The Leogane, Haiti Demonstration Project: Decreased Microfilaremia and Program Costs After Three Years of Mass Drug Administration

Home , Haiti

THE LEOGANE, HAITI DEMONSTRATION PROJECT: DECREASED MICROFILAREMIA AND PROGRAM COSTS AFTER THREE YEARS OF MASS DRUG ADMINISTRATION

MADSEN BEAU DE ROCHARS, SANJAT KANJILAL, ABDEL N. DIRENY, JEANNE RADDAY, JACK G. LAFONTANT, ELS MATHIEU, RICHARD D. RHEINGANS, ANNE C. HADDIX, THOMAS G. STREIT, MICHAEL J. BEACH, DAVID G. ADDISS, AND PATRICK J. LAMMIE* Hopital Ste. Croix, Leogane, Haiti; Lymphatic Filariasis Support Center, Emory University, Atlanta, Georgia; Division of Parasitic Diseases, Centers for Disease Control and Prevention, Atlanta, Georgia; Center for Tropical Disease Research and Training, University of Notre Dame, Notre Dame, Indiana

Abstract. To support the global program to eliminate lymphatic filariasis (LF), well-monitored demonstration projects are important for defining the relationship between coverage and reductions in microfilaremia. We are using mass treatment with diethylcarbamazine (DEC) and albendazole in an effort to eliminate LF from Leogane, Haiti. Wuchereria bancrofti microfilaremia prevalence at baseline ranged from 0.8% to 15.9% in four sentinel sites. After three rounds of DEC-albendazole mass drug administration (MDA), both microfilaremia prevalence and intensity decreased dramatically. Mild and moderate adverse reactions after treatment were common, especially after the first MDA, but decreased after subsequent MDAs. Drug coverage for the first year was estimated to be 72%, but concerns about adverse reactions appeared to decrease drug coverage in the second MDA. As a result of community education efforts that focused on providing a greater understanding of adverse reactions, coverage increased dramatically for the third round. Program efficiency increased substantially; the costs per person treated for three rounds of MDA were $2.23, $1.96, and $1.30 per person, respectively. The Leogane experience highlights the importance of adapting community education and mobilization campaigns to achieve and maintain good coverage.

