A summary report of the evaluation of ivermectin distribution in the OCP: January 1995

n {}srt;< WORLD IIIIALTII ORGANIZA'|ION ORGANISATION MONDIALE DE LA SAN'IE Onchocerciasis Control Programme in West Africa Programme de Lutte contre I'Onchocercose en Afrique de I'Ouest EXPERT ADVISORY COMMITTEE Sixteenth session Ouaeadoueou. 5-9 June [995 EACI6/tNF/DOC.I ORIGINAL: ENGLISH SUMMARY REPORT OF THE EVALUATION OF THE IVERMECTIN DISTRIBUTION IN THE OCP - January 1995 (by Dr F.K. Wurapa, Dr M. Sylla & Dr T. Diarra) CONTENTS INTRODUCTION ...I OBJECTIVES .....2 METHODOFWORK ..,.2 THE SURVEY IN GUINEA . . . .3 THESURVEYINMALI ......5 THE SURVEY IN SENEGAL . . .7 RESULTS ....9 DISCUSSION ....I0 RECOMMENDATIONS.. .... T3 Introduction 1. In 1990 the Onchocerciasis Control Programme's (OCP) objective was modified to take into account the need to devolve activities to national govemments and assist them in maintaining the gains of the Programme. Accordingly the original objective was modified to read as follows: 'to eliminate orrchocerciasis as a disease of public health importance and as an obstacle to socioeconomic development throughout the Programme atea and for participating countries to maintain this achievement". From the beginning of the Programme large-scale vector control operations based on aerial application of insecticides remained the main suategy employed. 2. Constant scrutiny and evaluation of operatioru and the subsequent adjustment of programme activities have assisted the Programme to respond to change and make progress. Vector control in the original sevenrountry area, started during the years 1974-1979, bas now ceased with one or two exceptions in small circumscribed foci. Sirrce then there has been no resumption of transmission, as demonstrated by post-control entomological surveillance and epidemiological evaluation, despite the return of blackflies. Also, as demonstrated by the analysis of monitoring data, there has been no long-term effect on non- target aquatic fauna. 3. The availability of ivermectin in 1987 as a drug suitable for the treatment of onchocerciasis provided another useful tool for the control of the disease. In the extension areas where vector control was only instituted during the years 1987 to 1989/1990, larviciding combined with large-scale distribution of, and community-based treatrnent with ivermectin by differcnt modalities has improved the epidemiological situation more rapidly than'was the case in the original programme area where larviciding alone was carried out. 4. Initially priority was given to individuals who were at greatest risk of onchocercal eye disease and or blindress. Prcvious epidemiological mapping showed that such people lived in areas with community microfilarial load (CMFL) of l0 mf/s or more. In areas of the Programme wherc vector control has not been undertaken, e.g., in parts of the westcrn extension area, the limit above which large-scale treatment was carried out was lowercd to Smf/s. [-ater on the strategy was modified to include all at risk of onchocercal disease. 5. The new definition of the target population included all people living in onchocercal endemic areas in the target treaunent group. Full-scale ivermectin distribution in most of the Programme extension areas started in 1990. Some river basins had biannual treatments while others had yearly treiatment. Over 70% of these treatments were carried out by national teams with the support of OCP, using mainly the large-scale distribution approach. Recently several NGOs have been collaborating with endemic countries adopting passive and community-based disuibution methods. 6. Since the initiation of iverrnectin distribution prograrlmes direct and indirect measurcments of impact have periodically been undertaken. Ophthalmological evaluations have shown very definite improvements in the eye lesions after treatment. Also coverage of treatment has bcen measured as an indirect evaluation of impact of the prograrnme throughout the treatr[ent areas. Over 70% average coverage has been reported. The aspect N.05.95 ',2 of programme evaluation which has not been carried out so far is the performance of the operational modalities and the evaluation of reactions of the treated population so that progress as well as areas that may need improvement could be identified. It was therefore in the usual management style of constant evaluation of Programme activities that this aspect of the evaluation of the ivermectin distribution in the OCP was requested. Objectives of the Evaluation: 7. The general objective of the evaluation is to assess the operational modalities of the Ivermectin Distribution Programme in order to determine its progress and identify areas that may need improvement. Specific objectives of the evaluation are to : - estimate the degree of coverage with ivermectin treatment in the target population; identify the perceived effects of ivermectin treatment in the selected communities; assess the mamgement of the national programmes in terms of their sustainability in the future; identify problems in the way of efficient running of ivermectin distribution prograrunes at ole community level, the solution to which will improve national programmes. I\{ethod of work 8. In order to accomplish the above objectives, tfuee consultants were recruited to carry out the evaluation. The consultants are: Dr. F. Wurapa (Epidemiologist); Dr. M. Sylla (Public Health Specialist); Dr. T. Diarra (Social Scientist). From 12 to 16 December 1994 the three consultants met at the OCP Headquarters to develop a framework for the evaluation. Using the guidelines provided by OCP for the evaluation exercise, a protocol was developed for the evaluation taking into account the constraints of time and field requirements. 