In Rural Ugandan Communities the Traditional Kinship/Clan System Is Vital to the Success and Sustainment of the African Programme for Onchocerciasis Control

Annals of Tropical Medicine & Parasitology, Vol. 94, No. 5, 485± 495 (2000)

In rural Ugandan communities the traditional kinship/clan system is vital to the success and sustainment of the African Programme for Onchocerciasis Control

BY N. M. KATABARWA* Carter Center, Global 2000 River Blindness Program, P.O. Box 12027, Kampala, Uganda

F. O. RICHARDS JR Global 2000 River Blindness Program, Carter Center, One Copenhill, Atlanta, GA 30307, U.S.A.

AND R. NDYOMUGYENYI Ministry of Health, P.O. Box 1661, Kampala, Uganda Received 5 April 2000, Revised 5 May 2000, Accepted 8 May 2000

In rural Ugandan communities where onchocerciasis is meso- or hyper-endemic, control of the disease is now being carried out using a strategy of community-directed programmes for the annual distribution of ivermectin to all persons eligible to take the drug. For these programmes to achieve their annual target coverage of at least 90% of the population eligible to take ivermectin, and to continue to sustain themselves for 10± 15 years or more, even after external donor funding ceases, it has been found essential to replace the initial community-based strategy, imposed from outside, by a community-directed strategy developed by the community members themselves. Furthermore, it is essential for success that full use be made of the traditional social system, which is very strong in all rural communities in Uganda. This system is based on patrilineal kinships and clans, governed by traditional law, and in it women pay an important role. If this system is ignored or by-passed by government health personnel or by the sponsors and promoters of the programme, the communities are likely to fail to reach their targets. When rural communities increase in size and complexity, following development and the arrival of migrant families, they become semi-urbanized. The kinship/clan system is then weakened, community- directed drug distribution is much more dif® cult to organize, and coverage targets are not often achieved. This effect is of minor importance in a rural disease, such as onchocerciasis, but is likely to be of greater signi® cance in the control of diseases, such as tuberculosis and lymphatic ® lariasis, which thrive in urban environments.

Several programmes of mass chemotherapy to ship: national health services participating control major parasitic and infectious diseases with international donor agencies, non- in developing, tropical countries are now be- governmental development organizations ing supported by the World Bank and exe- (NGDO), and major pharmaceutical compa- cuted by the World Health Organization in nies. Most of these programmes are based on partnership with the Ministries of Health in the fundamental, ® nal common pathway of the affected countries. An important feature of community participation and involvement this effort is that of public/private partner- (Katabarwa et al., 1999a). One of the most successful of these pro- * E-mail: [email protected]; fax: 1 256 41 250376. grammes is the use of ivermectin (as Mectizan