INTRODUCTION on centrally located posts. Intensive health education was used to sensitize and motivate people to participate. We re- Approximately 120 million persons are infected with the port here our results through the first three annual cycles of filarial worms Wuchereria bancrofti and Brugia malayi.1 De- drug distribution in Leogane and our analysis of the costs of spite these overwhelming numbers, there is substantial opti- the program. mism that lymphatic filariasis (LF) can be eliminated.2 First, there is historical evidence that transmission of LF can be interrupted directly as a result of public health interventions MATERIALS AND METHODS or indirectly through economic development.3–6 Second, new Study site. Leogane is located approximately 30 km west of tools have been developed to facilitate mapping of at-risk 7–9 Port au Prince, Haiti. The commune includes two distinct areas and implementation of mass treatment. Annual zones: a coastal plain and a mountainous area. Surveys of treatment with diethylcarbamazine (DEC) or ivermectin in school children showed that W. bancrofti antigen prevalence combination with albendazole leads to substantial reductions was as high as 50% or more in communities located in the in the prevalence and intensity of filarial infection in both 8,10 coastal plain; infection levels in the foothills and mountains humans and vector mosquitoes. Lymphatic filariasis elimi- 11 were significantly lower. Because of the lack of census in- nation programs based on this strategy have begun in more formation, neither the population of the plains nor of the than 30 countries. Because programs are at an early stage, commune of Leogane was known with certainty when the there are a number of scientific and programmatic questions project started; estimates ranged from 100,000 to 200,000. that have not been fully addressed. Are five rounds of mass Programmatic efforts were concentrated in the plains because treatment enough to eliminate transmission? What level of of the higher endemicity in this region. coverage is sufficient? Is high coverage sustainable in differ- Community education and social mobilization. Nearly one ent country settings? What costs are associated with develop- year was spent on the community mobilization process prior ing and operating LF elimination programs? to the first annual mass drug administration (MDA). Health The Leogane demonstration project, carried out by Hopital education messages that focused on transmission and preven- Ste. Croix, was developed to serve as the foundation for a tion of filariasis were developed and tested. These messages national filariasis elimination program in Haiti. As with other were used as the basis for training a team of health educators filariasis elimination programs, the goal of the program in and Hopital Ste. Croix community health workers. To dis- Haiti is to interrupt transmission by reducing the prevalence seminate the filariasis messages as broadly as possible, health and intensity of circulating microfilaria below the point where educators conducted hundreds of community meetings in transmission can be sustained. To achieve this objective, the schools, churches, and at vaccine posts to explain the nature entire at-risk population was offered treatment annually with of the program. In the weeks leading up to the annual MDA, DEC and albendazole, using a distribution mechanism based radio spots, posters, banners, and a sound truck were used to mobilize the public to participate in the MDA and to provide information on the dates of the MDA and location of the * Address correspondence to Patrick J. Lammie, Division of Parasitic Diseases, Mailstop F13, Centers for Disease Control and Preven- distribution posts. Health messages were modified annually, tion, 4770 Buford Highway, Atlanta, GA 30341. E-mail: pjl1@ based on community feedback obtained through knowledge, cdc.gov attitudes, and perceptions (KAP) surveys, focus groups, and 888 DECREASES IN MICROFILAREMIA AFTER MDA 889 informal discussions between project staff and community by nurses were established to treat moderate adverse reac- residents. For example, a television powered by a portable tions, including scrotal pain, itching and other problems that generator was used to show LF videos in evening community interfered with daily activities. Persons with severe adverse meetings after the MDA in the second year and flip charts reactions were referred to Hopital Ste. Croix for evaluation were used to address specific community concerns as dis- and hospitalization, if necessary. Surveillance for adverse re- cussed in this report. Community leaders also supported the actions in the first year represented a substantial investment program by recording public service announcements for local in project resources for the 2000 MDA, both to reassure radio stations. health authorities in Haiti and to collect information.12 Less Sentinel sites. To monitor programmatic progress, four emphasis was placed on surveillance for adverse events in plains communities were selected to serve as sentinel sites. 2001 and 2002. These communities were selected to be representative of the Coverage surveys. After the first round of drug distribution at risk population in terms of the type of community (e.g., in 2000, three different techniques were used to assess cover- rural versus urban) and the filarial antigen level. Sentinel sites age: two convenience methods and a more statistically rigor- served as the focus of research activities in the context of the ous method based on cluster sampling with probability pro- overall public health program. Research protocols were re- portional to size.13 As noted earlier in this report, the popu- viewed and approved by Centers for Disease Control and lation of the plains of Leogane was not known with any Prevention and University of Notre Dame Institutional Re- certainty. The cluster survey was used to generate an estimate view Boards and the Ethics Committee of Hopital Ste. Croix. of the population size for the plains area. This number was Filarial infection. Microfilaremia prevalence and intensity used as the denominator for coverage estimates for 2001. A as well as antigen prevalence were monitored at baseline and second cluster survey was performed after the third MDA in at annual intervals. Blood collections were done between 7:00 2002. Coverage in sentinel sites was assessed annually by in- PM and 9:00 PM. Thick films were prepared (20 ␮L), stained terviewing persons who participated in follow up sample col- with Giemsa, and examined for microfilaria. Antigenemia lections. These took place approximately 8–10 months after was detected with the immunochromatographic card test MDA. (ICT).7 For surveys conducted in 2000 and 2001, this was Assessment of health education. To assess the impact of the done with cards produced by Amrad-ICT (Melbourne, Vic- health education on the KAP of filariasis and the mass treat- toria, Australia); for the 2002 surveys, the cards were pro- ment program, KAP surveys were conducted after the first duced by Binax (Portland, ME). and third distribution, in parallel with the coverage surveys Mass drug administration. Health clinics, churches, and pri- described earlier in this report. An analysis of the results of vate houses were used as distribution posts. Schools were the first KAP survey has been reported elsewhere.14 In addi- added as distribution posts in 2001 and thereafter. Distribu- tion to the KAP surveys, focus groups were held after the tion posts were selected to cover the entire plains to provide second annual drug distribution to provide a deeper under- convenient access for persons in all communities. Numbers of standing of the reasons why some people refused treatment. posts varied from 103 in 2000 to 71 in 2001 to 133 in 2002. Cost analyses. Cost data were collected retrospectively be- Each post was staffed by community volunteers who received cause the program was already underway as the cost analysis a small payment (350–500 Gourdes, approximately $10–15) began. Costs were estimated from expenditure records and for the work. These persons distributed the DEC and al- interviews with administrators for each input that contributed bendazole and provided the first care for treatment-related to program operations in both the plain and the mountains. adverse reactions. Community health workers and program Primary outcome measures are the annual program costs staff supervised the distribution posts. Diethylcarbamazine (broken down by program activities), the cost per person was distributed to all persons greater than two years of age, treated and the cost per person at-risk. except pregnant women and those who were judged by drug There are five input categories that generate these costs: distributors to be too ill to receive treatment. The dose of personnel, capital costs, equipment and facilities, supplies, DEC was based on age for children less than 18 years old. We and transportation. The costs of any inputs (such as personnel previously showed that age and height were strongly corre- or transportation) that were used for multiple activities were lated with weight among children in the commune (Beach MJ distributed among the activities. Personnel, transportation, and others, unpublished data). Adults received a standard equipment, and facility costs were allocated based on the per- DEC dose of 400 mg. All persons treated with albendazole centages of time they spent on a given activity. Other non- received one 400-mg tablet (GlaxoSmithKline, Research Tri- monetary data that are necessary for the cost calculations angle Park, NC). At the request of the Ministry of Health, include the expected useful life and residual or scrap value of women of child-bearing age were not treated with albenda- capital investments (e.g., vehicles), which are necessary to cal- zole in 2000 or 2001 because of concerns about albendazole culate their annuitized costs. Cost data were entered into Mi- treatment of women in early stages of pregnancy. These ex- crosoft (Redmond, WA) Excel௡ worksheets for analyses. clusion criteria were removed in 2002. The MDA in the Costs for non-MDA activities (e.g., research projects) were mountains followed the drug distribution in the plains by ap- excluded from the analysis. proximately one month because of the limited supervisory The annual cost of any given activity is the sum of each staff. input’s cost (or use) for the year multiplied by its percent Adverse reaction surveillance. For the initial round of allocation to that activity for that year. Equipment/facility and MDA, a three-tiered referral system was developed for ad- transportation costs are comprised of operational costs and verse reactions.12 Treatment of minor complaints (e.g., head- annuitized capital costs. A straight-line depreciation method ache, low-grade fever) that developed after treatment was was used to annuitize capital costs. Personnel and supply costs available at the distribution posts. Ten referral centers staffed have only an operational cost component. 890 BEAU DE ROCHARS AND OTHERS

