the effect of treatment on infection and morbidity. For our worms to release microfilariae by 35% in treated individuals (with PLOS Neglected Tropical Diseases | www.plosntds.org 2 January 2013 | Volume 7 | Issue 1 | e2032 APOC 1995-2015: Health Impact and Cost cumulative effects for repeated treatments) [4,7]. Individual estimated from a program perspective, not accounting for participation in mass treatment was assumed to depend on age, community costs and costs of donated drugs. For the years sex (pregnant women and children under the age of five were 1995–2003 and 2010, cost data for national onchocerciasis task assumed to be excluded from treatment), random non-compliance forces were not available and were assumed to be proportional to (i.e., temporal factors), and systematic non-compliance (i.e., fixed APOC expenditures by a factor based on data available for other individual factors other than age and sex e.g. inclination towards years. Expenditures for 2011–2015 were estimated based on the participation). Systematic non-compliance was assumed to play a expected number of treatments in that period multiplied by the larger role when overall treatment coverage was lower (i.e. when estimated cost per treatment in 2010. All costs are reported in there is lower inclination to participate in general), and vice versa nominal values, by which we mean that the presented costs are the [6,8]. No simulations were performed for hypoendemic areas, as amounts that were actually spent (i.e. uncorrected for inflation, ONCHOSIM predicts that transmission of infection is unsustain- and undiscounted). able without migration of infected flies and/or human, and information on migration was lacking. Instead, we assumed that Results the prevalence of infection and morbidity in hypoendemic areas was 1/3 of that in mesoendemic areas, both pre-control and In 1995, the total population size in the APOC target area was during control. For non-endemic areas, we assumed that 71.5 million (Figure 1), with 30% of the APOC target population prevalence of infection and morbidity was always zero. living in hyperendemic communities, 31% in mesoendemic communities, 38% in hypoendemic communities surrounded by Calculating the health impact mesoendemic or hyperendemic areas, and 1% living in non- We combined the predicted trends in prevalence of infection, endemic communities. About 30% of the APOC population lived morbidity, and mortality with information on the number of in savanna areas and 70% in forest or forest–savanna mosaic areas people at risk, yielding an estimate of the absolute number of cases (Table 1). Before the inception of APOC in 1995, about 32 million of infection, morbidity, and deaths in each stratum. After people (45%) in APOC areas were infected with onchocerciasis, aggregation of these results over all APOC projects, we calculated with 404,000 people (0.6%) blind because of onchocerciasis, the burden of disease in terms of disability-adjusted life years another 889,000 (1.2%) suffering from visual impairment, and 10 (DALYs), which in our case is the sum of years lived in disability million people (14%) suffering from troublesome itch. In the same due to troublesome itch, visual impairment, and blindness, year, a total of 1.6 million DALYs (22.8 DALYs per 1,000 persons) weighted by the loss of quality of life due to each symptom: were lost due to onchocerciasis: 694,000 because of troublesome 0.068, 0.282, and 0.594, respectively [9]; and years of life lost due itch, 684,000 from blindness, and 251,000 due to visual to excess mortality from blindness (File S1). Every incident case of impairment. blindness was attributed 8 years of life lost, based on the average Mass treatment effectively started in 1997 (80,000 treatments) age of onset of blindness in ONCHOSIM, the associated life- and was scaled up over the years, reaching an overall therapeutic expectancy (16 years) of a healthy person of the same age, and an coverage of about 73% in 2010 (75.8 million treatments; Figure 1). estimated 50% reduction in remaining life-expectancy due to We estimated that the therapeutic coverage will increase to 78% blindness (File S1). The estimated annual burden of disease was by 2015 (92.5 million treatments). By 2010, about 65% of the compared to the burden in a counterfactual scenario in which the population lived in areas subjected to 10–13 rounds of mass pre-control prevalence of infection and morbidity did not change treatment, 17% in areas subjected to 6–9 rounds