9. It was agreed to cover five villages in each basin in Guinea, Mali and Senegal. later on during the pretest of the questionnaires in Mali, it was found that the time required for a satisfactory administration of the questionnairc would not permit five villages to be covered in each of the three countries . It was therefore agreed that the number of villages to be surveyed in Guinea and Mali should be limited to three in each of the four basins while that in Senegal should be four villages in each of the three zones. The questionnaires prepared for the village chiefs, school masters or other key informants, head of household and members of his family were pretested in Mali from 26-29 December 1994 and revisions were made accordingly. 3collection phase should be responsible for organising the evaluation survey in one of the three selected countries. Thus Dr. Diarra was assigned the task of conducting the evaluation survey in Guinea. Dr. M. Sylla and Dr. F. Wurapa were responsible for the field phase of the evaluation in Mali and Senegal respectively. The Survey in Guinea Background ll. The ivermectin treatment in Guinea was started in 1988 in the Milo Niandan basin rvith a trial covering 2500 people carried out by the national team. [t was aimed not only ar assessing the ability to undertake the distribution and the feasibility of the treatment but also at seeing the reactions of the populations to the drug. 12. In 1989, the ivermectin treatment was extended to other basins in addition to thrt of the Milo Niandan, i.e., the Dion, the Sankarani and the Niger Mafou. In 1990, the Koulountou, Koundara, Kolomba, Fiye Sankarani and Tinkisso basins were covered by the reatment. From that date to 1994, the ivermectin treatment in Guinea had, progressively covered more villages, and morc people in the river basins, the number of which continues to increase. The national onchocerciasis control programme in Guinea has diversified its treatment experiments and strategy and joined forces with other bodies, particularly NGOs like Sight Savers, OPC and Philafricaine Suisse. l. Active treatment. advanced strategy 13. Active treatment has been carried out by the national coordination office of the onchocerciasis control programme in Guinea since 1989. It has seven teams composed of two medical officers, seven heatth technicians and seven drivers and has seven vehicles and seven mopeds. In addition to thesc national-coordination teams, the VCU teirms in the nine operational bases in Guinea carry out active treatment in villages located on their itineraries rvhich are hard to get to. The VCU teams in the nine operational bases treat 12,500 people per year. At present, the active treiatrnent covers thousands of people. 14. In 1994, the coordination teams treated 92,890 people in 943 villages in the Kolente, Niger Mafou, Sankarani Dion, Tinkisso and Bafing basins. This gives an idea of the activities because we could not get information on the treatrnent in all the basins. We could not get data on the Milo Niandan basin. 2. Communitv-based treatment 15. The trcatment by the communities themsclves is initiated in villages where Sight Savers and OPC work. The strategy is the same for the two NGOs. They use the human resourrces of the health system and villages and provide logistics for the treatment. They give training to the staff involved in the distribution: district coordinator, supplier. The latter trains the village distributors who are chosen by the elders. Sight Savers works in the Bafing basin, in the Mamou district, mainly in the Saramoussaya. Kenieko and Oure Kaba su6istricts. OPC works in the Fiye Sankarani basin, in seven subdistricts of the Mandiana district: Central Mandiana, Faralako, Kinieran. Nyantanina, Morodou, Koundian. We could not get complete data on the treatment activities of the NGOs. OPC treated 59,135 persons in 9i villagis in 1994 while Sight Savers treated 109 villages in the Bafing basin in 1994. 4basin, in the Mamou district, mainly in the Saramoussaya, Kenick<-r ancl Oure Kaba subdistricts. OPC works in the Fiye Sankarani basin, in seven subdistricts of the Mandiana district: Central Mandiana, Faralako, Kinieran, Nyantanina, Morodou, Kounclian. We could not get complete data on the treatment activities of the NGOs. OPC treated 59,135 persons in 97 villages in 1994 while Sight Savers treated 109 villages in the llafing basin in 1994, Village distributors have not yet been trained in all the villages in which these two NGOs rvork so the supplier carries out the distribution in those villages. J. Passive treatment 16. Passive treatment is carried out in all the nine operational bases which have a stock of mectizan for that purpose. These bases have a notebook for recording cases. According to thc Coordinator, the nine operational bases treat a total of about 12,000 people per year. At the Mamou operational base, 139 people were treated from 28 December 1993 ro l0 September 1994. Most of them are referred to the base by the district health centre. 