The calculation of annual cost per person treated consisted of dividing the total cost of the program by the number of persons given drugs in that year of distribution. Annual cost per person at risk is the total program cost divided by the number of individuals residing in the commune. The population of Leogane was assumed to be constant over the three- year period that data were analyzed. Adjustments were made for inflation over the three rounds of the analysis. The base year of the dollar was set to 2002 and all costs previous to this date were inflated to 2002 dollars. Costs were collected in U.S. dollars and Haitian Gourdes. FIGURE 1. Changes in the number of persons treated and in those Final results are presented in 2002 US dollars. A Gourde- reporting adverse reactions. The number treated in each round of mass drug administration is shown in the bars. The percentage of Dollar exchange rate was calculated for each round of the persons reporting adverse reactions is plotted as a line. program. Rates were obtained from www.oanda.com and are the average of the daily rates between the first and last dates designated for that round. In response to the lower coverage in 2001, health commu- Statistical analysis. Parasitologic data were entered into an nications strategies were modified to place greater emphasis EpiInfo version 6.0 (Centers for Disease Control and Preven- on the relationship between adverse reactions and treatment. tion, Atlanta, GA) database. Changes in infection prevalence We also tried new methods to motivate the community, in- and intensity were analyzed by chi-square and Kruskal-Wallis cluding showing videos about filariasis in the communities tests, respectively. and broadcasting radio spots with statements from community leaders supporting the program. As a result of all these RESULTS changes, the number of persons treated and the drug cover- Baseline infection level. At baseline, antigen prevalence in age increased significantly in 2002 (Figure 1 and Table 2). the four sentinel sites ranged from 10.2% to 50.1% and mi- Reported adverse reactions decreased with each MDA, from crofilaremia prevalence ranged from 0.8% to 16% (Table 1). 23.1 per 100 persons treated in the first year to 9.0 in 2002 The communities represent urban, rural, and mixed settings. (Figure 1). The community that had the lowest level of W. bancrofti in- Effect on infection prevalence and intensity. Microfilare- fection, Mapou, was located near the foothills. Previous stud- mia prevalence decreased significantly in each of the sentinel ies have shown that antigen prevalence decreases with in- sites following MDA except Mapou (Figure 2). By 2003, mi- creasing altitude.11 crofilaremia prevalence had decreased to less than 2% in all Mass drug administration. For the first MDA in October of the sentinel sites except the urban setting of Leogane town. 2000, more than 70,000 persons were treated in the plains Of those who were microfilaria positive, mean (and median) during four days of distribution (Figure 1). Based on coverage microfilaria density also decreased significantly (Table 3). surveys, 72% of the plains population received treatment.13 Antigen prevalence decreased in 2003 to 8.6%, 20.1%, 15.4%, In the first year of the program, we implemented a system of and 29.8% for Mapou, Masson/Mathieu, Barrier Jeudi, and enhanced surveillance for adverse reactions. More than Leogane, respectively; however, it was not possible to analyze 16,000 people (23.1 per 100 persons treated) reported to a trends in antigenemia because the transition from AMRAD distribution post with complaints of adverse reactions, espe- to Binax ICT cards in 2002 was accompanied by an artifactual cially fever and headache (Figure 1). Most complaints were increase in antigen prevalence. minor; however, more than 2,500 men reported scrotal pain.12 Cost analyses. A multiyear comparison shows that the The KAP surveys after the first MDA led us to anticipate overall costs of the Leogane MDA decreased from round 1 to better coverage in year 2; however, on the first day of the round 2 and increased slightly again for round 3 (Figure 3). distribution, there was a substantial decrease in the number of The decrease in cost was related to a decrease in MDA ex- persons treated relative to year 1 (5,900 persons versus 26,000 penditures and a steep decrease in the costs of adverse event persons). In response to the lower numbers, the staff re- surveillance (Table 4). The decrease in MDA costs for rounds sponded vigorously with visits to churches, schools, and radio 2 and 3 was the result of lower personnel costs because of the stations during the remaining days of the drug distribution. reduction in the number of volunteers manning the posts after As a result, the numbers treated increased daily; nonetheless, the number of persons treated and the overall coverage decreased in 2001 (Figure 1 and Table 2). TABLE 2 Comparison of overall coverage in the plains of Leogane with that of the sentinel sites* TABLE 1