ISSN 0003-4983 (print) ISSN 1364-8594 (online)/00/050485-11 Ó 2000 Liverpool School of Tropical Medicine Carfax Publishing 486 KATABARWA ET AL. donated by Merck & Co., Inc.) to control personnel, from the government health-care onchocerciasis, with its associated skin lesions services, the NGDO concerned and the inter- and `river blindness’. In Uganda this pro- national health-care agencies that were programme started in 1992, as a co-operative moting onchocerciasis control, initially took venture between the Ugandan Ministry of very little account of the social structures in Health and three NGDO, namely the River Ugandan rural communities, or the skills Blindness Foundation (now the Global 2000 necessary to stimulate the involvement of River Blindness Program), and, to a lesser these communities in such a health pro- extent, the Christoffel Blinden Mission and gramme. Trained only in `modern’ or `west- Sight Savers International. Since 1997 the ern’ health-delivery systems, they were often efforts of these organizations have received reluctant to allow community members to additional support from the World Bank/ make decisions regarding the design and im- World Health Organization African Pro- plementation of programmes (Foster, 1987) gramme for Onchocerciasis Control (APOC), and, by portraying themselves as being in which is now active in a number of African charge, they reduced the capacity of com- countries where onchocerciasis is endemic, in- munity members to assume the ownership of cluding Uganda. The aim of this programme their programmes or to integrate control mea- is to control onchocerciasis by means of an- sures with their day-to-day activities. Some- nual mass distributions of ivermectin to all times they would also impose certain demands communities where human infection with on community members without considering Onchocerca volvulus is meso- or hyper- workable alternatives. It therefore came about endemic. Since control depends largely on the that, in many places, the programme, although ability of ivermectin to reduce and ultimately being nominally community-based, in fact to interrupt transmission, by killing mi- involved little more than a request for `com- cro® lariae and by interfering with their em- munity tolerance’ or `community compliance’, bryonic development, treatment must with little or no decision-making or responsi- continue for 10± 15 years or more in order to bility being devolved onto community mem- achieve success and eradicate the adult worms. bers. The failure of many community-based health programmes to become self-sustaining THE EARLY STAGES OF THE has also been attributed to the limited knowl- IVERMECTIN DISTRIBUTION edge of the programme staff, their consultants PROGRAMME IN UGANDA AND THE (whether expatriate or locally trained), and MISTAKES MADE their donors, concerning the community members’ culture. The tendency is to rate the The original strategy of the programme was to local culture as irrelevant or backward to the invoke community participation and to ap- development process, when compared with point and train community-based distributors `western’ knowledge and technology (Mander- (CBD) to distribute the ivermectin to the son, 1998). In the absence of any trained inhabitants of the affected communities. The anthropological or other experienced social- `annual treatment objective’ (ATO) was science expertise, no proper account is taken de® ned as the treatment of all persons in the of the depth and validity that existing tra- community who were eligible to take iver- ditional systems offer to the structuring of mectin. A coverage of at least 90% of the rural communities ATO within a period or 2 months each year In many rural communities in Uganda, it was considered to be the level that would need was observed that individual CBD were as- to be achieved if the programme were to be a signed various duties that occupied much of success. their time and yet were supposed to be under- Unfortunately, as is often the case with taken on a voluntary basis. Not unnaturally, such programmes (Stone, 1992), the relevant such CBD often asked for some remuneration KINSHIPS AND ONCHOCERCIASIS CONTROL 487 to compensate for their loss of time. Where the the ivermectin-distribution programme in programme refused such remuneration, the Uganda, which now receives support from CBD often dropped out of the programme; APOC, has adopted and implemented the where the programme did provide remuner- strategy of `community-directed treatment ation, the CBD tended to become more ac- with ivermectin’ (CDTI). This strategy, which countable to the programme sponsors and staff is described below, is leading to much higher than to their own community folk. As a result, community success rates and is raising good their performance did not reach, let alone prospects for continued, long-term, self- maintain, the expected level. sustainment of the programme. In rural Ugan- Some communities did indeed appoint their dan communities the strategy depends greatly own health workers, one per community or per on the understanding of the community `kin- parish, who were paid by the programme ship and clan system’, which is also described organizers to carry out the ivermectin distri- below. bution. In these circumstances, when external donor funding ends, as almost inevitably it The Roles of the National and of the will, it is likely that the CBD will lose their District Government Health Services in bene® ts. They will then stop working, the the Ugandan CDTI Programmes programme will collapse and, in all probability, At the national level, the Ministry of Health is the blame will be put unjustly upon the com- responsible for ordering supplies of ivermectin munity members. (donated free by the Mectizan Donation Pro- Where the externally supported, health-care gramme of Merck & Co., Inc.), for its duty- programmes lured or even coerced the com- free importation and for its distribution to the munity members to provide monetary incen- districts. Continued support from donors and tives to the CBD, so as to sustain the activity government health services at the national and after the donors have quit, the results have district levels is essential for the sustainment of been disastrous. Some sections of the com- the programme and the community-directed munity never contributed towards monetary distribution process. incentives while enjoying the bene® ts of the At district level in the CDTI programme, programmes. On the other hand, those who did the district health authorities are responsible contribute often received poor or no services. for: In addition, those CBD who received monetary incentives were usually selected only by certain (1) initiating the annual mobilization and sections of the community and hence were not health education of the communities; trusted by other sections. Therefore, the con- (2) training the community-directed distribu- tribution of monetary incentives to CBD was tors (CDD), who have been selected by perceived as a loss by those sections of the the community members; community that had not taken part in the (3) delivering supplies of ivermectin and selection of the CBD. This state of affairs other medicaments to ® xed points, from created mistrust and animosity within the com- which the communities can collect them; munity, and resulted in divisions that were (4) providing advice and support in the man- very dif® cult to heal and which hampered agement of any severe adverse reactions community involvement in the health-care associated with the ® rst-time therapy of programme. patients with onchocerciasis; (5) collecting reports from the CDD at the end of each distribution exercise; THE CHANGE TO COMMUNITY- (6) analysing data, writing reports and pro- DIRECTED TREATMENT WITH viding feedback to the communities; IVERMECTIN (7) accounting for the resources received from government or donors; and Since 1998, and learning from past mistakes, (8) providing, at appropriate and agreed times 488 KATABARWA ET AL.