17 . The greater part of the passive treatment is carried out by an NGO, Philafricaine suisse, based in Macenta. This NGO distributes mectizan in all the health posts in forest Guinea where people come to take the drug. It works in the Makona basin. We could not get data on the populations covered by Philafricaine suisse since the latest reports had not reached Kankan at the time of the survel'. 18. For the survey in Guinea, four basins were chosen according to the type of treatment and the body involved in the distribution activity. Thus, in the Bafing basin, the zone where Sight Savers operates was selected. This NGO has initiated a community-based treatment there. The Tinliisso and Niandan basins, where only the national team operates, were chosen. the fourth basin being that of the Fiye Sankarani where only OPC works. In fact, rvhile the national team undertakes treatment in the Sankarani basin also, OPC operates solely in the Fiye basin which was selected for the survey. 19. In each basin, the villages were selected at random at two levels. At the first level, either geographical entities (notably district) or operational entities of the treatment teams s,ere selected. At the second level, the villages for the survey were selected at random from thc list of villages of the selected entity having a population of at least 100 inhabitants. In each village, 20 families were selected at random from the exhausrive list of household heads. In cases in which the number of households was equal to or less than 20, all the households were selected for the survey. That was the case tbr the villages of Kelemassoya. Kenieko and Samamoudoula which had 11,20 and 12 households respectively at the time of rhe survel'. 20. In rhe villages selected, the village chief and a teacher, if there was one, rvere taken rhrough the questiomaire prepared tbr that purpose, the household head was intcrviewed using rhe household-head questionnaire while individuals aged more than five were intervierved using the individual questionnaire. In sonre cascs. the lttluseholcl head was absent on rhe day of the survey so a houschold-head questionnaire was not used. ll. Aparr trom thc questionniires. discussions wcrc hcld rvith ot't'icials of thc Ministry ol' Ilcalrh in Conukry, s'ith the Nationll Cottrdin:tt()r. \\'i(h the chicl's ol'tltc ()pcrational bitscs 5in the survey zones, with the exception of that of Kouroussa. Discussions were held also with the district health directors, the district coordinator of Mandiana, the Saramoussaya and Oure Kaba suppliers, the health workers of Ballagnoumaya and Noumandiana, and the distributors of Fode Hadj and lnila. There were discussions also with the village chiefs on the human potential that could be involved in the community-based distribution and on rcacrions to the treatment. Discussions were likewise held with the deputy national coordinator in Kankan, and with the Kankan EPI laboratory chief who was on a treatllrcnt mission to Kenieko. 22. The distribution notebooks were examined, in villages where there was a community- based distributor. Also the records and a report made in Kankan were reviewed. 23. After Conakry, the survey started first of all in the Bafing basin (Mamou district), rhcn it covercd the villages of the Tinkisso basin in the Diroguiraye district. Therc were discussions in Dabola before leaving for Diguiraye. The field work was then completed in the villages in the Niardan and Sankarani basins. The Survey in Mali 24. The ivermectin distribution programme in Mali for the control of onchocerciasis covers two very distinct zones: the original area and the extension area. 25. The original area, which covers 202,655 km2 of savanna with 5,078,721 inhabitants in the villages of the river basins of the Niger (right bank), the lotio, the Farako, the Wassoulou, the Banifing IV, the Konifarawa (Bani), the Kankelaba and the Baoule, has been under aerial larviciding since 1975. The original area entercd its devolution phase in August 1989, the prevalence, which was estimated at more than 60% with a blirdness rate of about 12%, having been reduced by 70-80%. Only the Baoule and Bagoe river basins are eligible for ivermectin treaunent when vector control operations ceased therc. 26. The extension area covers the villages in the river basins of the Niger (left bank), the Baoule, thc Bakoye, the Faleme, the Bafing, the Senegal, the Kolinkine ard the Terekole. They are villagei where 1,065,503 inhabitans live, in an arca of 146,E44 knf with blindness rates ranging between 3% and l0%. The whole of this extension area is under ivermectin treatrnent. 27. In Mali, a country which is at p.rcsent considered as an example, the three modes of ivermectin treatment are used, i.e., the "OCP-type' large-scale treatment, community-brsed treatment and passive treatnrent. Many bodies are also involved in the treatment in the field (OCP, NGOs, communities, etc.). 