Onchocerciasis Control Programme in V/est Africa Programme de lutte confte l'onchocercose en Afrique de l'Ouest African Programme for Onchocerciasis Control Programme africain de lutte contre I'onchocercose JOINT ACTION FORUM Office of the President FORUM D'ACTION COMM{'NE Bureau du Président JOINT PROGRAMME COMMITTEE Office of the Chairman COMITE CONJOINT DU PROGRAMME Bureau du Président JPC - JAF CCP. FAC JOINT SESSION / SESSION CONJOINTE The Hague/La Haye, 8 Dec. 1999 JPC2O-JAF5/INF/DOC.2 AN ECONOMIC EVALUATION OF INTEGRATING LYMPHATIC FILARIASIS ELIMINATION INTO APOC Preliminary Report
AI\ ECONOMIC EVALUATION OF INTEGRATING LYMPHATIC FILARIASIS ELIMINATION INTO APOC Preliminary Report * Please send correspondence to: Dr. Anne Haddix Deparünent of International Health Rollins School of Public Health of Emory University 1518 Clifton Road Atlanta, GA30322 USA 404-727-3558 (voice) 404-727-4590 (fax) achaddi@sph.emory.edu * Appendix alrailable upon request
EVALUATION ECONOMIQUE DE L'INTEGRATION DES ACTIVITES D'ERADICATION DE LA FILARIOSE LYMPHATIQUE DANS L'APOC RESUME ANALYTIQUE Introduction : La filariose lymphatique (FL) est une maladie transmise par des moustiques. Elle touche 120 millions de personnes dans le monde, dont près de 40 milüons présentent des symptômes. Après une période d'infection prolongée, le système lymphatique est atteint, la personne infectée souffre d'une lynphadenite aiguë et présente des symptômes chroniques qui revêtent le plus souvent la forme d'un lymphoedème des extrémités inférieures et d'une hydrocèle. Les personnes ayant des symptômes chroniques ont un niveau d'activité limité, et celles qui souffrent d'une lynphadenite aiguë sont souvent totalement incapables d'accomplir la moindre tâche pendant leurs attaques qui peuvent durer de trois à cinq jours. Bien que la FL ne soit pas mortelle, les personnes atteintes présentent souvent des symptômes chroniques pendant leurs années les plus productives, ce qui impose une lourde charge économique à la société. Elles subissent, par ailleurs, une perte considérable sur le plan social (elles ne peuvent se marier, ou perdent leur conjoint, et ne peuvent plus avoir d'enfants) et vivent dans la honte. Selon les estimations, un tiers des personnes ayant contracté la FL vivent en Afrique et près de 500 millions d'Africains courent le risque de contracter cette maladie. Objectif : La présente étude a pour objet d'examiner les coûts et les avantages de l'intégration de l'éradication de la FL dans les activités du Programme africain de lutte contre l'onchocercose (APOC). Elle analyse aussi les coûts et les avantages des prograûrmes d'éradication de la FL dans l'aire du Programme de lutte conEe l'onchocercose (OCP). Pour chaque région, elle indique la charge économique actuelle de la FL, les valeurs actualisées des coûts et des avantages du programme données par les projections sur 30 ans, les taux de rentabilité économique (TRE) et les seuils de rentabilité. Elle indique aussi les coûts (économies) du(e)s à l'intégration de l'éradication de la FL dans les activités actuelles de I'APOC. Population exposée à un risque : Selon les estimations d'un modèle epidémiologque basé sur des études antérieures de la FL en Afrique subsaharienne, le nombre de personnes susceptibles de contracter cette maladie est de 251 millions dans les 19 pays couverts par I'APOC, et de 73 millions dans les l1 pays couverts par I'OCP. Nous avons posé en hlpothèse que 50 o delapopulation vivant dans des régions où l'onchocercose est endémique courent aussi un risque. Étaot donné l'incertitude dont est entachée cette estimation, nous avons également considéré differents taux compris entre 0 et 100 oÂ.La prévalence de la FL varie fortement dans la région. La présente analyse sr.rypose un taux d'incidence annuel de lymphadenite aiguë de 33 episodes pour I 000 personnes, un taux de prévalence d'hydrocèle de 75 pour I 000 et un taux de prévalence de lyrrphodème de l0pour I 000. Programme d'éradication de la FL : Le programme examiné ici domre lieu à une campagne de traitement de masse pendant une période de cinq ans. Un cachet de
Mectizan et un cachet d'Albendazole sont administrés une fois par an pendant cinq ans dans le cadre de programmes de traitement communautaires à toutes les personnes vivant dans des "o.-*"uiés infectées à l'exclusion des femmes enceintes et des enfants de moins de cinq ans. Le projet consiste, au départ, à évaluer rapidement la situation pour identifier les iégions diendémie, mais le prograrnme prévoit aussi la poursuite d'un suivi pendant cinq ans après la fin du traitement pour s'assurer que le cycle de transmission a Lien été interrompu. Le coût du programme englobe celui de [a cartographie epidémiolo$que;les frais généràux de I'APOC,,le coût des programmes par pays et celui des activités de surveillance post-traitement. Les deux médicaments étânt fournis gratuitement par leurs fabricants, leur coût n'est pas pris en compte. Nous avons supposé [ue 10 % deiapopulation << courant un risque » était traitée la première année, lO %o de ptus ta deuxièmi arurée et ainsi de suite jusqu'à ce que la population tout entière soit touverte au bout de dix ans. Nous avons également supposé que 80 o delapopulation susceptible d' être soi gnée l' était effectivement. Efficacité du programme : L'efficacité du programme est uniquement mesurée par la diminution du nombre d'épisodes de tymphadenite aiguë et de nouveaux cas d'hydrocèle et de lymphoedème. Selon les estimations effectuées au moyen du modèle epidémioiogrque, le nombre d'épisodes de lymphadenite aiguë diminue de l0 Yo par an à partir de la deuxième arurée de traitement et la prévalence d'hydrocèle et de iymphoedème baisse de 5 oÂpar an à partir de la sixième arurée pour remonter à 10 oÂpar an après la dixièrire année. Avantages : L'analyse ne prend en compte que la perte de productivité engendrée par des attaqueJde lymphaàenite aiguë et la morbidité associée à l'hydrocèle et la lymphoedàne. Elle n'inclut pas la baisse de productivité due à d'autres effets éventuels de la FL, tels qu'une mortalité précoce par suite d'une dénutrition associée à une chylurie, à une insuffisance rénale chronique et à des helminthes instestinaux. Le calcul des pertes de productivité annuelles est basé sur la