Baseline Wuchereria bancrofti prevalence in the sentinel sites, 2000* MDA treatment coverage (%)

Year Leogane plains† Sentinel sites‡ Community Microfilaremia (%) Antigenemia (%) Mapou 5/587 (0.8) 59/579 (10.2) 2000 73 78.4 Masson/Mathieu 125/1,140 (11.0) 423/1,149 (36.8) 2001 52 87.8 Barrier Jeudi 78/1,155 (6.7) 392/1,134 (34.6) 2002 78 92.9 .mass drug administration ס Leogane town 98/617 (15.9) 302/603 (50.1) * MDA † As assessed by cluster survey.16 * Microfilaria and antigen prevalence differed significantly by sentinel site (P < 0.001 for ‡ As determined by interview of persons providing blood specimens during follow-up both sets of comparisons). surveys. DECREASES IN MICROFILAREMIA AFTER MDA 891

TABLE 3 Changes in intensity of microfilaremia (per 20 ␮L) of microfilaria (Mf)–positive persons*

Year No. tested No. Mf positive Mean† Median Range 2000 3,499 302 18.9 8 1–322 2001 1,861 70 8.1 4 1–84 2002 1,439 31 10.6 3 1–78 2003 1,260 15 4 2 1–15 * Microfilaria intensity decreased significantly (P < 0.001) over time. † Arithmetic mean of Mf positive persons.