28. In 1994, 1900 villages were treated with ivermectin in eight river basins (Baoule, Bakoye, Bafrng, Niger, Bagoe I, Bagoe II, Faleme and Banifing II). 29. Out of a total of 512,581 persons treated, community-based treatment covered 340,344 persons, "OCP-type" large-scale treatment 171,131 persons and passive heatment I,106 persons. 630. , The "OCP-type' large-scale treatment was instituted in Mali in 1988 and is carried out by the national 'oncho' teams. Community-based distribution was initiated some three years ago and Mali serves at present as a model as regards this mode of ivermectin treatment. Community-based treatment is supported by non-governmental organizations whiph take part in the financing of activities related to training, the supply of ivermectin to the communities, the evaluation of the impact of the treatment and the supervision of ivermectin distribution in the villages. They are the NGO Sight Savers which supports self- treatment activities in 898 villages in the Kati, Bougouni, Kolondieba, Kangaba and Kouliboro districts, the NGO OPC (Organization for the Prevention of Blindness), which supports the activities in some 100 villages in the Bafoulabe, Kayes and Kenieba districts, the NGO Save the Children which operates at the Kolondieba centre, the Dutch Cooperation which supports the Kangaba district and CFAR (Centre for the Training of Female Extension Wo'rkers), an instirution under the Ministry of Health, which was responsible for 30 villages of the Kourouba district in the Kati region 31. For the survey to be conducted, a reasonable choice of four river basins was made with the EPI Chief in Ouagadougou, taking into account the different modes of treatment and the.presence of various actors. In Bamako, after discussions with the National Coordinator and the EPI Chief/Bamako, and on the basis of the lists available on the treated villages, a multi-cluster random sampling method was used. Thus, in the Niger basin where Sight Savers operates, out of 5l villages, each having more than 100 inhabitants, in the Kalaban- Coro district, the villages of Serwala, Djinconi and Siracoro-Ntekedou were chosen. In the Bafing basin where OPC works, among the 17 villages, each having more than 100 inhabitans, in the Koundian district, Sekotonding, Tiliba and Kabada were selected. On arrival at Manantali, it was observed that it would be risky to cross the lake because its level was quite high. Three villages were thereforc selected at random on the spot from the 29 villages having more than 100 inhabitants each, in the Bamofele district, to replace those of Koundian which were inaccessible. They were Ougoundigo, Goungoudala and Bantadioke. 32. In the Bale tI river basin, where only the national noncho" team operates, out of 6l villages, each having morc than 100 iphabitans, in the Sagabary district, Farabale, Magoabougou and Bandjougoubouria werc chosen. 33. Finally, in the western Baoule basin, where both the national "oncho" team and Sight Savers operate, out of the 62 villages each having more than 100 habitants, in the Siby district, Bedeya, Dinkou and Nioumamakana werc selected. It should be noted that Nioumamakana is under community-based treatmeut with support from Sight Savers while Bedeya and Dinkou are under the OCP-type large-scale treatment. 1 34t On 9 January 1995, discussions were held in Bamako with the Chief of the Prevention Division, the national coordinator and his deputy and the Regional Coordinator of Sight Sayers. The discussion guide on ivermectin treatment prcpared for the survey was followed in qthesc discussions. 35. [n view of the deadline, the mission headed by the consultant and comprising two interviewers who speak Bambara and Malink6 and thc driver left Bamako on l0 January 1995 tbr Manantali. I i 736. The field survey which was started on I l/l/95 in the Bamafele district, was continued in the Sagabary district, passing through the regional capital of Kita, then the Siby district and. finally the Kalaban-Coro district. The team returned to Bamako on 2l January 1995 in the morning. 37. On arrival in each district, the district chief, the health post chief and the health worker/supplier were immediately contacted and discussions held with them on the basis of the guide. In the villages, it was the village chief who took care of the mobilization of the household heads in the presence of whom a household list was established using the distriburion cards held by the village distributors and/or family cards (sometimes difficult to get because of taxes). In case the number of households was more than 20, a random selecrion of 20 households was made in front of all the household heads. This was done in the villages of Oungoudigo, Nioumamakana and Serwala. In the nine other villages, all the households were surveyed, their number being less than 20. The questionnaires were then administered, according to the forms, to the village chiefs, teachers (only two), household hcads, arxl household members present aged more than five. In the community-based treatment villages of Nioumamakana, Sirakoro, N'Tekedou, Serwala, Djinconi, Ougoundigo, Banradioke ard Goungoudala, discussions were held with the distributors on the basis of the guide., Everyday, the checking of the survey forms and their filing were done, taking advanrage of the interval between the village surveys. The second villages were always surveyed up to the night and even midnight, using lanterns and torchlights, a constraint due to the free time of the villagers and the quite short duration of the whole evaluation. 