médiane du PIB par habitant des pays couverts par I'APOC et I'OCP. Taux dtactualisation et monnaie : L'analyse du programme de FL repose sur un tau:t social d'actualisation de 3 Yo etun coût d'opportunité du capital de 10 %o.I*s resultats sont, dans les deux cas, présentés en dollars de 1998. Résultat : Étant donné les hlpothèses retenues aux fins de I'analyse, la charge économique annuelle actuelle de la FL mesurée en termes de productivité perdue est de l'ordre de 1,3 milliard de dollars dans les pays couverts par I'APOC. La mise en oeuwe d'un prograûtme de FL coûterait 524 millions de dollars (chiffre non actualisé) sur une période de 19 ans, et ce prograrnme coûlmencerait à faire sentir ses effets dès la àeuxième année. La VAN du programme calculé pour des taux d'actualisæion annuels de 3 Yo etde 10 o/o sur 30 ans s'établit à, respectivement,4,3 milliards et 1 milliard de dollars. Dans ces mêmes conditions, le TRE est de 25 % à la fin de la période d'investissement (2019) et de 28 % 30 ans après le lancement du programme. Le seuil de rentabilité est atteint la dixième année. Dans l'aire de I'APOC, le progranrme bénéficie de I'existence d'une structure administrative et du chevaucheme,lrt des zones dans tt
lesquelles I'onchocercose et la FL sont endémiques. L'intégration des activités d'éradication de la FL dans les activités actuelles de I'APOC permettra de réduire de 45 millions de dollars le coût du programme (valeurs non actualisées) pour la FL. Dans l'aire de I'OCP, les pertes aruruelles de productivité dues à la FL se chiffrent à près de 390 millions de dollars. Un programme similaire à celui décrit dans le cas de I'APOC coûterait 169 millions de dollars. La VAN, moyennant des taux d'actualisation annuels de3 o et de 10 o/o sur 30 ans serait de, respectivement, 1,2 milliard de dollars et 302 millions de dollars, tandis que le TRE attein&attz3 o/o àla fin de la période d'investissement (2019) et26 o 30 ans après le lancement du programme. Analyse de sensibilité : Le coût total du programme dépend dans une tès large mesure de la taille de la population exposée à un risque, qui est la variable du modèle dont la valeur est la plus entachée d'incertitude. L'analyse dépend également des degrés d'incapacité, du taux d'actualisation et de la valeur de la productivité. Les résultats ne se modifient toutefois pas nettement lorsque ces paramètres fluctuent dans une mesure raisonnable. Dans presque tous les scénarios, le programme atteint son seuil de rentabilité la dixième ou la onzième année. Conclusion : I1 serait avisé, sur le plan financier et humain, d'intégrer les activités d'éradication de la FL dans les activités actuelles de I'APOC. Les investissements consacrés à l'éradication de la FL donnent des résultats satisfaisants par rapport à d'autres investissements dans le secteur de la santé ou dans d'autres secteurs. Par ailleurs, ils permettraient d'améliorer les conditions de vie de millions d'Africains pendant plusieurs générations. Les résultats de la présente analyse sous-estiment la réalité car ils sont basés sur des évaluations très prudentes. Nous avons retenu des taux de diminution du nombre de cas de morbidité assez faibles et n'avons pas pris en compte un certain nombre des avantages des interventions, tels que l'impact de l'association de deux médicaments sur les helminthes intestinaux faute de données suffisantes. Nous n'avons pas non plus pris en considération l'impact de la FL sur les systèmes de santé des pays d'endémie. L'établissement au plus tôt d'une carte couwant l'aire de I'APOC permettra de préciser l'analyse. L'identification rapide et le ciblage des zones qui seront couvertes par les programmes d'éradication de la FL dans les régions où I'onchocercose est endémique et celle où la FL est prévalente pourront contribuer à accroître les TRE. Lll
AN ECONOMIC EVALUATION OF INTEGRATING LYMPHATIC F'ILARIASIS ELIMINATION INTO APOC EXECUTIVE SUMMARY Introduction: Lymphatic filariasis (LF) is a mosquito-borne disease that affects 120 million people worldwide. Approximately 40 million people show visible signs of the disease. Over a prolonged period of infection, damage to the lymphatic system produces both acute lymphadenitis (ADL) and chronic symptoms of which lymphedema of the lower extremities and - hydrocele are the most common conditions. Individuals with chronic symptoms are severely limited in their activity level, while ADL suffers are often completely incapacitated for 3-5 days during episodes. Although LF is not fatal, the chronic symptoms often afflict individuals during their most productive years and therefore impose a significant economic burden on socieÿ. Additionally, those who suffer from the disease experience a tremendous social loss (lost or broken marriages, lost parenthood opportunities) and personal shame. It is estimated that one- third of persons affected by LF live in Africa and that almost one-half billion Africans are at risk ofthe disease. Objective: The objective of this study is to examine the costs and benefits of integrating LF elimination into the activities of the African Program for Onchocerciasis Control (APOC). We have also examined the costs and benefits of LF elimination programs in the countries where the Onchocerciasis Control Program (OCP) has been active. We report the current economic burden ofLF in each region, the predicted present values ofprogram costs and benefits over a 30-year period, the economic rates of return (ERRs), and the break-even points. We also report the shared costs (savings) associated with integrating LF elimination into APOC's 'current programs. Population At Risk: Using epidemiological modeling based on studies that have previously reported LF in sub-Saharan Africa, we have estimated that there are 251million people at risk of LF in the 19 APOC countries. Another 73 million people are at risk of LF in the I I OCP countries. We have assumed that 50Yo of the population in the onchocerciasis-endemic areas are also at risk of LF. Because of the uncertainÿ of the estimate, we have varied this from 0-100%. The prevalence of LF varies greatly across the region. In this analysis, we use an annual ADL incidence of33 episodes per 1000 persons, a hydrocele prevalence of 75 per 1000, and a lymphedema prevalence of 10 