FIGURE 2. Prevalence of microfilaremia before and after three rounds of mass treatment. The prevalence of infection decreased other materials that dealt specifically with issues related to significantly (P < 0.001) for all of the sentinel sites except Mapou adverse reactions and the relationship between the adverse -reactions and the therapeutic action of the drugs. The in .(0.14 ס P) creased coverage that resulted in the third MDA suggests that these efforts were effective and serves as a reminder of the round 1 and a reduction in supply expenditures. For example, critical importance of adapting health communications mes- whereas $22,295 was spent on distribution supplies for round sages to the needs of the program. The KAP surveys have 1, an average of only $9,435 was spent in the next two rounds emphasized that participation in MDA is associated with (a reduction of 58%). The cost for treatment of adverse re- knowledge of LF14; unfortunately, social mobilization activi- actions also decreased markedly after the first round of the ties are typically the first to be cut in the face of budgetary MDA. These costs initially consumed 24% of total program shortfalls. costs but decreased to only 4% and 7% in the following Coverage surveys provide essential information for moni- rounds. Social mobilization costs increased steadily from 10% toring program implementation and for validating coverage to 28% of total costs. estimates derived from distribution posts. It is important to Table 5 shows the cost per person treated. Cost per person assess coverage patterns as well as raw numbers. Coverage treated decreased substantially each round. Round 3 was the surveys conducted after the third MDA were designed to most cost-efficient round of the Leogane MDA with cost per assess the degree to which noncompliance was systematic. A person treated decreasing to $1.30. significant proportion of the adult population (approximately 18%) reported that they had not taken the drugs in any round DISCUSSION of the MDA (Mathieu E and others, unpublished data). In settings where initial microfilaremia prevalence is high (e.g., Lymphatic filariasis elimination programs hold the promise > 10%), it is clear that significant levels of systematic non- of reducing the burden of filarial infection and disease while compliance represent a threat to the target of elimination. providing important public health benefits through reduction Preliminary assessments of serologic responses after the first in intestinal helminth infections. After only three rounds of MDA showed that antifilarial antibody responses were influ- DEC-albendazole treatment, Leogane has experienced dra- enced by proximity to untreated infected persons.17 Thus, matic reductions in both filarial and geohelminth infections noncompliant persons are likely to represent a reservoir of while achieving concurrent gains in program cost and effi- infection that can lead to focal transmission. Defining the ciency; thus, broad public health benefits are being realized conditions under which microfoci can persist is an important through implementation of the program.15 Based on our ex- objective for future research. Similarly, understanding more perience, we suggest that social mobilization plays an impor- about noncompliance and about how to motivate noncompli- tant role, both in developing a distribution post-based MDA ant persons to participate in MDA is an urgent research issue. program and in addressing community concerns that result in We observed substantial reductions in microfilaremia lower coverage. prevalence after three cycles of mass treatment with DEC and A key programmatic indicator for public health programs is albendazole (Figure 2). The site where microfilaria preva- coverage. Coverage decreased from 72% in the first MDA to lence remained the highest was the urban setting. Although only 52% in the second MDA (Table 2). Although the rea- we assumed that lower coverage was the explanation for the sons for this decrease are undoubtedly complex, it appeared difference between the urban (Leogane town) and more rural from focus groups that fear of adverse reactions was a com- setting (e.g., Masson/Mathieu), coverage surveys indicate that ponent of the community’s aversion to treatment in the sec- coverage was comparable in these settings (Mathieu E and ond year. Since the coverage needed to eliminate transmis- others, unpublished data). It is possible that population mi- sion is thought to be much higher than 50% of the total popu- gration into Leogane town, from communities where no lation,16 this decrease represented a real threat to the MDA is conducted, contributed to the maintenance of higher program. The program increased the investment in social mo- infection levels. Higher mosquito densities in the urban set- bilization to address these issues (Table 4). Enhanced social ting may provide an alternative explanation. mobilization can overcome problems with decreased cover- The reduction in microfilaria level that we observed, age if the root causes of the problem are adequately ad- though encouraging, must be interpreted with some caution. dressed through messages tailored to the community. Health Because some persons in sentinel sites refused to be re-tested educators working with the program used feedback from after baseline data collection in 2000, bias was introduced into KAP surveys and focus groups to develop flip charts and the follow up sampling. In general, persons who agreed to be 892 BEAU DE ROCHARS AND OTHERS

mass drug ס FIGURE 3. Program costs by year of distribution. The costs of the indicated program activities are plotted by year. MDA administration. re-checked were also more likely to participate in the MDA. planned national LF programs in other countries. In Haiti, Coverage among persons sampled in the sentinel sites was where little infrastructure exists, start-up costs are higher than consistently higher than that of the general population (Table in other settings where support from the national government 2). These differences suggest that spot checks, as called for in or other existing health programs helps divert much of the the Program Manager’s Guidelines, may be a useful tool for cost burden away from the program. Program efficiency in- verifying programmatic progress.18 creased over time as the program management gained prac- Decreases in adverse reactions were noted in the second tical experience with the program. Decreasing costs (Table 4) and third years of the program, consistent with the decrease in and increased coverage led to lower costs per person treated microfilaremia prevalence. We did not make any systematic in 2002 (Table 5). attempt to establish a link between filarial infection and ad- A comparison of the costs in Haiti to the few cost studies verse events; thus, it is likely that a substantial number of done on MDA programs in India and Tanzania showed that visits to the health posts reflect health needs of the population the Leogane program costs started out higher but are ap- that are unrelated to MDA. The rapid decrease in adverse proaching values that are comparable to those in the other reactions (Figure 1) suggests that enhanced surveillance for countries. Krishnamoorthy and others in their cost effective- adverse reactions is not needed beyond the first year if cov- ness analysis of an MDA program in southern India obtained erage is adequate, even in settings such as Leogane where an annual cost per person treated of $1.21 and $1.18 for two initial infection prevalence is high. Decreasing the surveil- rounds of distribution, respectively, and $0.65 and $0.69 per lance for adverse events led to substantial cost savings in 2001 person at-risk.19 Similarly, Michael and others conducted a and 2002 (Table 4). study in Tanzania that showed a cost per person at-risk of The cost analysis provides essential insight for existing and $0.70 over the course of three years.20 A systematic study of