of mass (i.e., as if there were no mass treatment), yielding an estimate of the treatment, 18% in areas subjected to 3–5 rounds of mass averted disease burden. All DALY estimates in the present study treatment, and less than 1% in areas subjected to only 1–2 rounds are undiscounted. of mass treatment (Table 1). Cumulatively, about 500 million treatments with ivermectin were given between 1995 and 2010, Sensitivity analysis with another 430 million expected to follow in the period 2011– We assessed the influence of uncertain model assumptions on 2015. Considering the differences between projects in start year the estimated health impact, by means of univariate and and patterns of scaling up of mass treatment, the prevalence of multivariate sensitivity analyses (File S1). In a univariate sensitivity infection for APOC as a whole declined gradually and non-linearly analyses, we assumed extreme, though plausible parameter values over time, from 45% in 1995 to 31% in 2010, and to 18% in 2015 for each of the selected parameters. In a multivariate sensitivity (Figure 2). Similarly, the prevalence of troublesome itch was analysis, the analysis was repeated, based on 200 sets of random reduced from 14% to 6% to 2%, and prevalence of visual parameter values. Parameter values were randomly drawn from impairment was reduced from 1.2% to 0.8% to 0.6%. Because of triangular distributions with modes equal to the values used in the excess mortality among the blind and the fact that ivermectin main analysis, and minimum and maximum values equal to those prevented blindness in individuals who were already visually used in the univariate sensitivity analyses. To arrive at a crude impaired, the prevalence of blindness declined more rapidly than estimate of the uncertainty in the estimated health impact, the that of visual impairment: from 0.6% to 0.3% to 0.2%. results of the multivariate sensitivity analysis were expressed as the In the counterfactual scenario without mass treatment, in which 2.5 and 97.5 percentiles of results from 200 repeated analyses. levels of infection and morbidity were stable, the absolute number of DALYs lost due to onchocerciasis would have increased over Estimating the cost of APOC the years with population growth. In contrast, in the scenario that We estimated the financial costs for coordination of ivermectin considers mass treatment with ivermectin, the absolute number of mass treatment taken on by APOC and national onchocerciasis DALYs lost was predicted to decrease over the years. Due to these task forces (beneficiary governments and non-governmental divergent trends, the number of DALYs averted by mass treatment development organizations), based on APOC financial reports with ivermectin was predicted to increase year by year (Figure 3). for The World Bank, which acts as fiscal agent for APOC. Because Overall, mass treatment with ivermectin averted 8.2 million governments of beneficiary countries will eventually have to DALYs between 1995 and 2010 (3.2 million due to itch, 4.4 finance and coordinate ivermectin mass treatment, costs were million due to blindness, 0.6 million due to visual impairment). PLOS Neglected Tropical Diseases | www.plosntds.org 3 January 2013 | Volume 7 | Issue 1 | e2032 APOC 1995-2015: Health Impact and Cost Figure 1. Population at risk and treated in areas covered by the African Programme for Onchocerciasis Control. Dots represent time points for which data were available; projections for 2011–2015 (shaded area) are based on the assumptions that populations continue to grow according to the latest known growth rates and that all projects scale up therapeutic coverage by 10 percentage points per year (up to a maximum coverage of 75%). doi:10.1371/journal.pntd.0002032.g001 Moreover, we expect that APOC will avert another 9.2 million number of DALYs averted could be up to 25% higher or lower, when DALYs in the period 2011–2015, adding up to an expected total we considered the separate sources of uncertainty simultaneously of 17.4 million averted DALYs by 2015 (Table 2). In relative (6.0–9.8 million DALYS cumulatively averted by 2010, and 13.1– terms, the disease burden of onchocerciasis was reduced from 22.8 21.3 million DALYs cumulatively averted by 2015; Figure 4). DALYs per 1,000 persons in 1995 to 9.6 DALYs per 1,000 Between 1995 and 2010, coordination of mass treatment cost persons in 2010, and is expected to be further reduced to 5.0 roughly US$257 million (Table 2), of which US$175 million was DALYs per 