per 1000. LF Elimination Program: The program analyzed in this study is a S-year mass treatment program. Everyone in infected communities is üeated once a year for 5 years with a single dose each of Mectizan and albendazole. The mode of delivery is community-directed treatment. Pregnant women and children under the age of 5 are excluded. The first step of the project is to conduct rapid assessment to identiÿ the LF-endemic areas. The program also includes 5 years of post-treatment surveillance to assure against a recurrence of transmission. Program costs include those for mapping, APOC overhead, country programs, and post-treatment surveillance. Because the two drugs are donated, their cost is excluded. Iüe have assumed that 10% of the "at risk" population has access to treatment in the first year and that an additional l0oÀ are added each year until 100% coverage is achieved in the tenth year. We further assume that 80% of the population with access to treatment actually receives treatment. Program Effectiveness: Effectiveness measures are limited to the decline in ADL episodes and new cases of hydrocele and lymphedema. Using epidemiological modeling, we estimated that ADL episodes decline lÙYo per year beginning in the second year of treatment; and the prevalence hydrocele and lymphedema declines 5Yo per year beginning in the sixth year and increasing to I.0% pei year after the tenth year. lv Benefits: Benefits in this analysis are limited to the loss in productivity associated with incapacitation from ADL episodes and disability associated with hydrocele and lymphedema. We have not included-lost productivity associated with other possible manifestations of LF such as early mortality from denutrition caused by chyluria and from chronic renal failure and intestinal helminthes infections. Annual productivity losses are valued using the median per capita GDP for the APOC and OCP countries. Discount Rate and Currency: We have analyzedthe LF program using both a social discount rate of 30 and a capitalrate of l0%. We report the results of both in 1998 US dollars. Results: Under the assumptions made in this analysis, the current annual economic burden of LF measured in lost productivity is approximately $1.3 billion in the APOC countries. Implementing an LF program would cost $524 million (not discounted) over l9 years. Benefits would begin accumulating in the second year. The NPVs for the program using 30% and l0%o annual discount- - rates over a 3O-year period are $4.3 billion and $l billion respectively. Under the same conditions, the ERRs are25o at the end of the investment period in2019 and2SYo 30 years after the initiation of the program. The program breaks even in the lOth year. In the APOC countries, the program takes advantage of the existing administrative structure and the overlap in onchocerciasis- and LF-endemic areas. lntegrating LF elimination into existing APOC activities will save $45 million in program costs (not discounted) for LF. In OCP countries, annual productiviÿ losses from LF approach $390 million. An LF program similar to that described for APOC would cost $ 169 million. The NPVs for the program using 3oZ and l0%o annual discount rates over a 30-year period are $1.2 billion and $302 million respectively and the ERRs are 23Yo atthe end of the investment period in 2019 and26Yo 30 years after the initiation of the program. Sensitivity Analysis: Total program costs are most sensitive to the size of the at risk population, the variable for which there is the most uncertainÿ in the model. The analysis is also sensitive to disability levels, discount rate, and value of productiviÿ. However, reasonable variation in these parameters does not produce substantially different results. Under almost all scenarios, the program breaks even in either the lOth or the I lth year. Conclusions: Integrating LF elimination into current APOC activities is a sound financial and human investment. Investment in LF elimination compares favorable with other health and nonhealth investments. Additionally, such an investment will improve the lives of millions of people in Africa for generations to come. The results of this analysis are understated because of the conservative approach taken. We have assumed conservative rates of disease decline and have not included many of the benefits-including the affect to the two-drug combination on intestinal helminthes-because of lack of data. We have also not included the impact of LF on the health care systems of the endemic countries. Early mapping of APOC countries will allow the refinement of this analysis. Initially identiÿing and targeting LF elimination programs in the onchocerciasis-endemic areas and the areas with high LF prevalence can improve the ERRs. INTRODUCTION Lymphatic filariasis (LF) is a mosquito-borne disease, which affects 120 million people worldwide. Approximately 40 million people show visible signs of the disease (WHO 1994). Over aprolonged period of infection, damage to the lymphatic system produces both acute lymphadenitis (ADL) and chronic symptoms of which lymphedema_ of the lower extremities and hydrocele are the most cortmon conditions. Individuals with chronic symptoms are severely limited in their activity level, while ADL suffers are often completely incapacitated for 3-5 days during episodes. Although LF is not fatal, the chronic symptoms often afflict individuals during their most productive years and, therefore, impose a significant economic burden on socieÿ. Additionally, those who suffer from the disease experience a tremendous social loss (lost or broken marriages, lost parenthood opportunities) and personal shame. It is estimated that one-third of persons affected by LF live in Africa and that almost one-half billion Africans are at risk of the disease. The objective of this study is to examine the costs and benefits of integrating LF elimination into the activities of the African Program for Onchocerciasis Control (APOC). V/e have also examined the costs and benef,rts of LF elimination programs in the countries where the Onchocerciasis Control Program (OCP) has been active. Vy'e report the current economic burden of LF in each region, the predicted program costs and benefits over a 30-year period, the economic rates