and places, the supporting resources, skills directed drug distributors (CDD) and treat- and services that are absent at the com- ment centres. Members of the programme munity level. staff then tell the selected distributors and the community leaders how to store the ivermectin safely, how to determine dosage, Implementing CDTI in the how to manage adverse side reactions, how to Communities and Integrating it with keep proper records, and how to prepare the Community Agenda reports. The communities are then left to The objective of promoting community direc- organize their own distribution exercises. In tion is to render the ivermectin-distribution Uganda, a community, once prepared in this programmes capable of being sustained by the way and allowed to plan and implement its community members at their level. To put own CDTI, almost always achieved and sus- this into effect successfully demands a sound tained the desired coverage of 90% of the understanding of the cultural factors that eligible target population (Katabarwa and in¯ uence the involvement of community Mutabazi, 1998). members in health-care programmes. Account The approach to the communities targeted needs to be taken of vital community aspects, for CDTI starts with meetings in the com- such as the social structures, legal systems, munity, to explain the purpose and the strat- resource mobilization, and sharing systems. egy of treatment. Success depends on meeting The concept of community-directed treat- with groups of a signi® cant number or a ment with ivermectin has been developed to `critical mass’ of community members and replace the vaguer term: community-based their leaders, in order to inform them about treatment programmes. The latter, in the con- onchocerciasis and its control and the need for text of Ugandan onchocerciasis control, were community-directed activities to be adopted. usually associated with inadequacy or failure. The CDTI strategy includes the community The newer, CDTI strategy involves searching assuming responsibility for the following: for the correct and appropriate information that can be used to maximize community (1) selecting their own CDD, who must be involvement, both in decision-making and in members of the community, and having the assignment of appropriate programme re- them trained by the district health author- sponsibilities to community members for the ities at a time and venue chosen by the betterment of their own health. The results of community; a multi-country study of CDTI for onchocer- (2) deciding whether the distribution shall be ciasis control (WHO, 1996) and of the work of from house to house or from a central site Katabarwa et al. (1999b) in Uganda have re- and, if the latter, choosing its location; vealed that communities are better able to (3) deciding how many CDD there shall be, achieve their target coverage when the com- which sections of the community each munity members themselves actually make the shall cover, when the distribution shall decisions as to how the programme should be take place, and how the CDD shall be organized within the community. supported; The CDTI programme functions in the (4) assuming responsibilities for collecting following way. The district health personnel ivermectin from a central place not very ® rst explain the purpose, principles, and far from the community; bene® ts of the programme to the communities (5) safely storing a supply of ivermectin for by means of participatory health education. the subsequent treatment of absentees and That done, the communities are then empow- non-eligible individuals (such as the preg- ered to make all the local management deci- nant or sick) who could not take the drug sions and carry out the treatment, without at the time of mass treatment in the com- external interference. The community mem- munity; bers ® rst select their own, community- (6) recognizing the rare severe adverse reac- KINSHIPS AND ONCHOCERCIASIS CONTROL 489