38. A halfway review with the two interviewers took place on 15 January 1995 at Kita and a review meeting on 20 January 1995 at Djinconi. The OCP/Bamako staff were given a brief report on the survey on 21 January 1995. The Malian authorities could not be met for debriefing since the day the team returned to Bamako was a Saturday. 39. On the whole, the survey went on smoothly with full support by the hcalth and administrative authorities and, particularly, the village chiefs, household heads and village distributors. The Survey in Senegal 40. Since 1988 ivermectin distribution in the two river basins of the Gambia and the Fal6md in Senegal has been carried out mainly by the national team with the support of the OCP. The national ivermectin treatment team is based at Tambacounda - the Regional headquarters of the endemic districts for onchocerciasis. The team consists of six technicians and two drivers headed by a medical officer and a senior technician. During field visits up to four local health suff are rccruited to assist the team. Two Toyota four-wheel drive double cabin pick-ups are available to the team. This team has been carrying out large- scate rreatment of ivermectin once a year in the Gambia and Fal6m6 basins since 1988 and in the Gambia basin twice a year since 1990. .1. A programme of community-based distribution has been started. The collaboration bctween ttre OpC and the national team in ivermectin distribution in 1995 is currently being planned. In order to select twelve villages for the evaluation, a multi-stage cluster random 8sampling method was used. For the large-scale ivermectin distribution programmes, one of the five teams in each basin was randomly selected. Then atl the villages treated by the team were serially numbered and five villages among the list were randomly selected for the evaluation. In villages with more than 20 households, a random sample of 20 from the village household list was taken. Where the number of households are less than 20, the whole village was included in the survey. Selection of households took place in seven villages while total village coverage was carried out in five viltages. 42. After discussions with the national authorities, it was found necessary to review the selected villages in which community-based treatment with ivermectin was being carried out. It was pointed out that the programme of ivermectin distribution in the four villages randomly selected before arrival in Dakar had acnrally not yet started a corununity-based distribution programme. [t therefore became necessary to repeat the selection of the five villages in which community-based distribution has been started. The selection was carried oud in Tambacounda from a list of 30 villages provided by the national field rcam leader" l 43'.t Following a random selection of five villages from the group of villages with community-based treatment, Senoude'bou, Naye, Sara Woura, Thies and Nionghani were picked to replace the original group. The other ten villages where large-scale treatmenr with ivermectin was being carried out remained the same i.e. Ngari, Thiobo, L:ngi Tanda, Dindefelo Tanda, Noumoufouga, Sansela, Magnaf6 and Demboli. 44. The evaluation team consisting of the consultant, the two interviewers and the driver left Dakar on 1l January for Tambacounda. Following courtesy call on the Regional Chief Medical Officer and the Governor, a detailed bricfing session was organised with the national ivermectin distribution team. 45. The survey was started on 12 January in Senoude'Bou in the northern Faldmd river basin. On entering every village the village chief was contacted to inform him of our mission. A village household list was obtained for sclecting households to be interviewed. Invariably the chiefs provided the team with one or two guides to assist the team. 46. The interviewers administered the thrce questionnaires always starting with the head of household and then systematically through the rcst of the members of the family. The questionnaire for the village chief, school master or other key informants in the village was completed as convenient to the respondents. In the absence of the head of household, an appropriate member of the family provided the information on who lived in the household. 47. The major problem encountered in the administration of the questionnaire was the inadequate advancc information about the evaluation exercise at the village level. This led to an increased number of absences in some households. This was particularly so in villages in the Fal6md basin wherc villagers have seasonal rcsidences, one in the main village and the other on the farm. The other contributing factor to the increased absences was the fact that most villagers were completing their harvest. In any case a participation rate of 80% was achieved. 