of return (ERRs), the net present values §PVs), and the break-even points. We also report the shared costs (savings) associated with integrating LF elimination into ApoC,s current programs. BACKGROUND The disability, pain, and suffering caused by lymphatic filariasis impose a considerable social and economic burden on individuals and, ultimately, society. In this section, we will review the socio-economic impact of acute and chronic conditions of Iymphatic filariasis. Although this review focuses on Africa, we also examine the economic burden literature from other regions, primarily India, as relevant literature is scant. As these studies show, there is sufficient data to demonstrate that in highly endemic areas, lymphatic filariasis can have a devastating impact on affected individuals and, subsequently, communities. The acute manifestations of LF are characterizedby episodic attacks of ADL associated with fever and malaise with the presence of localized pain, tenderness and warmth (WHO 1992). In men, ADL may be localized in the genitals. Although well- known terminology exists for the acute clinical manifestations, the etiology is not well understood (Gyapong, and others 1996a) (Gasarasi, and others 1,999). Data suggest that bacterial infections, immune responses to parasite antigen, and repeated infections with filarial infective larvae are possible etiological agents (Gasarasi, and others 1999). An ADL sufferer may experience one or more episodes per year, which can last from2 to 2l days (mean is 8.6 days). The mean duration of total incapacitation is 3.7 days. ADL episodes occur in both men and women although are more frequent in men. The risk of having an ADL episode increases with age (Gasarsi). The highest incidences of acute ADL attacks and disability from chronic disease peaked during the rainy and agricultural season, thereby increasing the productivity losses. (Gyap_ong, and others 1996b) These conditions can severely impact the draining in the retroperitoneal and subdiaphragmatic areas. The risk of hydrocele in the LF endemic areas increases markedly with age. [n some areas, as many as 60 - 80% of older men suffer from hydrocele. Pain and discomfort from swelling of the scrotal sac and recurrent bacterial infections cause significant life-long disability in hydrocele sufferers. The second most common debilitating chronic condition is lymphedema or elephatiasis. In Africa, lymphedema manifests as a swelling of the lower limbs and arms in both men and women. Recurrent episodes of limb lymphedema are the result of lymphatic blockages. Lymphedema becomes progressively worse over time with pitting oedema, loss of skin elasticity and fibrosis. Secondary bacterial and fungal skin infections are common, particularly in people who do not have access to adequate hygiene and do not wear footwear. Repeated episodes of lymphedema and infection results in progressive disability of suf[erers. ADL episodes, hydrocele, and lymphedema reduce the labor productivity of infected individuals. In the Tanzanian study, those experiencing an ADL episode reported 1.4 hours of work per day, compared to 5.0 hours of work reported by the controls, those who experienced other illnesses during that period. (Gasarasi, and others 1999) InGhanqanimpact of l.5Yo femalelaborinputwasassumedlost,whileT.ÙYoof the male labor was lost. Most affected individuals felt that they were an economic burden to the rest of their family. Where possible, they worked in short bursts on the farms or switched to sedentary occupations. (Gyapong, and others 1996b) Very few of the 19.3o of men with hydrocele could afford an operation, which cost $30.00. The individual, virtually bringing all economic activiÿ to a standstill. Although the Gyapongs estimated a0.790 loss in community labor inputs due to ADL episodes in the Ghanaian study , the impact is proportionally greater because the attacks took place during the rainy season. In other areas, the economic burden was exacerbated when alternatives to agricultural work, e.g., weaving, were not available. For example, a study in India found a significant loss of 1.36 hours per day in male farmers as compared to non-farmers (1. 10 hours).(Ramaiah 1 998b) Although the range and mean duration of ADL episodes varies across studies, the period of incapacitation is fairly constant. In the Tanzanian study, length of ADL episodes ranged from2-21days; the mean was 8.6 days(Gasarasi, and others 1999). The mean length of an ADL episode in Ghana was 5.1 days(Gyapong, and others 1996a). Investigators in India have reported mean durations of 3.39 days (Ramaiah 1998a), 3.9 days and 4.9 days (Sabesan, and others 1992). In each episode there was a period of time where the affected individual was unable to work and may have required assistance with their own physical care. In the Ghana study, the individuals were physically unable to participate in economic and domestic activities for an average of 3.0 (Gyapong, and others 1996a). Gasarasi reported an average of 3.7 days of incapacitation in Tanzania, ranging from 1-14 days. Ramaiah's study was also consistent with an average of 3.58 days. The chronic manifestations of LF considered in this analysis are hydrocele and lymphedema. Hydrocele is the most common chronic condition. It manifests clinically as a swelling of the perineal lining that surrounds each of the testicles. Usually, hydrocele fluid accumulates in this closed sac as a result of blockage in the lymphatics mean daily wage in the nearest big town was $1.10. So, given this assumption, it would take an average 6f 33 days for one individual to pay for an operation. Other studies showed significant impacts on productivity losses. The Ghanaian study estimates that chronic conditions result in disability ranging from2}oÂto 600Â. (Gyapong, and others 1996a). In the Indian studies, Dr. Ramaiah reported a loss o13.74 hours of work per day and an 0.08% loss in labor supply in India. Ramaiah assumed thaf at a loss of .74 hours a day, 52-65 productive days would be lost.