tions and referring individuals suffering bers whom they were to treat had to be from them to local health authorities; and short, so that the task of distribution (7) changing the treatment approach if it is could be ® tted in conveniently with the found to be unsuitable after the ® rst daily chores of the CDD and other com- round of treatment. munity members. (4) When the areas of distribution of iver- When attempting to integrate a health-care mectin were divided along the lines of programme into the community, there are kinship or of the traditional social-support certain issues that the organizers of donor- groups, known as engozi (Katabarwa, supported and government-sponsored pro- 1999), the communities achieved their grammes, and the health personnel employed target coverage and sustained the pro- by them, must understand if they are to obtain gramme from year to year. By contrast, their desired objectives. Chief among these are when they followed the demarcation of the following: communities by local government of® cials (1) Provision of correct information to com- using government-determined village munity members. boundaries, the result was a failure to (2) Encouraging the community members to reach the annual coverage target. meet, discuss and take decisions which These processes allow the community affect their performance. members to become stakeholders in the CDTI (3) Understanding the social structures (e.g. and to see themselves as partners in health- kinship/clan groups) and cultural systems care delivery. The communities investigated (e.g. social codes) of the communities by Katabarwa et al. (1999b) are in four dis- which, in turn, involves close contact and tricts of Uganda: Adjumani, Moyo and Nebbi continued dialogue with community in the north± west and Kisoro in the south± members. west. These communities meet when necess- (4) Prioritization of health needs and inter- ary to identify their responsibilities, solve ventions. problems and take decisions needed to achieve (5) Obtaining the trust of community mem- their objectives (Table 1), and they maintain bers. and enjoy the bene® ts that accrue from the In the Ugandan CDTI programmes, it was successful implementation of what they can found that there were several other important truly come to regard as their own programme. considerations (Katabarwa and Mutabazi, The mean coverages achieved by these com- 1998): munities, as proportions of the eligible population, increased signi® cantly when the (1) Having more than one trained, com- community-based strategy of ivermectin dis- munity-selected CDD/50± 100 persons in- tribution in 1997 was replaced with a CDTI in creased the chance of good integration. 1998 (83.8% v. 93.8%; P 5 0.02). This helped the CDD to accomplish their duties within an acceptable time, without affecting their domestic chores and, at the same time, it encouraged delegation of the THE IMPORTANCE OF THE KINSHIP/ work. CLAN SYSTEMS IN THE SOCIAL AND (2) The selection of CDD from within their CULTURAL STRUCTURES OF RURAL kinship/clan groups greatly increased the UGANDAN COMMUNITIES AND ITS acceptability of the health-care pro- RELATIONSHIP TO HEALTH-CARE gramme, as well as facilitating the mobi- PROGRAMMES lization of other group members and in¯ uencing their compliance. In all rural communities in Uganda, patrilineal (3) The distance between the homes of the kinshipÐ the successive links between the CDD and those of the community mem- male parent and his childrenÐ is the most 490 KATABARWA ET AL.

TABLE 1 Mean treatment coverages in the meso- or hyper-endemic communities of four Ugandan districts, using community-based treatments (CBT) in 1997 and community-directed treatments (CDT) in 1998

Treatment coverage (%)

District No. of communities 1997 1998

Adjumani 79 86 97 Kabale 26 89 96 Kisoro 31 84 86 Nebbi 637 76 96 All four 773 83.8 93.8