1,000 persons by 2015. disbursed by APOC and US$82 million by national onchocerciasis Univariate sensitivity analyses identified the following parameters task forces (cost of donated drugs and government salaries not as having the most influence on the estimated health impact: the included). Assuming that costs will rise proportionally with the population at risk, pre-control levels of infection, and the associations number of treatments, mass treatment was expected to cost between infection and itch and eye disease (Figure 4). The another US$221 million between 2011 and 2015, adding up to a multivariate sensitivity analysis showed that the estimated cumulative total cost of US$478 million by 2015. Table 1. Size and distribution of population in APOC target areas (thousands and fraction of total). Number of treatment rounds provided through 2010 Onchocerciasis type Endemicity class 1–2 3–5 6–9 10–13 Total Forest/mixed Non-endemic 3 0.0% 129 0.1% 119 0.1% 342 0.3% 593 0. 6% Forest/mixed Hypoendemic 155 0.1% 5,669 5.4% 5,245 5.0% 14,170 13.6% 25,239 24.3% Forest/mixed Mesoendemic 71 0.1% 4,179 4.0% 4,210 4.0% 11,768 11.3% 20,228 19.4% Forest/mixed Hyperendemic 13 0.0% 4,128 4.0% 5,428 5.2% 15,201 14.6% 24,770 23.8% Forest/mixed Total 243 0.2% 14,104 13.6% 15,002 14.4% 41,481 39.9% 70,831 68.1% Savanna Non-endemic 0 0.0% 1 0.0% 1 0.0% 18 0.0% 19 0.0% Savanna Hypoendemic 0 0.0% 871 0.8% 1,048 1.0% 12,837 12.3% 14,756 14.2% Savanna Mesoendemic 0 0.0% 1,695 1.6% 1,143 1.1% 9,402 9.0% 12,240 11.8% Savanna Hyperendemic 0 0.0% 1,900 1.8% 255 0.2% 4,049 3.9% 6,203 6.0% Savanna Total 0 0.0% 4,467 4.3% 2,446 2.4% 26,306 25.3% 33,219 31.9% Grand Total 243 0.2% 18,571 17.8% 17,449 16.8% 67,787 65.1% 104,050 100.0% Populations were stratified by onchocerciasis type, endemicity class and the history of mass treatment. The history of mass treatment is expressed as the number of treatment rounds provided through 2010. doi:10.1371/journal.pntd.0002032.t001 PLOS Neglected Tropical Diseases | www.plosntds.org 4 January 2013 | Volume 7 | Issue 1 | e2032 APOC 1995-2015: Health Impact and Cost Figure 2. Predicted prevalence of onchocercal infection and morbidity in APOC areas from 1995 to 2015. Please note the different scales for the y-axes in the four panels. Shaded areas represent projections for 2011–2015. A) Prevalence of infection is defined as infestation with at least one adult female worm, or alternatively, presence of detectable microfilariae in the skin. B) Prevalence of troublesome itch, caused by onchocerciasis. C) Prevalence of onchocercal visual impairment, defined as corrected visual acuity (i.e. measured with glasses on or through pinhole) of ,18/60 and $3/60 in the better eye. D) Prevalence of onchocercal blindness, defined as corrected visual acuity (i.e. measured with glasses on or through pinhole) of ,3/60 or restriction of visual field to less than 10u in the better eye. doi:10.1371/journal.pntd.0002032.g002 Discussion utes to the disease burden of onchocerciasis and is known to be reduced by ivermectin [10–13]. Further, epilepsy may be We estimated the health impact and cost of mass treatment with associated with onchocerciasis, as suggested by a growing but ivermectin for the 20-year period that APOC is scheduled to run still uncertain base of evidence [14]. However, we chose to as a morbidity control program (1995–2015). Our simulations include only the most important disease manifestations for which suggest that mass treatment with ivermectin has markedly reduced data were available for model calibration (i.e., eye disease and the prevalence of infection with O. volvulus, troublesome itch, visual troublesome itch). Furthermore, we did not include the effect of impairment, and blindness in APOC areas, averting an estimated ivermectin on diseases that are co-endemic with onchocerciasis, 8.2 million DALYs due to onchocerciasis by 2010 at a nominal such as soil transmitted helminthiases, ectoparasitic infections, financial cost of about US$257 million (excluding cost of donated and lymphatic filariasis [15]. Other minor factors leading to an drugs). We expect that APOC will avert another 9.2 million underestimation of the health impact are that we only considered DALYs between 2011 and 2015, at a nominal financial cost of the effect of ivermectin on the capacity of adult female worms to US$221 million. release microfilariae and its microfilaricidal effect, whereas Our estimate of APOC’s health impact only considered eye ivermectin may additionally