(Ramaiah, and others 1998; Ramaiah 1998a). Eighty-five percent of those with ADL said that they felt their condition impeded their economic activity. (Ramaiah, and others 1997) An unpublished study from Malaysia, Kessel (1957) reports that a group of 600 rubber-tappers \ryas affected by 150 ADL episodes over a year (incidence 250 per 1000), causing a loss of 450 work days. Furthernore, Giglioli (1960) examined workmen's compensation claims on an agricultural estate in (then) British Guiana, in which hydrocele exacerbation and inflammation were attributed to work-related trauma or strain. Disabiliÿ ranged from four to 60 days and constituted a significant expense to the estate. METHODS In this study, we used cost-benefit analysis to compare the costs and benefits associated with integrating LF elimination programs in to APOC. The analysis is done from the perspective of APOC, in that all of the costs to APOC, regardless of the source of funds, are included. The costs of the two drugs used for LF elimination are not included. The drugs, ivermectin and albendazo\e, are donated by Merck and Company and SmithKline Beecham, respectively. Benefits are measured as increases in productivity associated with decreases in the prevalence of the debilitating clinical manifestations of LF. We examined the costs and benefits for two different time horizons. First, we report the results for the investment period of 2000 to 2019. Next, we extended the time horizon to 2019 to capture benefits from treatment during the investnent period. We report the decrease in the prevalence of clinical LF conditions associated with - the implemetation of the LF elimination program, the annual economic burden imposed by LF in the APOC countries in the absence of a program, and the undiscounted costs and benefits from a program. V/e also report two summary measures: the discounted net present values §PVs)t and the economic rates of return (ERR)2 for the two time horizons. Finally, we report the savings associated with integrating LF elimination into the existing APOC structure compared with beginning a new program. We report results discounted at an annual rate of 3%o, a societal discotrnt rate, and again at l0oÂ, a rate that reflects the opportunity cost of capital. All monetary units are expressed in 1999 US dollars. We have not adjusted any costs or benefits to anticipate future inflation. V/e repeat this analysis for the OCP countries. Population at Risk for Lymphatic Filariasis The population at risk of lymphatic filariasis infection is defined as all of the residents of geographic areas with active lymphatic filariasis transmission. The ' The NPV is the sum that results when the discounted value of the expected costs of the program is deducted from the discounted expected benefits. ' ERR is a measure of the rewards to investment, taking into account the timing of the investment and the retums. It is the-discount rate which makes the NPV of a project equal to zero. population at risk in the 19 APOC countries is estimated at250.56 million (64Yo of the total population) Table I shows the estimates of population at risk in the 19 APOC countries. All APOC have at-risk populations. Populations at risk range from 0.4 million in Equatorial Guinea to over 78 million in Nigeria. Populations at risk for lymphatic filariasis represent fromZgo of the total population in Uganda to 100% of the population in Burundi, Equatorial Guinea, Gabon, and Rwanda. In the l1 OCP countries, the population at risk is 73.3 million (80% of the total population). Country estimates are shown in Table 2. Populations at risk range from 14.3 million in the Cote D'Ivoire to 1.1 million in Gineau Bissau. Estimates of populations at risk are derived from expert opinion. The estimates for the populations at risk were calculated by considering the populations of all Level One administrative units for which evidence of the presence of microfilarial infection appears in the published literature plus all those of the unsurveyed adjoining units to be potentially at risk. Administrative units in which surveys published in the literature indicated no infection are not included in the populations at risk for infection. For those countries for which there were no reports of infection status, estimates of the populations at risk were derived from data from the 4th report of the WHO Expert Committee on Filariasis (WHO 1984). The total populations for the APOC countries are taken from the World Health annual statistics for 1995. Much of the area endemic for onchocerciasis also has lymphatic filariasis infection. However, the overlap between the two diseases has not been accurately determined. We have assumed in this evaluation that there is a 50Yo overlap in populations at risk for lymphatic filariasis and onchocerciasis. Because of the degree of uncertainty in the size of the population at risk for lymphatic filariasis and the overlap in areas with endemic lymphatic frlariasis and onchocerciasis, sensitivity analyses were performed. LF Elimination Program Description The program analyzed in this study is a 5-year mass chemotherapy program. Everyone in infected communities is treated once a year for 5 years with a single dose each of ivermectin and albendazole. Pregnant \ /omen and children under the age of 5 are excluded. The mode of delivery is community-directed treatment similar to that used in the onchocerciasis programs. The first step of the program is to conduct rapid assessment of the 19 APOC and the l1 OCP countries to identifu the LF-endemic areas. The program also includes 5 years of post-treatment surveillance to assure against a recurrence of transmission. Program costs include those for rapid assessment, the APOC administrative structure, country chemotherapy programs, and post-treatment surveillance (Table 3). Because the two drugs are donated, their costs are excluded. Country-level budgets are included in the appendix of this report. We have assumed that l0% of the at-risk population has access to treatment in the first year and that an additional lÙYo are added each year until l00o/o coverage is achieved in the tenth year. V/e further irssume that 80% of the population with access to treatment actually receives treatment. I Program Effectiveness We examined the impact of the LF elimination program on the prevalence of three clinical LF conditions: ADL episodes, hydrocele, and lymphedema. Other clinical manifestations of LF, including chyluria, renal disease, and tropical pulmonary