basic structure organizing individuals into so- is based on descent from a common ancestor, cial groups. It is within and through these the land is generally held collectively. Where structures that marriage, property, inherit- family homesteads are scattered, the land be- ance, and community welfare of the social longs to individual families, but it can be system are organized. Community loyalties subdivided and handed over to sons, especially still divide along kinship lines, rather than in after they marry. However, under this tra- accordance with political or administrative ditional, family-homestead culture, the right dictates. The typical kinship group, or clan, to sell a family’s land to outsiders does not comprises 50± 100 persons, depending on the rest solely with the individual owner. The numbers of wives taken by its male members other kinsmen may refuse to allow the sale or and the children produced by these marriages. demand that they have the ® rst chance to As kinship groups grow, they may split (along purchase it before an outsider is allowed to matrilineal lines in polygamous families, or buy the land. Thus, in most districts, out- when brothers separate as a result of disputes siders do not have easy access to lands already or in search of more land for their children), occupied by the kinship group or clans. In thus forming sub-clans which still maintain some cases, a father’s gift of land to his their allegiance to the original family clan in married son only carries the right to cultivate the event of any external threat. In the north- and not the right to sell, and most of the ern and central districts of Uganda, families cultivated land is considered as belonging to tend to be organized into homesteads that are the clan. clustered relatively close together (i.e. within a In rural communities, kinship groups may radius of 50± 100 m). In the rest of the coun- also have a much wider array of functions. try, homesteads are mainly scattered, with the They often serve as the basic units for pro- exception of Kabale district, where clustering duction and distribution of produce, storage of is observed. In every case, groups of home- cultural, technical and `magical’ knowledge, steads tend to belong to close relatives who religious cults that worship spiritual beings can trace their descent from one male individ- (who are themselves considered as members of ual and who belong to the same clan or the kinship group), and even political repre- sub-clan, with the exception of those women sentation. It is the kinsmen who ensure that from other clans who have married into these their elders and the sick are looked after, that patrilineal families. mothers and children are protected and pro- In areas where family homesteads are lo- vided for, and that there is enough labour to cated in kinship clusters, whose membership produce food for the community. This `social KINSHIPS AND ONCHOCERCIASIS CONTROL 491 safety net’ is admirably exempli® ed by the Service to one’s own kinship group falls traditional, social-support systems, known as under the division-of-labour and distribution- the engozi in south± western Uganda of-wealth functions of the traditional system. (Katabarwa, 1999). If such service is decided by the traditional It follows that, in rural Uganda, a sound kinship institution, it is given happily without knowledge of the role of kinship is essential if question, and without negotiating `incentives’, one is to understand the social dynamics of for it is the means of survival of the kinship any community and the way in which these members. In contrast, the demanding of mon- will in¯ uence the acceptability, management, etary incentives for services rendered to those sustainability and ultimate success or failure of outside one’s group is quite in order, and is, any community-directed, health-care pro- indeed, encouraged. Interestingly, however, gramme. To date, rural health programmes items (such as food, drink, or labour in the have not taken the kinship issue seriously or home or on the farm) that are provided in the even bothered to consider its importance in course of communal service are considered as health-care delivery. As a result, it is not the due rights of anyone who provides this surprising that community members’ apparent service. Their role is to reduce pressure on the `refusal’ to participate fully in these well-in- individual providing the service, while at the tended programmes has frustrated both same time strengthening kinship ties with the government and donor-supported health pro- recipient of the service. grammes. In their frustration, the programme Avoidance of kinship duties, or refusal to personnel usually give up truly meaningful recognize kinship-authority institutions, is a attempts to involve the community. Instead, serious offence, with dire, even draconian, they `hire’ one or two community members consequences. For example, if a member of and pay them to accomplish the necessary the engozi social-support system does not re- tasks. As a short-term remedy, this may well spond when requested to carry a patient to provide donors with quick results, but it is a hospital, he could face a ® ne equivalent to policy that is incompatible with long-term U.S.$5.00 or 20 litres of local beer. If he failed sustainability of the health interventions. to pay this ® ne, it would be increased after 1 day to U.S.$20.00. If he still refused to pay, he and his family would face the wrath of his kinsmen, and could be denied communal THE IMPORTANCE OF SOCIAL LEGAL labour in their ® elds or even be banished from SYSTEMS the community. Women who marry into or belong to a In the rural Ugandan communities there are kinship group are required to care for the two legal systems. There is the common law, children of other mothers who are sick, tend which was established ® rst by the colonial their crops, prepare food for the sick, and to government and is now enforced by the local- provide food at burial ceremonies. Any government structures. The common law ap- woman refusing to perform these duties with- plies to everyone in the community. The out good reason would be branded a witch and second system is the traditional legal system, thus, according to traditional belief, as a per- enforced by clans and kinship groups. These son who will bring bad omen and death to her traditional systems have ensured the survival kinship group or clan. The consequences of of the communities throughout the centuries this are dire. She might no longer bene® t from and retain a great in¯ uence in society. Their community labour in her ® elds, or she might codes govern, among other things, gover- be denied carriage to hospital if she fell sick. nance, resolution of con¯ ict, the acquisition The other womenfolk might withdraw their and distribution of wealth and land, distri- emotional support from her, and the resultant bution of labour, care of the sick and children, loss of face may even lead her husband to and choice of marriage partners. marry another woman. 492 KATABARWA ET AL.