TABLE 4 Program input costs by activities and round of mass drug administration (MDA) (US$)

Activities

Adverse Social event General Mapping mobilization MDA treatment Monitoring administration Total Round 1, 2000 Personnel 8,979 11,540 63,461 32,371 11,540 11,540 139,432 Transportation 4,755 4,755 7,263 0 7,266 4,755 28,793 Equipment/Facilities 935 4,479 7,143 3,527 3,416 1,199 20,699 Supplies 1,360 1,675 22,295 20,036 1,910 103 47,379 Total 16,029 22,449 100,162 55,933 24,132 17,597 236,302 Round 2, 2001 Personnel 7,906 23,995 55,390 5,752 6,167 2,082 101,293 Transportation 0 6,385 6,116 126 6,509 4,510 23,646 Equipment/Facilities 641 5,049 5,138 546 2,607 200 14,181 Supplies 31 2,986 8,893 223 4,580 88 16,801 Total 8,579 38,415 75,537 6,647 19,862 6,881 155,921 Round 3, 2002 Personnel 3,806 28,478 44,809 10,691 10,672 8,251 106,705 Transportation 172 7,786 5,354 289 4,725 7,049 25,375 Equipment/Facilities 390 3,659 2,331 426 2,162 632 9,600 Supplies 62 3,624 9,976 124 2,102 103 15,991 Total 4,429 43,547 62,469 11,530 19,661 16,034 157,672 DECREASES IN MICROFILAREMIA AFTER MDA 893