have a modest effect on adult worm disease and troublesome itch, and would be even higher if other viability [16,17]. Furthermore, we ignored between-village clinical manifestations of onchocerciasis would have been taken variation in coverage, which is perhaps most extreme in the into account. For instance, disfiguring skin disease also contrib- phase of scaling up: in some projects, treatment started in a PLOS Neglected Tropical Diseases | www.plosntds.org 5 January 2013 | Volume 7 | Issue 1 | e2032 APOC 1995-2015: Health Impact and Cost Figure 3. Disability-adjusted life years (DALYs) lost due to onchocerciasis from 1995 to 2015. The total height of the bars (colored plus blank) represents the estimated number of DALYs lost in a counterfactual scenario without ivermectin mass treatment (increasing trend due to population growth). The colored part of each bar represents the estimated actual number of DALYs lost (declining trend due to ivermectin mass treatment). The blank part of each bar therefore represents the annual number of DALYs averted by ivermectin mass treatment in the total APOC population. doi:10.1371/journal.pntd.0002032.g003 subpopulation with high coverage, while the other part of the The validity of our results, as in any simulation study, depends population did not yet receive mass treatment (which is more on the quality of the model and its assumptions. ONCHOSIM was efficient than treating the entire project population at an first developed in the early nineties and has earned trust over the equivalent average coverage). We may have somewhat overesti- years from the large scale control programs. ONCHOSIM has mated the number of life years lost due to excess mortality from been used to successfully mimic observed epidemiological data blindness during and after mass treatment, causing a small from various locations [4,20–22], and has been used for policy underestimation in the number of DALYs averted. This is making in the West-African Onchocerciasis Control Programme because we appointed a fixed number of life years lost to every [7]. Efforts to validate the model continue. We have recently new case of blindness, while regular ivermectin treatment is compared ONCHOSIM predictions to longitudinal data from expected to postpone the onset of blindness to a higher age, Senegal and Gambia [23] and found that model-predicted trends reducing the number of life years lost due to blindness. in mf prevalence during 14 to 16 years of mass treatment were Furthermore, we did not consider a possible association between broadly consistent with the observed trends, although the mf excess mortality and (high) microfilarial load [18,19]. prevalence sometimes seemed to decline slightly faster than There are several factors that may (partly) counterweigh the predicted (unpublished data). Furthermore, our model predictions underestimation of the health impact of APOC described above. for trends in itch were comparable to the reported average trend in Therapeutic coverage may have been over-reported by commu- APOC sentinel areas [13]; after five to six years of mass treatment nity members responsible for the distribution of ivermectin, either at 70–80% coverage, itch prevalence was reported to decline from because of incomplete estimates of the community population or 16% to 7%, and we predicted a decline from 14% to 6.5% for to inflate their own performance. Yet, the estimated health impact areas with similar pre-control levels of infection and history of of APOC by 2015 would decrease by only 0.8 million averted mass treatment. Likewise, our model adequately reproduced DALYs if we assume that coverage were to be systematically 10% trends in onchocercal blindness during vector control in West lower than reported. Also, we ignored any mass treatment prior to Africa (File S1). Although the above suggests that our model the inception of APOC, whereas in reality, ivermectin distribution predictions are realistic, our estimates remain subject to uncer- had already started in a limited number of foci (here morbidity tainty and it would be good to have them confirmed by more field levels had already been reduced somewhat, but not on account of data, especially regarding trends in morbidity during mass APOC). Taking all above sources of under- and over-estimation treatment. into account, we believe that the true health impact of APOC is Even though the model seems to be reliable, we should consider still slightly higher than our calculations. potentially