KINSHIP FACTORS IN COMMUNITY- a result of the cost associated with its distri- DIRECTED TREATMENT WITH bution. IVERMECTIN Another reason why community members refuse to pay monetary incentives to externally Incentives appointed CBD lies in a cultural phenomenon In Uganda, the mean per-capita income is less in rural Uganda, known as the levelling mech- than U.S.$200/year and the per-capita anism. Basically this means that when a mem- government allocation to health services is ber of a kinship or a clan acquires wealth, he , U.S.$10/year (Anon., 1999). Disease pre- is expected to share it with his kinsmen for the vention and control is costly and can only be greater good of them all. On this basis, the afforded if everyone in each community col- kinship members would steadfastly refuse to laborates and contributes, not only to the costs pay any monetary incentives to a CBD who of the user fees in government- or missionary- was not a member of their clan, for any money sponsored clinics, but also towards the intra- that he received, whether from them or from community management of control the promoters of the programme, would be programmes, such as CDTI. Although iver- lost to the kinship and/or clan. In contrast, mectin is provided free of charge to the Ugan- when a CDD was a member of the clan, he dan onchocerciasis-control programme, its would be bound to serve the clan free of distribution costs (transport, `per diems’ for charge, but all would understand that, if the staff, time spent and labour lost in childcare CDD did receive any money from the external and the gardens etc) can only be affordable by programme organizers, it (at least in part) these generally poor communities through uti- would eventually `trickle down’ from the lization of the kinship system. Utilization of CDD, under the levelling mechanism, to the the kinship/clan system, and its associated general bene® t of his kinship/clan. traditional laws, greatly facilitates mobilization of the communities for resource sharing in the Acceptance of Tablets from Kinsmen course of the control programme. The fear of witchcraft or poison being admin- Katabarwa et al. (1999b) observed that the istered along with the medicine that is being members of those communities with success- handed out by an outsider is often suf® cient ful CDTI programmes (that achieved and sus- grounds for refusing to take the ivermectin tained their annual target coverage) had that is being offered. As one woman in Kisoro usually selected their kinsmen as CDD for district put it: `Suppose they put something distributing the ivermectin. Those communi- harmful in this medicine. I can’t give a chance ties that relied upon externally appointed to the devil where my life and those of my CBD (i.e. CBD who are not relatives of the family members are involved. I will only get community members) generally failed to meet medicine from the community members I the desired treatment coverage. In most com- know and trust’. She trusts mainly those re- munities, externally appointed CBD could, lated to her in the community and her hus- with impunity, demand monetary or other band; that is the `world’ she knows and material incentives (e.g. umbrellas, boots, T- understands. Traditional clan or kinship- shirts, coats, bags etc) from all those who were group leaders, or the leaders of local engozi not their kith and kin, and they would often systems, are usually able to sort out such withhold their services if these were not forth- problems among their followers and persuade coming. The CBD were not constrained by them to accept the tablets. They are also of the social code that governs the behaviour of great help in identifying patients with severe kinsmen, and once monetary incentives were adverse reactions to ivermectin, counselling given, the demand for more was triggered, them and the community about why the side- and a cycle of alienation of the community effects have occurred, and helping those affec- from the programme continued until treat- ted to reach health