eosinophilia are not included in this analysis because of lack of data on prevalence, impact on productivity, and response to treatment. V/e used epidemiological modeling to estimate the rate of decrease with an LF elimination program that successfully intemrpts transmission of microfilaria. Overall, there is some correlation between microfilaraemic (mf) rates, ADL episodes, and chronic disease. Several studies in India reported an average microfilaremia prevalence rate of l3o with 96.5 ADL episodes per 1000 persons. (Ramaiah, and others 1998; Ramaiah, and others 1997; Ramaiah 1998b) Gyapong and others investigated a district in an upper-eastern region of Ghana with a microfilaremia prevalence rate of 4l%o of people above I 0 years of age. In this area, the ADL prevalence was 95.9 cases per 1000 persons and 19.3% of the sample size (5097 persons over 9 years ofage) reported hydrocele. (Gyapong, and others 1996a). (Gyapong, and others 1996b) In the study in southeastern Tanzaru4 a microfilaremia rate of l8%o was found in 3 villages (3000 persons). The prevalence of ADL episodes was 33 per 1000 persons was reported. Twelve percent of the men reported that they had hydrocele, and3.lYo (141 people) reported that they had any chronic disease. (Gasarasi, and others 1999) The risk of developing clinical manifestations of LF increases with age. Gasarasi found the prevalence of ADL episodes ranging from l4 per 1000 population in persons 10-19 years ofage to 88 per 1000 in persons 60 years ofage and older. Vy'e use estimates of decline in prevalence of ADL, hydrocele, and lymphedema from epidemiological modeling done by Broun at WHO (unpublished report, 1998) to estimate the impact of ADL episodes on disability adjusted life years. We have repeated- - that modeling for this exercise. The epidemiological parameters used in this model are given in Table 3. First, we constructed a hypothetical population of 1000 residents in the APOC region. This group has a similar gender and age distribution to the residents in the APOC region. We assume that this group resides in an LF-endemic area with a baseline microfilaremia prevalence of 100 per 1000 residents. This is a conservative estimate of the mean microfilaremia prevalence in the at-risk population in APOC and OCp countries. We then constructed an age- and gender-specific prevalence profile for ADL, hydrocele, and lymphedema in this community, based on reported prevalences in the literature and adjusted to reflect the 10% microfilaremia rate. Because of the uncertainÿ in the relationship between microfileremia and overt clinical manifestations of LF, we were conservative in our estimation in that we have, most likely, underestimated prevalences. We then implemented an LF progftlm in this hypothetical community. We have assumed that we have achieved an 80% coverage rate, sufficient for intemrpting transmission. We have also assumed that the program will have no impact on ADL episodes in the first year but that the incidence will decline lO%o per year beginning in tho second year. For hydrocele and lymphedema, we have assumed that there will be no new 10 cases in children who are under 6 at the time treatment is initiated. We also assumed that an increased proportion (but not all) of individuals in the younger age groups without overt clinical disease but who may be microfilaria postitive, will not progress to clinical disease. In this model, the LF program has no affect on individuals with overt clinical disease although there is some evidence in the literature to suggest otherwise. Thus, our estimations of the rate of decline in the prevalence of chronic conditions may be underestimated. Using the assumptions described above, we estimate that there will be no change in the prevalence of hydrocele and lymphedema in the first 5 years of the program, a 5Yo decrease each year between years 6 and 10, and a l\Yo decrease in the subsequent years. The age, gender, and disease profile for this hypotheticat LF community is provided in the appendix to this report. Program Benefits Program benefits are limited to the increases in labor productivity associated with the decline in the prevalence of ADL episodes, hydrocele, and lymphedema. Based on data reported in the literature on the age of onset, duration, and degree of disability associated with each of these conditions. Based on days of total incapacitation and inabiliÿ to work reported in the Tanzanian and Ghanaian studies, we estimate that each ADL episode results in a20 loss in annual productivity. Disability associated with hydrocele and lymphedema reported in the Ghana study ranges from20o to 60%. We have used 20Yo inthis analysis (Table 3). 11 We used two measures to value increases in productivity. We have reported the results using themedian per capita gross domestic product (GDP) of $300 for the APOC and the OCP countries. Per capita GDP in this region ranges from $80 in Mozambique to 54,230 in Gabon. We have used per capita GDP rather than the marginal productivity of agricultural labor (MPL) reported in the economic evaluations of the onchocerciasis program (Benton and Skinner 1995, Benton 1997) because of the potentially large LF- endemic urban populations. Because per capita GDP may overestimate the value of labor productiviÿ and to make the results comparable with the onchocerciasis evaluations, we have also analyzed the model using the MPL in sensitivity analyses. We estimated the losses in productivity that accrue each year with and without an LF program. Because this is a prevalence-based model, we only measure increases in productivity that occur within the time horizon. We do not measure the lifetime losses in productivity in persons who have ADL episodes or who have chronic manifestations, as is done in incidence-based models. Because this model is truncated at 30 years, a prevalence-based approach does not capture all of the benefits associated with the program and produces a lower estimate of benefits than an incidence-based model. Thus, the benefits measured here should be considered a lower bound estimate. RESt]LTS APOC Under the assumptions made in this analysis, the current annual economic burden of LF measured in lost productivity is approximately $1.3 billion in the APOC countries. The