48. As much as possible, all completed questionnaires were reviewed by the interviewers Ir the completion of each village before the forms were passed on to the supervisor for 9verification. The team moved down to the Gambian and lower Fal6m€ river valleys on 14 January. The selected villages in the two basins were covered using Kedougou as the base. The team completed the survey in the twelve villages and returned to Tambacounda on l8 January. On 19 January the authorities in Tambacounda werb debriefed about the field work and the team departed for Dakar in the afternoon. A similar debriefing of authorities in Dakar including the National programme coordinator, the Director of the Control of Communicable Diseases, the Director of Health Services, and the Director of the Cabinet. Results During the survey 49. 47 vitlage chiefs and other key informants were interviewed: 18 in Mali, t3 in Guinea and 16 in Senegal. On the basis of information provided by these informants, scales and height measuring apparatus were commonly used in the villages, as confirmed by 667o of the respordents. 76.6% of these respondents confirmed that the last treatment cycle took place within the eleven months prior to the current survey. It was reported Olat only 30% of members of their communities participated in ivermectin distribution activities. This indicates the inadbquate involvement of the communities in this activity. Only 55 % of the respondents werc told that ivermectin was used for treating orrchocerciasis. 50. In general most of these respondents (83%) did not list any major problems of the distribution programme. The only problem identified by 8.5% of respondents was the shortage of the tablets in the villages during treatrnent cycles. Almost all the respondents (98%) statd that they would like the ivermectin programme to continue. When ttrc 47 resp<indents werc asked whether they knew any neighbouring villages that were not treated, 18 of them gave names of 40 such villages - 24 in Guinea, 4 in Mali and 12 in Senegal. 51. 541 heads of household werc interviewed: 203 in Guinea, 153 in Mali and 185 in Senegal. Orny 52% of ttre heads of family knew about the ueaunent of onchocerciasis with ivermectin. 85.6% of them stated that they were treated by the visiting teams. 5O.7% of respondents rcported drug reactions following the first treatrnent while 16% claimed they expcrienced zuch reactions aftcr the last treatment. The most common side-rcaction was troublesome itching. 86% of the respondents stated that they were satisfied with the treatment prcgrarnme. But t heads of household said they did not find the treatment beneficial. 52. There were 4106 household members interviewed during the survey. The participation rate in the survey was 78.7% (41614932). Out of 4106 respondents, 3184 (77.5%) confirmed that they had been treated with ivermectin. A comparison of treatment coverage with the type of Ueatrnent showed that in the total surveyed population the overall treatment coverage by large-scale treatment was 80.8% while the coverage by community- based tneatrnent was 74.5%. 53. Only 21.4% indicated that they had never taken ivermectin and 1.07o did not respond ro ttre question. 879 persons indicated that they did not receive treatment during any of the treatment cycles and only 19 later received their treatment. The follow-up rate of 2% l0 (19/879) is low. Out of the 19 treated tater, 14 (73.7%) were treated in the community- based treatment programme while five (26.3To) were treated in the large-scale treatment programme. There was a sutistically significant difference between the proportion treated by the two methods of distribution. The ivermectin treatment coverage rate from this survey compares favourably with the over 7O7o coverage ra(e reported by OCP. The ivermectin coverage rate in Guinea was 66.6%. The coverage rate in Mali and Senegal were77.8% and 85.9% respectively. 54. 63.9% of respondents confirmed having swallowed the ivermectin tablets at the treatment site. The dose of ivermectin taken varied: 5.8% of respondents took one half tablet; L9.4% took one tablet; 17.7% took one and a half tablets and23.47o received two ivermectin tables. There was a higher proportion of individuals that swallowed the tables at the treatment site in the large-scale treatment progranune (67%) as comparedto 6l% in the communiry-based treatment programme. 55. Overall, ivermectin appears to be well tolerated by the treated population. Out of 3184 treated individuals ,2,13L (68.8%) had no secondary reaction following their treatment. The main reaction of significance recorded was troublesome itching which was reported in L4.5% of the treated population. Other minor complaints after taking ivermectin included abdominal pains, sleeplessness and dizziness. When asked whether respondents would be wilting to take the next ivermectin treatment 65.8% confirmed their willingness to continue. Only I .8% of the surveyed population indicated that they would not like to take the drug. The main reasons given for not uking the drug were that they did not have onchocerciasis and did not like swallowing tablets. The follow-up of untreated individuals by method of treatment, i.e., large-scale treatment and community-based treatment was very interesting. Overall, it was found that while only 19.5% of individuals not treated in the large-scale treatment zone were satisfactorily followed up, the proportion in the community-based treatment zone was 52.3%. The difference between these two systems was statistically signiticant. 