centres, when necessary. In ment with the `free medicine’ was withheld as this way, the traditional system prevents the KINSHIPS AND ONCHOCERCIASIS CONTROL 493 rare adverse effects of treatment from escalat- level interfered with community decision- ing into a general rejection of the treatment making, such as the selection of CDD and programme. treatment centres, this coverage was never achieved. Role of Gender Dif® culties based on the gender of the CBD THOSE WITH MORE THAN ONE CLAN, EACH also arise when treatment is being offered OCCUPYING A SPECIFIC AREA (60%) across kinship or clan boundaries. CBD were In this category, community members had to seen as potential sexual partners for any adult be correctly assembled within clan-speci® c ar- person of the opposite sex in another kinship eas. Clans were then empowered to select group or clan, especially when house-to-house their distributors and treatment centres, con- distribution was the mode of ivermectin distri- veniently and appropriately located according bution. If male, their presence in the home- to clans. When this approach was successfully stead was not welcomed by the males of the executed, communities achieved their target of group visited; if female, their presence was 90% coverage. When programme staff inter- resented by the females. In either case, the net fered with these decisions, or did not mobilize result was often a drop in treatment coverage, a suf® cient number of speci® c clan members and a feeling that the distribution programme to attend and make decisions, the distribution was intrusive or threatening. was beset by mistrust, accusations and coun- Although relatively few women have been ter-accusations, and the target coverage was involved in ivermectin distribution, they have never reached. been keenly interested in having access to information on the programme, and in being THOSE WHERE INDIVIDUAL HOUSEHOLDS FROM involved in the decision-making processes. MANY DIFFERENT CLANS, OR EVEN FROM DIF- Within their kinship groups, women are im- FERENT TRIBES, RESIDE IN ONE COMMUNITY portant opinion leaders, and their ideas are (3%) equally respected by their men folk. Thus, These communities were usually of the semi- support of the programme by women has been urban type. Since onchocerciasis is primarily a essential, even though women have not often rural disease, experience with applying CDTI been physically involved in distribution of the in a semi-urban environment is limited. In ivermectin. Women usually attend health- semi-urban communities one ® nds a mixture education meetings with community mem- of families from different clans and tribes, bers, and are often very active participants. displaced from their villages, and no longer Most comments and questions during such necessarily linked to land ownership. Close meetings come from women, and often proximity and lack of known kinship lines women will boo and stop drunken or stubborn leads to mistrust. As Katabarwa et al. (1999b) men from talking nonsense. reported, mobilization of the population in such communities was much more dif® cult Effects of Migration and Urbanization and cumbersome than in rural communities During distribution of ivermectin in 1998 and (Table 2). In 1998, rural communities 1999 to a total of 1730 communities, three achieved a mean coverage of 87.8% for the categories of community were observed. eligible population whereas the semi-urban communities achieved only 63.3% THOSE WHERE ONLY ONE CLAN DOMINATES (P 5 0.049). Similarly, during 1999, rural (37%) communities treated 94.3% of the eligible In this category, when the CDTI approach population whereas the semi-urban communi- was used and the dominant clan was engaged ties treated only 71.0% (P 5 0.028). More in the process, 90% coverage of the treat- health-education sessions and visuals (poster ment-eligible population was achieved. How- and pamphlets), more video shows, radio jin- ever, when programme staff from the district gles and other activities were required to mo- 494 KATABARWA ET AL.