TABLE 5 Leogane, the people living in the sentinel sites, colleagues at the Cost per person treated* Centers for Disease Control and Prevention, and especially the team at GlaxoSmithKline for their generous support for the LF program in Persons Haiti and in other countries through their donation of albendazole. Endemic covered Total cost Cost/person We especially thank Dr. Marie Denise Milord for her support of the MDA round population (% of total) (US$) treated program in Leogane, Gladys Mayard for her work with the focus 1 150,000 105,750 (71) 236,302 2.23 groups, Joyanna Wendt for her work developing health communica- 2 150,000 79,713 (53) 155,921 1.96 tion messages, and Wendi McAfee and Michelle Sexton for essential 3 150,000 121,139 (81) 157,672 1.30 support with logistics. -mass drug administration; includes distribution in mountains. Financial support: This project was supported by the Emerging In ס MDA * fections Program of the Centers for Disease Control and Prevention and by a grant from the Bill & Melinda Gates Foundation to the costs, using standardized data collection methods, is needed University of Notre Dame. to understand variations in costs across programs. Authors’ addresses: Madsen Beau de Rochars, Hopital Ste. Croix, Intensively monitored demonstration projects provide ideal Leogane, Haiti, Telephone: 509-555-5246, Fax: 509-235-1845, E-mail: [email protected]. Sanjat Kanjilal, Lymphatic Filariasis Support opportunities to learn how to build and monitor LF elimina- Center, Emory University, 1518 Clifton Road, Atlanta, GA 30322, tion programs. Many lessons have been learned from the ex- Telephone: 770-488-1187, Fax: 770-488-1148, E-mail: perience in Leogane (Table 6). For example, initial concerns [email protected]. Abdel N. Direny, Hopital Ste. Croix, Leogane, about treating pregnant women with albendazole led to the Haiti, Telephone: 509-551-6445, Fax: 509-235-1845, E-mail: decision to exclude all women of child-bearing age from al- [email protected]. Jeanne Radday, Division of Parasitic Diseases, Mailstop F22, Centers for Disease Control and Prevention, 4770 Bu- bendazole treatment, which prevented women from receiving ford Highway, Atlanta, GA 30341, Telephone: 770-488-7538, Fax: the benefits of this drug.15 After two years of MDA, this 770-488-7761, E-mail: [email protected]. Jack G. Lafontant, Ho- decision was reversed when it was noted that other Ministry pital Ste. Croix, Leogane, Haiti, Telephone: 509-555-7692, Fax: 509- of Health programs were using questionnaires to ascertain 235-1845, E-mail: [email protected]. Els Mathieu, Division of Parasitic Diseases, Mailstop F22, Centers for Disease Control and pregnancy status. This simplified treatment decisions at the Prevention, 4770 Buford Highway, Atlanta, GA 30341, Telephone: post level and reduced confusion about women taking part in 770-488-3603, Fax: 770- 488-7761, E-mail: [email protected]. Richard D. the MDA that led to women being inappropriately excluded Rheingans, Lymphatic Filariasis Support Center, Emory University, from MDA. 1518 Clifton Road, Atlanta, GA 30322, Telephone: 404-727-2425, Clearly, there is much we still need to learn. We do not yet Fax: 404-727-5530, E-mail: [email protected]. Anne C. Haddix, Lymphatic Filariasis Support Center, Emory University, 1518 Clifton know how many years of MDA will be required to eliminate Road, Atlanta, GA 30322, Telephone: 404-498-3337, Fax: 404-498- transmission. As noted earlier in this report, the effect of 1111, E-mail: [email protected]. Thomas G. Streit, Center for Tropical systematic noncompliance also must be carefully evaluated. Diseases, University of Notre Dame, 351 Galvin Hall, Notre Dame, Nonetheless, even at this stage of the program, MDA has IN 46556, Telephone: 574-631-3273, Fax: 574-631-7413, E-mail: [email protected]. Michael J. Beach, Division of Parasitic Diseases Mail- delivered significant public health gains to the population of stop F22, Centers for Disease Control and Prevention, 4770 Buford Leogane. We should not allow concerns about current finan- Highway, Atlanta, GA 30341, Telephone: 770-488-7763, Fax: 770- cial challenges facing the LF elimination program to diminish 488-7761, E-mail: [email protected]. David G. Addiss, Division of Para- the intensity of our efforts to achieve this goal. sitic Diseases, Mailstop F22, Centers for Disease Control and Pre- vention, 4770 Buford Highway, Atlanta, GA 30341, Telephone: 770- 488-7770, Fax: 770-488-7761, E-mail: [email protected]. Patrick J. Received April 25, 2005. Accepted for publication June 29, 2005. Lammie, Division of Parasitic Diseases, Mailstop F13, Centers for Acknowledgments: We thank the demonstration project staff in Disease Control and Prevention, 4770 Buford Highway, Atlanta, GA 30341, Telephone: 770-488-4054, Fax: 770-488-4108, E-mail: [email protected].

TABLE 6 Lessons learned* REFERENCES 1. Molyneux DH, Zagaria N, 2002. Lymphatic filariasis elimination: 1. A distribution post-MDA strategy can work as well as a progress in global programme development. Ann Trop Med house-to-house strategy. Parasitol 96: S15–S40. 2. Excluding all women of child-bearing age from getting 2. Ottesen EA, Duke BOL, Karam M, Behbehani K, 1997. Strate- albendazole decreases coverage and denies women benefits of gies and tools for the control/elimination of lymphatic filaria- treatment. sis. Bull World Health Organ 75: 491–503. 3. Adverse reactions can be managed effectively at the community 3. Cao WC, Xu JF, Ren ZX, 1994. Epidemiological surveillance of level and they decrease in incidence after the first cycle of filariasis after its control in Shandong Province, China. South- MDA. east Asian J Trop Med Public Health 25: 714–718. 4. KAP surveys are important as a tool to refine community 4. Webber RH, 1979. Eradication of Wuchereria bancrofti infection mobilization strategies. through vector control. Trans R Soc Trop Med Hyg 73: 722– 5. Simple coverage surveys are important for program monitoring 724. and for defining patterns of noncompliance. 5. Rawlins SC, Siung-Chang A, Baboolal S, Chadee DD, 2004. Evi- 6. Sentinel sites are useful for demonstrating reductions in dence for the interruption of transmission of lymphatic filari- microfilaremia, but sentinel site data may not be representative asis among schoolchildren in Trinidad and Tobago. Trans R over the life of the program because community residents tire of Soc Trop Med Hyg 98: 473–477. annual sample collection. 6. Chernin E, 1987. The disappearance of Bancroftian filariasis 7. MDA provides broad public health benefits because of the from Charleston, South Carolina. Am J Trop Med Hyg 37: de-worming action of the drugs. 111–114. 8. MDA costs decrease and program efficiency improves with 7. Weil GJ, Lammie PJ, Weiss N, 1997. The ICT filariasis test: A successive rounds of treatment as program managers gain rapid format antigen test for diagnosis of Bancroftian filariasis. experience with personnel and supply costs. Parasitol Today 13: 401–404. -knowledge, attitudes, and perceptions. 8. Brown KR, Ricci FM, Ottesen EA, 2000. Ivermectin: effective ס mass drug administration; KAP ס MDA * 894 BEAU DE ROCHARS AND OTHERS