important sources of uncertainty in our analysis. An PLOS Neglected Tropical Diseases | www.plosntds.org 6 January 2013 | Volume 7 | Issue 1 | e2032 APOC 1995-2015: Health Impact and Cost Table 2. Health impact and cost of ivermectin mass treatment, 1995–2015. Health impact in number of DALYs Year averted (millions) Costs for coordination of mass treatment (million US$) National onchocerciasis APOC task forces* Total** 1995 0.00 0.0 0.0 0.0 1996 0.00 2.4 1.1 3.6 1997 0.00 2.4 1.1 3.6 1998 0.03 9.3 4.4 13.7 1999 0.13 9.3 4.4 13.7 2000 0.21 9.2 4.3 13.5 2001 0.29 9.2 4.3 13.5 2002 0.39 9.1 4.3 13.3 2003 0.47 11.3 5.3 16.7 2004 0.58 12.6 5.1 17.8 2005 0.69 13.5 4.0 17.6 2006 0.79 11.0 6.0 17.0 2007 0.92 13.7 7.7 21.4 2008 1.05 13.7 7.5 21.3 2009 1.23 21.2 10.0 31.1 2010 1.41 26.7 12.5 39.2 2011 1.56 40.2 2012 1.70 42.5 2013 1.83 44.4 2014 1.98 46.3 2015 2.13 47.9 Subtotal 1995–2010 8.20 174.8 82.1 256.9 Total 1995–2015 17.39 478.1 The health impact is expressed as the annual number of DALYs averted. Costs include those taken on by the African Programme for Onchocerciasis Control (APOC) and national onchocerciasis task forces (including beneficiary governments and non-governmental development organizations). All costs are expressed in nominal US$ (i.e., uncorrected for inflation and undiscounted), and do not include cost of donated drugs or government salaries. *National onchocerciasis task force expenditures for the years 1995–2003 and 2010 were unknown; they were assumed to be equal to 47% of APOC expenditures, based on known expenditures for the years 2004–2009. **Expenditures for 2011–2015 were estimated based on the expected number of treatments in that period multiplied by the estimated cost per treatment in 2010 ($0.52). doi:10.1371/journal.pntd.0002032.t002 often debated factor concerns the effect of ivermectin on adult impact (625%), the magnitude of the predicted impact was always worms. The univariate sensitivity analysis showed that the large. assumed treatment effects of ivermectin on the capacity of adult With an estimated 8.2 million DALYs averted in a 15-year worms to release microfilariae influenced the estimated health period and a predicted doubling in the subsequent 5 years, the impact only marginally. We did not study the effects of assuming predicted health impact of APOC is impressive. According to our no cumulative effects of ivermectin on worm fecundity, whereas it calculations, mass treatment against onchocerciasis cost about a has been suggested that the latter may be the case [24]. However, nominal US$31 per undiscounted DALY averted between 1995 if we had, ivermectin efficacy parameters would have been and 2010. According to World Health Organization guidelines calibrated such that the model-predicted trends in mf prevalence [25], this is highly cost-effective, as it is below the per capita gross and density were still in agreement with observed trends [4,22], domestic product of most countries covered by APOC (27–1,545 and therefore predicted trends in infection levels and morbidity international dollar per capita; Global Health Observatory Data should not have differed much from the current model’s Repository, accessed 2 August 2012). Furthermore, this cost- predictions. The sensitivity analysis showed that alternative effectiveness is comparable to or even better than those for several assumptions for the effect of ivermectin on itch (the only reversible other public health interventions. For example, the life-time cost- symptom under consideration) also influenced the estimated health effectiveness of prophylaxis against mother-to-child transmission of impact only marginally. The most influential assumptions in our HIV in a resource-limited setting has been estimated at US$52 per analysis were related to the estimated size of the population at risk, undiscounted DALY (incremental cost-effectiveness ratio of World pre-control levels of infection, and the assumed associations Health Organization guidelines versus minimal standard of care) between infection and morbidity, which were all based on data. [26]. The cost-effectiveness of large-scale, long-term (30-year Even though the multivariate sensitivity analysis suggested period) public health interventions targeting other neglected considerable overall uncertainty in our estimate of the health tropical diseases has been estimated at US$4–US$29 per DALY PLOS Neglected Tropical Diseases | www.plosntds.org 7 January 2013 | Volume 7 | Issue 1 | e2032 APOC 1995-2015: Health