TABLE 2 Mean, community-directed treatment coverages in the meso- or hyper- endemic communities of four Ugandan districts

Treatment coverage (%)

Semi-urban communities Rural communities (%)

District 1997 1998 1997 1998

Adjumani 67 70 98 93 Kabale 69 67 83 93 Kisoro 42 59 85 95 Nebbi 75 88 85 96 All four 63.3 71.0 87.8 94.3

tivate the semi-urban communities. The more mectin treatments carried out are to achieve complex lifestyles and time-demands on famil- their target coverage and become self- ies in the urban environment made it more sustaining, they need to be based on com- dif® cult to bring together a suf® cient number munity-directed distribution. Such distri- of community members to make the meaning- bution, if it is to succeed, must make full use ful decisions needed to implement a CDTI of the existing local kinship/clan system. programme. It was clear that, although neigh- Similar social systems exist in other African bourliness and kinship/clan systems were de- countries and their recruitment into the iver- termining factors in implementing satisfactory mectin distribution process is likely to be ivermectin distribution, the degree of these critical for success. qualities as `natural resources’ in semi-urban In Uganda, it becomes more dif® cult to communities was very varied. achieve satisfactory ivermectin distribution The dif® culties in establishing effective coverage in communities that are becoming CDTI in semi-urban and urban environments semi-urbanized, which contain many migrant are to some extent counteracted by the fact families, and in which the kinship/clan system that onchocerciasis usually becomes less en- is much less strong. This weakening of the demic as communities enlarge and become kinship/clan system may become an increas- more urbanized. As human population densi- ing problem for community-directed health ties increase, pollution of local Simulium programmes aiming to control diseases that, breeding sites also increases, and there is a unlike onchocerciasis, thrive in an urban or consequent reduction in man± ¯ y contact. semi-urban environment. However, good penetration of semi-urban Most health interventionists concerned with communities is vital to the success of control enlisting human behaviour and social struc- programmes for other diseases, such as tu- ture in the battle against infectious diseases berculosis and lymphatic ® lariasis, which are have con® ned themselves to studies of the transmitted in urban environments and which `knowledge, attitudes and perception’ (KAP) also require community direction and owner- of community members. These KAP studies ship. aim to identify `false beliefs’, and then provide a guide to the best approach to replacing them CONCLUSIONS with `accurate knowledge’ (Manderson, 1998). The attitude of the interventionist tends to be In rural Ugandan communities where that of a saviour of the people, whose mission onchocerciasis is endemic, if the annual iver- is to rid them of a `backward’ culture that KINSHIPS AND ONCHOCERCIASIS CONTROL 495 promotes disease. Experience in Uganda, on whether the bonds of friendship and neigh- the other hand, shows that these so-called bourliness can replace this, are factors that `backward’ social and cultural systems are im- have yet to be studied. The optimal manage- portant `natural resources’, that can and ment of disease-control programmes in these should act as a powerful motivational force for new environments may depend upon the out- the prevention and control of disease (and come of these future investigations. indeed for the general advancement of the communities). Increasingly more and more people are mi- ACKNOWLEDGEMENTS. We are grateful to the grating from their present rural Ugandan community members and health workers who communities towards larger towns or other willingly provided the vital information for areas, in search of new opportunities and for- this study. We are also indebted to Carter tune. In the process they become separated Center, Global 2000, APOC and the Ugandan from their kinship groups. At the same time Ministry of Health, for ® nancial contributions, the rural communities are gradually changing expertise and their promotion of community- and becoming `modernised’. The effects of directed treatment programmes for the control these changes on the kinship system, and of onchocerciasis in Uganda.

REFERENCES

ANON. (1999). National Strategic Plan for the Prevention of Blindness in Uganda. 5-year Strategic Plan (Jan. 2000±Dec. 2004). Kampala: Ugandan Government. FOSTER, G. M. (1987). Bureaucratic aspects of international health agencies. Social Science and Medicine, 25, 1039± 1048. KATABARWA, M. (1999). Modern health services vs the traditional `engozi’ system in Uganda. Lancet, ii, 343. KATABARWA, N. M. & MUTABAZI, D. (1998). The selection and validation of indicators for monitoring progress towards self-sustainment in community-directed, ivermectin-treatment programmes for onchocerciasis control in Uganda. Annals of Tropical Medicine and Parasitology, 92, 859± 868. KATABARWA, N. M., MUTABAZI, D. & RICHARDS, F. O. (1999a). Monetary incentives are detrimental to community-directed health programmes in less-developed countries. Lancet, ii, 1909. KATABARWA, N. M., MUTABAZI, D. & RICHARDS, F. O. (1999b) Controlling onchocerciasis by community- directed ivermectin treatment programmes (CDITP) in Uganda: why do some communities succeed and other fail? Annals of Tropical Medicine and Parasitology, 94, 343± 352. MANDERSON, L. (1998). Anthropology in infectious disease control. Tropical Medicine and International Health, 3, 1020± 1027. STONE, L. (1992). Cultural in¯ uence in community participation in health. Social Science and Medicine, 35, 409± 417. WORLD HEALTH ORGANIZATION (1996). Community-directed Treatment with Ivermectin. Report of a Multi-country Study. Geneva: WHO.