ness in lymphatic filariasis. Parasitology 121(Suppl): S133– 15. Beau de Rochars M, Direny AN, Roberts JM, Addiss DG, Rad- S146. day J, Beach MJ, Streit TG, Dardith D, Lafontant J, Lammie 9. Ottesen EA, Ismail MM, Horton J, 1999. The role of albendazole PJ, 2004. Community-wide reduction in prevalence and inten- in programs to eliminate lymphatic filariasis. Parasitol Today sity of intestinal helminths as a collateral benefit of lymphatic 15: 382–386. filariasis elimination programs. Am J Trop Med Hyg 71: 466– 10. Bockarie MJ, Tisch DJ, Kastens W, Alexander ND, Dimber Z, 470. Bockarie F, Ibam E, Alpers MP, Kazura JW, 2002. Mass treat- 16. Stolk W, Swaminathan S, van Oortmarssen GJ, Das PK, ment to eliminate filariasis in Papua New Guinea. N Engl J Habbema JD, 2003. Prospects for elimination of Bancroftian Med 347: 1841–1848. filariasis by mass drug treatment in Pondicherry, India: a simu- 11. Boyd HA, Waller LA, Flanders WD, Beach MJ, Sivilus JS, Lo- lation study. J Infect Dis 188: 1371–1381. vince R, Lammie PJ, Addiss DG, 2004. Community- and indi- 17. Washington CH, Radday J, Streit TG, Boyd HA, Beach MJ, vidual-level determinants of Wuchereria bancrofti infection in Addiss DG, Lovince R, Lovegrove M, Lafontant JG, Lammie Leogane Commune, Haiti. Am J Trop Med Hyg 70: 266–272. PJ, Hightower AW, 2004. Spatial clustering of filarial trans- 12. McLaughlin SI, Radday J, Michel MC, Addiss DG, Beach MJ, mission before and after a mass drug administration in a set- Lammie PJ, Lammie J, Rheingans R, Lafontant J, 2003. Fre- ting of low infection prevalence. Filaria J 3: 3. quency, severity and costs of adverse reactions following mass treatment for lymphatic filariasis using diethylcarbamazine 18. WHO, 2000. Preparing and Implementing a National Plan to and albendazole, Leogane, Haiti, 2000. Am J Trop Med Hyg Eliminate Lymphatic Gilariasis. Geneva: World Health Orga- 68: 568–573. nization. 13. Mathieu E, Deming M, Lammie P, McLaughlin S, Beach M, 19. Krishnamoorthy K, Ramu K, Srividya A, Appavoo NC, Saxena Deodat DJ, Addiss D, 2003. Comparison of methods for esti- NB, Lal S, Das PK, 2000. Cost of mass annual single dose mating drug coverage for filariasis elimination, Leogane, Haiti. diethylcarbamazine distribution for the large scale control of Trans R Soc Trop Med Hyg 97: 501–505. lymphatic filariasis. Ind J Med Res 111: 81–89. 14. Mathieu E, Lammie PJ, Radday J, Montilus W, Beach MJ, Streit 20. Michael E, Meyrowitsch DW, Simonsen PE, 1996. Cost and cost T, Wendt J, Addiss DG, 2004. What influenced people to par- effectiveness of mass diethylcarbamazine chemotherapy for ticipate in a mass treatment campaign against lymphatic fila- the control of Bancroftian filariasis: comparison of four strat- riasis? Ann Trop Med Parasitol 98: 703–714. egies in Tanzania. Trop Med Int Health 1: 414–426.