greatest proportion-83 %o-is due to disability in men with hydrocele. Eleven t2 percent is due to disability in both men and women with lymphedema and the remaining 6 percent is the loss from incapacitation during ADL episodes. Implementing an LF program would cost $524 million (not discounted) over l9 years. The first-year cost is approximately $12 million, increasing to $55 million in the tenth year, and then decreasing to just over $l million in the final-nineteenth-year. Figure 1 shows the investment stream over the life of the program. The distribution of costs and benefits over the 30-year time period is provided in the appendix. Benefits would begin accumulating in the second year of the program. During the first five years of the program benefits accumulate from the decrease in ADL episodes. In the 6th year, benefits from reducing the prevalence of hydrocele and lymphedema begin to accumulate. Figure 2 shows the decrease in ADL, hydrocele, and lymphedema prevalence over the 3O-year period in which program benefits are captured. The benefits (not discounted) over a 3O-year period approach $8.6 billion. The NPVs for the program using 30 and 10%o annual discount rates over a 30- year period are $4.3 billion and $1.1 billion, respectively. Underthe same conditions, the ERRs arc25Yo at the end of the investment period in 2019 arrd28%o 30 years after the initiation of the program. The program breaks even in the 13ü year using a 3% discount rate and in the 14th year with a lTYo rate. Tabte 4 summarizes the results. In the APOC countries, the program takes advantage of the existing administrative structure and the overlap in onchocerciasis- and LF-endemic areas. lntegrating LF elimination into existing APOC activities will save $45 million in program costs (not discounted) for LF. 13 OCP In OCP countries, annual productivity losses from LF approach $390 million. An LF program similar to that described for APOC would cost $169 million, §4 million in the first year, increasing to $ l7 million by the l0û year and decreasing to under § 1 million in the final year (Figure 3). The total undiscounted benefits over a 30-year period are $2.5 million. The NPVs for the program using 3olo and l|Yo annual discount rates over a 3O-year period are $1.2 billion and $302 million respectively and the ERRs are 23Yo at the end of the investment period in 2019 and26Yo 30 years after the initiation of the program (Table 5). Although we have assumed that the program in the ocp countries does not take advantage of an existing regional administrative structure similar to APOC, there are considerable financial savings from integrating LF elimination with the ongoing onchocerciasis control programs. We have estimated that these amount to savings of about $5.5 million (not discounted) over the life of the LF elimination program. The annual costs and benefits for the OCP countries are provided in the appendix. Sensitivity Analysis Total program costs are most sensitive to the size of the at risk population, the variable for which there is the most uncertainÿ in the model. The analysis is also sensitive to disability levels, discount rate, and value of productivity. However, reasonable variation in these parameters does not produce substantially different results. Under almost all scenarios, the program breaks even in either the 10th or the I lü year. l4 Sensitivity analysis also shows that the greatest program retums can be achieved with early investments in areas with high LF prevalence and in the areas already under treatment for onchocerciasis. Figure 4 shows the relationship between LF prevalence and ERR. CONCLUSION Integrating LF elimination into current APOC activities is a sound economic and human investment. Investment in LF elimination compares favorably with other health and nonhealth investments. Additionally, such an investment will improve the lives of millions of Africans for generations to come. Investing in LF is also an opportunity to take advantage of an existing public health infrastructure and build upon it, thus, saving millions of dollars compared with building a ne\^/ program from the ground up. The results of this analysis are understated because of the conservative approach taken for estimating benefits. We have assumed conservative rates of disease decline and have not included many of the benefits-including the affect to the two-drug combination on intestinal helminthes-because of lack of data. Vy'e have also not included the impact of LF on the health care systems of the endemic countries. For example, hydrocele surgeries in some hospitals in Ghana account for one-third of all operations. An LF elimination program will eventually reduce the burden LF currently imposes on these systems. We have conducted this analysis from the perspective of APOC. Therefore, we have not included the costs borne by ministries of health in endemic countries to oversee 15 LF program efforts. We have also not included the costs of the donated drugs. Because these costs are excluded, care must be taken in comparing the results of this analysis with other economic evaluations of LF programs. An analysis from the societal perspective should include these costs. Our analysis is based on the premise that the goal of the LF program is the elimination of transmission of microfilaria. Chemotherapy ends after five years of treatment. The benefits of an elimination program are not just limited to those receiving medication but accrue to the entire community--current and future members. For this goal to be accomplished high community coverage rates must be achieved. Thus, we have assigned a higher per person cost for the LF program than for onchocerciasis control. If transmission is not eliminated, ongoing chemotherapy will be required and only those receiving treatment will benefit. Schemes to share program costs such as cost- recovery programs, must be carefully evaluated for their impact on coverage rates. Retums on investments in LF control programs are likely to by substantially less than in elimination programs. Early mapping of APOC countries will allow the refinement of this analysis. Initialty identifying and targeting LF elimination programs in the onchocerciasis-endemi0 areas and the areas with high LF prevalence can improve the ERRs. Regardless, LF elimination provides a sound economic investment opportunity with benefits for generations to come. 16 REFERENCES- Benton B. 1998. "Economic Impact of onchoceriasis control through the African Programme for Onchoceriasis Control: an overview". Annals of Tropical Medicine & Parasitology 92(l ):33-39. Benton B, Skinner ED. 1990. "Cost-benefits of onchocerciasis control". Acta Leidensia 59:405-41 1. Gasarasi, D. 8., Premji, 2.G., Mujinja P. G. M., and Mpembeni, R. "Acute Adenolymphangitis due to Bancroftian Filariasis in Rufiji District, South East Tanzania". in press . 