56. Among the individuals who did not receive full treatment at all treatment cycles 5l % was due to their absence, 8% was due to pregnancy and9.8% was due age exclusion. Only 2% of the population refused treatment indicating that rhe treatment programme was well accepted. 57. Also among the individuals who had never received treatment 46% of such non- treatment was due to absences, LlVo was due to pregnancy, LL% was due to age exclusion and l0% was due to the fact that respondents lacked information on the programme. Only 3% of the non-treated group was due to outright refusal. 58. When the non-trearment was analysed according to the type of treatment, it was tbund that there was a higher non-participation rate, 61% in the large-scale ireatment as compared to 35% in the community-based treaunenl programme. Discussion 59. Considering the loltowing issues. findings of thc study. attention should bc drawn ttt the 11 60. One of the important observations from tlre data is that the monitoring of the trqrtment by one ana tne same individual is only partial. In fact, in any village where ivermectin treameff has been carried out no multiple cycles of treatment has auained full coverage. This situation is more serious in the case of passive treatment. The,laa obtained in tbe surdy is only a rorgh estimate since this question depended on recall aod in some basins the treatmeffs start€d over five years ago. Although the treatment is an important event, the number of treatments in which those surveyed had participated could be difficult to remember if therc had been many treameffs. On &at assumption, among those who did not answer tbe question there would be a grearer mrmber of individuals who did not participate in all the trreahenis and thercfore the uumber of persons partially trreated would be even greuitar. The facts call'for a morc effective system geared to increased participation of indivirtuals in all tbe treaments for it is not the fact of being treated once or twicg whi-cl matters hr that of being treated regularly over a given period. 61. The village chiefs have good hwledge of and inforrration on tbe neighbouring villages since therc are family and matrimonial relations between the villages. There are morc uotreated neigtrboring villages than the villages srrveyed. However, it is possible that ore particular vi[age may have been mentioned by two viltags chiefs if the villagcs zurveyed are not very far from each other. Under these conditions, the number of untreated neighbouring villages could be less. Nevertheless, tbe places wherc those absent were at the time of heatmeDt" often in the same zone, give additional information. If they arc taken into a@ount, they would ircrease the number of untreated villages, if we have inforrnation on the siuration in those villages with regard to the treatment. Thus, the number of untreated reighbouring villages is a retative irdication because other parameters must be taken into account, ttre mosp important of which is the level of infonnation which tbe village chiefs in the survey had. 62- Tb results,of tfu survey havehighlighted several important problems in the planning and organisation as well as implementation, monitoring and evaluation of ivermectin distribution activities- A major wealcness of current programmes regardless of the method of ivermectin distribution is the lack of information on the prograrnmes among the affected communities tbeir iustifrcatioa arrd the need for the participation of the communities in -thc planning an4'exciution of the progrtrrmes. An effort to irrcrease information, irducation and communication on the important endemic diseases including onctrocerciasis in tbese communities would contribute immensely to the success of the programmes. 63. A major reason for non-participation in the trea&nent programmes was absence from the village duriqg treatmeqt. Some of these absences were due to travel outiide ttre village for personal and family Feasons. Others were due to abeence of whole families living in their seasonal-residence homes on their farms. [t was clear from discussion with villagers that provided information on the schedule of treatment were available, such families could be inforrred to be present for large-scale trcatments. In the case of community-based treatrnent such seasonal rcsidents wotrld need to be identified and included in the prograrnme. &. It will be important in future development of the programme that a critical assessment of the advantages and disadvantages of the various treatment methods available should be made. The choice of a method should be guided by the efficiency and the cost effectiveness tz of thc method. There has been anrple demonstration from the results of the survey that contnruniry-based distribution programmes meet these important criteria. 