Impact and Cost Figure 4. Sensitivity analysis for the estimated cumulative number of DALYs averted by 2015. The multivariate sensitivity analysis (last item) consisted of 200 repeated analyses, based on 200 sets of random parameter values, which were drawn from triangular distributions with modes equal to parameter values used in the main analysis, and minimum and maximum values equal to parameter values used in the univariate sensitivity analysis (first eight items of this figure). The results of the multivariate sensitivity analysis are expressed as the 2.5 and 97.5 percentiles of results from 200 repeated analyses. doi:10.1371/journal.pntd.0002032.g004 (mass drug administration against lymphatic filariasis), US$38 per predictions of health impact, these figures suggest that by 2015, DALY (case detection and treatment for leprosy), US$260 per APOC will have averted a staggering US$2.2 billion due to DALY (vector control against Chagas disease), and US$48– productivity losses from blindness (US$517 million) and itch US$303 (vector control against lymphatic filariasis) [27]. Mass (US$1.7 billion, assuming productivity losses in 25% of people treatment against onchocerciasis is of even better value (US$27 per with itch). In other words, beneficiary countries should expect DALY) if expected health gains and costs for the period 2011– economic benefit from mass treatment that outweighs any costs. 2015 are included. In view of the anticipated elimination of Clearly, all of the above calculations apply only under the infection so that mass treatment can be stopped altogether, the condition that countries do not themselves pay for the drug cost-effectiveness will be even better than our calculations suggest ivermectin. The amount of ivermectin donated up to 2010 [23]. represents a value of US$2.1 billion, assuming 2.8 tablets per The objective of APOC is to establish country-led systems for treatment and a commercial price per tablet of US$1.50 plus onchocerciasis control by 2015, which means that countries and US$0.005 shipping costs (personal communication with Dr. A. their partners will have to carry full financial responsibility by that Hopkins, director of the Mectizan Donation Program). This year. Our results indicate that cost per treatment with ivermectin amount is eight times the program costs for coordinating mass in APOC areas is affordable (US$0.51 per treatment, excluding treatment. Likewise, for the period 2011–2015, the value of cost of donated drugs) and comparable to the costs of existing donated ivermectin will be an additional US$1.8 billion. There- national mass treatment programs for the elimination of lymphatic fore, mass treatment with ivermectin can be sustained only with filariasis (US$0.06–US$2.23 per treatment) [28]. Mass treatment donation of ivermectin, which Merck has pledged to continue for with ivermectin, however, also involves costs for society not as long as necessary. covered by the program. From published data for two Nigerian We expect that levels of infection in the APOC target area will communities, we derived that these costs are about US$0.23 per have fallen drastically by 2015 (overall prevalence of adult female treatment (excluding start-up costs) [29]. Based on this estimate, worms 18%). The implication is that by that time, transmission of the sum of program and community costs for mass treatment with infection may be almost interrupted in areas with favorable ivermectin was approximately US$370 million from 1995 to 2010 conditions for elimination, such as high coverage of mass and will be another US$320 million for 2011–2015. In addition to treatment, sufficient treatment rounds, and/or low to medium costs, there are significant benefits for society that countries need pre-control levels of infection [31]. Until recently, elimination of to take into account, such as prevented productivity losses resulting onchocerciasis from Africa was thought to be impossible by means from blindness and itch. Blindness in rural Africa has previously of mass treatment alone, considering the large size of the been assumed to result in an annual productivity loss of US$150 transmission zones, mobility of the vectors and human popula- per case [30]. Likewise, the productivity loss due to itch among tions, and poor compliance with mass treatment [32]. Following coffee plantation workers in an Ethiopian site has been estimated reports of elimination of onchocerciasis from foci in Mali and at around US$5.32 per month per case [12]. Combined with our Senegal by mass treatment alone [23], however, interest has PLOS Neglected Tropical Diseases | www.plosntds.org 8 January 2013 | Volume 7 | Issue 1 | e2032 APOC 1995-2015: Health Impact and Cost renewed in elimination of onchocerciasis from Africa [33]. be read when referred to from the corresponding methods section Following this, WHO has recently been advised to extend APOC of the main article (i.e. the supplement is not intended to be read mandate by ten years to 2025 with the new aim of eliminating as a stand-alone document). infection with O. volvulus, where possible. With this new (PDF) motivation, we may indeed expect focal elimination of infection, resulting in even more health gains from mass treatment with Acknowledgments ivermectin in the future and the possibility of being able to end mass treatment altogether. We thank Drs. M.C. Asuzu, M. Hagan, W.H. Makunde, P. Ngoumou According to our simulations, APOC has had a remarkable (deceased), K.F. Ogbuagu, D. Okello, G. Ozoh and J.H.F. Remme for their contributions to pre-control data on nodule prevalence and impact on population health in Africa between 1995 and 2010. prevalence of itch. We thank Dr. K.Y. Dadzie for contributing data on This health impact is expected to double during the subsequent trends in blindness during vector control. We thank Dr. A.D. Hopkins, five years. Further, APOC is a highly cost-effective public health director of the Mectizan Donation Program, for providing estimates of cost programs, and given the anticipated elimination of onchocerciasis per tablet ivermectin and average number of tablets needed per ivermectin from APOC areas, we expect even more health gains and a more treatment. profitable cost-effectiveness of mass treatment with ivermectin in the near future. Our study fully supports the advice to continue Author Contributions APOC activities for another ten years. Conceived and designed the experiments: LEC WAS JLV DH SJdV. Performed the experiments: LEC. Analyzed the data: LEC. Contributed Supporting Information reagents/materials/analysis tools: HGMZ KBA MEM MN GF UVA DAPB. Wrote the paper: LEC WAS JLV DH SJdV HGMZ KBA MEM File S1 This document describes the details of the methods that MN GF UVA DAPB JHR. were used in this study, both in text and figures. This supplement is divided into appropriately named sections, which are intended to References 1. Duke BO (1993) The population dynamics of Onchocerca volvulus in the 18. Little MP, Breitling LP, Basa ˜ ez MG, Alley ES, Boatin BA (2004) Association ´n human host. Trop Med Parasitol 44: 61–68. between microfilarial load and excess mortality in onchocerciasis: an 2. Boatin B (2008) The Onchocerciasis Control Programme in West Africa (OCP). epidemiological study. Lancet 363: 1514–1521. Ann Trop Med Parasitol 102 Suppl 1: 13–17. 19. Walker M, Little MP, Wagner KS, Soumbey-Alley EW, Boatin BA, et al. (2012) 3. Noma M, Nwoke BE, Nutall I, Tambala PA, Enyong P, et al. (2002) Rapid Density-dependent mortality of the human host in onchocerciasis: relationships epidemiological mapping of onchocerciasis (REMO): its application by the between microfilarial load and excess mortality. PLoS Negl Trop Dis 6: e1578. African Programme for Onchocerciasis Control (APOC). Ann Trop Med 20. Plaisier AP, van Oortmarssen GJ, Remme J, Alley ES, Habbema JD (1991) The Parasitol 96 Suppl 1: S29–39. risk and dynamics of onchocerciasis recrudescence after cessation of vector 4. Plaisier AP, Alley ES, Boatin BA, Van Oortmarssen GJ, Remme H, et al. (1995) control. Bull World Health Organ 69: 169–178. Irreversible effects of ivermectin on adult parasites in onchocerciasis patients in 21. Plaisier AP, van Oortmarssen GJ, Remme J, Habbema JD (1991) The the Onchocerciasis Control Programme in West Africa. J Infect Dis 172: 204– reproductive lifespan of Onchocerca volvulus in West African savanna. Acta 210. Trop 48: 271–284. 