1999. Gyapong JO, Evans D, Aikins M, Gyapong M, Adjei s. 1996a. "The economic burden of lymphatic filariasis in northern Ghana". Annals of Tropical Medicine and Parasitology 90(1 ):39-48. Gyapong JO, Gyapong M, Adjei S. 1996b. "The epidemiology of acute adenolymphangitis due to lymphatic filariasis in northern Ghana". American Journal of Tropical Medicine and Hygiene 54(6):591-5. Ramaiah KD, Ramu K, Guyatt H, Vrjay Kumar KN, Pani SP. 1998. "Direct and indirect cost of the acute form of lymphatic filariasis to households in rural areas of Tamil Nadu, south India". Tropical Medicine and International Health 3(2):108-15. Ramaiah KD, Vijay Kumar KN, Ramu K, Pani SP, Das PK. 1997. "Functional impairment caused by lymphatic filariasis in rural areas of south India". Tropical Medicine and International Health 2(9) :832-8. Ramaiah KDea. 1998a. "The impact of lymphatic filariasis on labor inputs in south India: implications of a multi-site study". Ramaiah KDea. 1998b. "Treatment costs and loss of work time to individuals with chronic lymphatic filariasis in rural communities in south India". Sabesan S, Krishnamoorthy K, Pani SP, Panicker KN. 1992. "Mandays lost due to repeated attacks of lymphatic filariasis". Trends in Life Sciences 7(l):5-7. World Health Organization, Geneva. 1984. "Lymphatic Filariasis. Fourth report of the WHO Expert Committee on Filariasis". WHO Technical Report Series, No.702. t7 Table I Population at LF Risk in APOC Countries APOC Country Angola Burundi Cameroon CAR Chad Qongo Equatorial Guinea Ethiopia Gabon Kenya Liberia Malawi Mozambique Nigeria Rwanda Tanzania Uganda Zaire Sudan Total Total Population (in millions) I 11.1 6.4 t3.2 J.J 6.4 2.6 0.4 5s.0 1.3 28.2 3.0 11.1 16.0 nl.1 7.9 29.7 2r.0 43.9 28.0 389.1 Estimated at Risk Percentage of totfl Population (in population at ris$ millions) 2 I1.1 6.4 9.6 1.7 5.1 2.1 0.4 46.3 1.3 7.4 2.8 4.2 I 1.8 78.6 7.9 tr.7 6.0 26.2 10.0 t00% 100% 73% 5r% 80% 8l% 100% 84% 100% 26% 93% 37% 74Yo 70% 100% 400 29% 60% 36% 239.5 ' based on World health statistics annual, Geneva, World Health Organization, 1995 2 from prevalence data reported in Michael et al (1996) Parasitology 112:409-428 18 Table 2 Population at Risk in OCp Countries OCP Country Total Population (in millions)r Estimated at Risk Population (in millions) I Percentage of total population at risk Benin Burkina Faso Cote D'Ivoire Ghana Guinea Guinea Bissau Mali Niger Senegal Sierra Leone Togo Total 5.4 r0.3 14.3 17.5 6.7 1.1 10.8 9.2 8.3 4.5 4.1 92.2 a-). t 10.3 14.3 4.0 6.7 1.1 9.6 9.2 8.3 4.5 1.6 73.3 69% t00% 100% 23% r00% t00% 89% 100% 100% 100% 39% 80' I based on l4/orld health statistics annual, Geneva, World Health Organization, 1995 19 Table 3 Economic and Epidemiological Parameters Included in the Economic Evaluation of Integrating LF Elimination into APOC Enidemiolosical Parameters Value Prevalence of LF Conditions Microfilaremia ADL Hydrocele Lymphedema Annual Rate of Decrease in Prevalence ADL Year I Years 2-5+ Hydrocele and Lymphedema Years 0-5 6-10 ll+ Disability from LF Conditions ADL Hydrocele Lymphedema Treatment Compliance Rate Economic Parametersl 100/1000 33l1000 7sl1000 10/1000 0% t0% 0% 5% t0% 2% 20% 20% 80% Value Mapping2 APOC Administrative Overhead3 Treatment Cost3 Year 1 Years 2-5 Po st-treatnent surveill*""4 Per Capita GDP Marginal Product of Labor s897,000 $1,893,000 $ 0.49 $ 0.32 $ 0.0s $300 $1s0 'in u.s. s2 one-time costI annual cost n annual cost per person 20 Table 4 Costs, Benefits, ERRs and NPVs for LF Elimination in APOC Countries Time Period -Total Costs I Total Benefits t ERR 25% 28% NPVat3To NPVat l07o(cumulative) (cumulative) sl,582,342,304 S505,308,900 §4,07 4,503,414 $1,017,293,696 2000-2019 2000-2029 s504,207,223 $3,095,804,806 $504,207,223 $9,230,156,496 rNot discounted 27 Table 5 Costs, Benefits, ERRs and NPVs for LF Elimination in OCP Countries Time Period - Total Costs I Total Benefits I ERR NPV at 37o NPV at l07o(cumulative) (cumulftive) 2000-2019 $169,047,499 §947,642,580 23% §471,818,297 S145,f96,910 2ooo-2029 $169,047,499 §2,519,295,375 26% §1,234,682,280 $302,F 18,220 rNot discounted 22 o\ e.lO c.l (ô c.lO c.l ONON tôL o-oc\/ O N ra)OO c.l O o c.t qqqqq OTÔOTôOô.IHÈ ü) o o(J trrI o q) c.) () GI 9)È o ào 6l U clL à0oLÈ o c! H frl E Ft EI6 È rË, o) 6l q) o U)O ü) o "é U tÉÈ () c) {) c! o)LÈ â q) ah G o)tro c)â €)li à0 ,\ 23 Oôtoc\ oo O cil \o O c.lÿ Oôt N O e.l OL OJ)c\r' ooO N \o ô1 saa(\ c.lOO c.l O Oc! qqq<?qq OOOOOO(ô$cîNÊ $'s'oJo suomlntr U oÊr I o) H Ér Lq) o an ü)oU t! à0 ÊÈ C!I 0 U oÀ o cÉ E1 tu Fl L € U' ü) oU Gt ê0oliÊr G 6) a N q) L â0 tiEr !o oo ! o z 24 oc{O c.l oo o c.l \o O c! s oN N o c.l OliHGIoJ)c\/ ooôOôi \Ô c! \f, o(.I C.lOoôl OO c.l qqqqaooooo oo \O .t c.l O od \Cj + c.i OHÊ-HH $'slJo §uoml^tr ÈU oIq) H fr§) o ar) U)oU tg àDp È fH o cÉI a aq) L oUÈU o q) (l) Fi ,i o) o rn U'o O 6üfr à0oLÈ c€ c)L a (1 q) L êo tu .o o o() oË oz o d Ê.oÊ. ooo L a) a. (l) q oÀ (d o (É oro o ro-Ê. OO cl 6l ÈoÊ o OF{Otrc'.t I q)(.) C) cl O o.)olr cÉ I cl ùi .9 l'{à çÀq) I q) ctl I Êr ll 9) l-{ â cl q) ü U) 0) Gtü() o() r-l 6tL â0oLÈ o «l rc E F] c\! li e)É o q) ù() o c) f-l o q) q,) q) «t c)LÈ E(l) L cÉ oL e, 13à o o é)qi EI ç c)L èo f* GZOZ-OOOZ srze§ UUtr \O \O \O \O \O \O \O 6\ 6\ 6\ q\ 5\ 6\ 6\OOOOOOO\O ra) $ c.| c..l È 26
t-1. ^ OO-o3 3FUZ-(D U c _o-lA r- { C} @ o -DO m :D rn C o rn 3i --z U)o:rm C)_D -om I .-o@ Om)Soffi ô'8 eë=Pgiio ,*. z. aU;iC{-oôTz-o C m I o . 2". o -i- o. :o -o --t-t ooC @ ô) ô r§ È' l"'' -È l'|- {*' :' t§ 1*'".1Jô , I ['l- \ t,À l) fl v'üa eO o_ !*r @ ''â aI :' o) c-) a'! :. ç"=t) .fi,* ,[1 .n- \-) o- o- (D <n tn ml €() !.I <(D ':g ol 6(D =- æè:-1gâ -r(D .= o) =U,Y o -J -D C3p. c, aô ,- 1- tsa