65' As regards the performance or reliable indices on the monitoring of the rreatment, the intbrrttation obtained was of a high quality because atnrost all those surveyed answerecl the question on cover:rge. While mobile team treatment rate is higher than tommunity-basecl treatment rate, the bcst perfonnance or efficacy index should be measured more by the treatnrent monitoring rate. This is why despitc its lower coverage rate, community-base<l treatnlent is in our opinion the most effective system. The shortcoming of the raw coverage rate is that it is diachronic. It only gives the situation for a given peiioO. The monitoriig rate. on the other hand, which is synchronic, makes it possible to analysc the evolving situarion and takes into account all constraints and events which could have an influence at any time. 66. Another fact concerning the coverage which this study underscores is its assessment by the number of tablets given. In fact it happened that the distribution has an effect orr the real coverage because some tablets given to some of the villagers werc still in the households u'hilc their recipients appeared de facto as having been trcated. Under these conditions, the cl'ttctiveness of tlte treatntent remains a serious problem. One can therefore ask whether solne coverage rates reported by the different teams are not anticipated coverage rates calculated on tltc basis of having succeeded in placing ivermectin in the fiel6 and thcrefore covcrage ratcs envisaged or sirnulated on the basis of this assumption. P.rtrsnects 67. Thc survey carricd out in the threc countries - Guinea, Mali an4 Senegal - in I I river b;rsins had :r good participation rate. 68' The no-response rate was negligible. However, it should be noted that although the overull non participation ratc is low, it con'ceals.differences between the countries concerned. For example, the covemge rate per ivennectiri ircatment proved to be very high (in generat) during the survey but it varied in the ihree countries studied. It is -theiefore iasy to understand why although the survey resuls arc very useful because of the informition obtained on the dift'erent aspects of the ivermectin treatment programme, they cannot be generalized as such to the other OCP Participating Countries. This evaluarion merhodologyis therefore an approach which must be applied in each country in order to define the problems in each country pr,pgramme. .The evaluation team therefore recommends that OCp organize a similar evaluation .g[,.the ivermg.ctin.treatment programme in r]re other eight Pa rt ic ipating Countries. ,1 ,,iRJ .-i1r;J - ' ,,i, , l3 Recommendations l. There is a need to increase information and education of affected communities about the importance of prevention of onchocerciasis through community-based treatment with ivermectin. 2. Advantage should be taken of the considerable goodwill and enthusiasm in the affecred communities in ivermectin treatment by involving them in the design and implementation of ivermectin distribution in order to irrcrease the success and sustainability of the programme. 3- [n line with the national devolution plan, the health centre staff should become actively involved in the education of affected communities on the prevention of the selected priority endemic diseases in their catchment areas. 4. ' ," Therc is a rped to increase rcsponsibility of the health centre and hospital staff for the orientation and supervision of community trealth workers involved in the ivermectin d istribution prograrune. 5. As the ivermectin distribution programme becomes fully operational without the financial ard logistic support of OCP, it will be important to channel available support from orher partners into critical programme activities such as support to improved supervision of community health workers. 6. In order to accelerate the coverage of ivcrmectin distribution, it will be necessary to increase the mobile outreach distribution prografirme of the health centres and the Communicabte Disease teams. This component of ttre programme could be combined with the current multidisease surveillarrce activities being planned. 7. It is important that even in itre large-scale trcatment programmes tablets of ivermectin should be left in charge of appropriate persons for the treatment of absenteix and other persons who carmot take the drug at the time of distribution because of pregnancy or illness. 8. .. Supervision of ivermectin distribution activities at the community level needs to be improved particularly,in the correct keeping of simple records on the work done and on individuals excluded from treatment. 9. Community-based treatment appears to be the method of distribution most appropriate in terms of effective coverage and follow-up as well as cost-effectiveness and every effort should be made to encourage it. 10. Based on our experierrce in carrying out this evaluation, the positive reaction of the authorities in the three evaluated programmes and the findings of this evaluation, we would tike to strongly recommend that similar evaluation should be carried out in the other cight national prograrnmes.