5. Basa´n˜ ez MG, Pion SD, Boakes E, Filipe JA, Churcher TS, et al. (2008) Effect of 22. Alley ES, Plaisier AP, Boatin BA, Dadzie KY, Remme J, et al. (1994) The single-dose ivermectin on Onchocerca volvulus: a systematic review and meta- impact of five years of annual ivermectin treatment on skin microfilarial loads in analysis. Lancet Infect Dis 8: 310–322. the onchocerciasis focus of Asubende, Ghana. Trans R Soc Trop Med Hyg 88: 6. Plaisier AP, van Oortmarssen GJ, Habbema JD, Remme J, Alley ES (1990) 581–584. ONCHOSIM: a model and computer simulation program for the transmission 23. Diawara L, Traore MO, Badji A, Bissan Y, Doumbia K, et al. (2009) Feasibility and control of onchocerciasis. Comput Methods Programs Biomed 31: 43–56. of onchocerciasis elimination with ivermectin treatment in endemic foci in 7. Plaisier AP, Alley ES, van Oortmarssen GJ, Boatin BA, Habbema JD (1997) Africa: first evidence from studies in Mali and Senegal. PLoS Negl Trop Dis 3: Required duration of combined annual ivermectin treatment and vector control e497. in the Onchocerciasis Control Programme in west Africa. Bull World Health 24. Bottomley C, Isham V, Collins RC, Basa ˜ ez MG (2008) Rates of microfilarial ´n Organ 75: 237–245. production by Onchocerca volvulus are not cumulatively reduced by multiple ivermectin treatments. Parasitology: 1–11. 8. Habbema JDF, Oostmarssen GJ, Plaisier AP (1996) The ONCHOSIM model 25. WHO Commission on Macroeconomics and Health (2001) Macroeconomics and its use in decision support for river blindness control. In: Isham V, Medley and health: investing in health for economic development. Report of the G, editors. Models for infectious human diseases - their stucture and relation to Commision on Macroeconomics and Health. Geneva: World Health Organi- data. Cambridge: Cambridge Universitry Press. pp. 360–380. zation. 9. World Health Organization (2004) Global Burden of Disease update 2004: 26. Shah M, Johns B, Abimiku A, Walker DG (2011) Cost-effectiveness of new disability weights for diseases and conditions. Geneva: World Health WHO recommendations for prevention of mother-to-child transmission of HIV Organization. in a resource-limited setting. AIDS 25: 1093–1102. 10. Murdoch ME, Asuzu MC, Hagan M, Makunde WH, Ngoumou P, et al. (2002) 27. Remme JHF, Feenstra P, Lever PR, Me ´dici A, Morel C, et al. (2006) Tropical Onchocerciasis: the clinical and epidemiological burden of skin disease in Africa. Diseases Targeted for Elimination: Chagas Disease, Lymphatic Filariasis, Ann Trop Med Parasitol 96: 283–296. Onchocerciasis, and Leprosy. In: Jamison DT, Breman JG, Measham AR, 11. Alonso LM, Murdoch ME, Jofre-Bonet M (2009) Psycho-social and economical Alleyne G, Claeson M et al., editors. Disease Control Priorities in Developing evaluation of onchocerciasis: a literature review. Soc Med 4: 8–31. Countries (2nd Ed). Washington DC: The World Bank/Oxford University Press. 12. Kim A (1997) Health and Labor Productivity: the economic impact of 28. Goldman AS, Guisinger VH, Aikins M, Amarillo ML, Belizario VY, et al. (2007) onchocercal skin disease. The World Bank. National mass drug administration costs for lymphatic filariasis elimination. 13. Ozoh GA, Murdoch ME, Bissek AC, Hagan M, Ogbuagu K, et al. (2011) The PLoS Negl Trop Dis 1: e67. African Programme for Onchocerciasis Control: impact on onchocercal skin 29. Onwujekwe O, Chima R, Shu E, Okonkwo P (2002) Community-directed disease. Trop Med Int Health 16: 875–883. treatment with ivermectin in two Nigerian communities: an analysis of first year 14. Pion SD, Kaiser C, Boutros-Toni F, Cournil A, Taylor MM, et al. (2009) start-up processes, costs and consequences. Health Policy 62: 31–51. Epilepsy in onchocerciasis endemic areas: systematic review and meta-analysis of 30. Benton B (1998) Economic impact of onchocerciasis control through the African population-based surveys. PLoS Negl Trop Dis 3: e461. Programme for Onchocerciasis Control: an overview. Ann Trop Med Parasitol 15. Tielsch JM, Beeche A (2004) Impact of ivermectin on illness and disability 92 Suppl 1: S33–39. associated with onchocerciasis. Trop Med Int Health 9: A45–56. 31. Winnen M, Plaisier AP, Alley ES, Nagelkerke NJ, van Oortmarssen G, et al. 16. Awadzi K, Attah SK, Addy ET, Opoku NO, Quartey BT (1999) The effects of (2002) Can ivermectin mass treatments eliminate onchocerciasis in Africa? Bull high-dose ivermectin regimens on Onchocerca volvulus in onchocerciasis World Health Organ 80: 384–391. patients. Trans R Soc Trop Med Hyg 93: 189–194. 32. Dadzie Y, Neira M, Hopkins D (2003) Final report of the Conference on the 17. Gardon J, Boussinesq M, Kamgno J, Gardon-Wendel N, Demanga N, et al. eradicability of Onchocerciasis. Filaria J 2: 2. (2002) Effects of standard and high doses of ivermectin on adult worms of 33. Mackenzie CD, Homeida MM, Hopkins AD, Lawrence JC (2012) Elimination Onchocerca volvulus: a randomised controlled trial. Lancet 360: 203–210. of onchocerciasis from Africa: possible? Trends Parasitol 28: 16–22. PLOS Neglected Tropical Diseases | www.plosntds.org 9 January 2013 | Volume 7 | Issue 1 | e2032