Validation of participatory appraisal for use in animal health information systems in Africa
Catley, Andrew Paul
Participatory appraisal (PA) is a methodology for problem description and analysis that has been widely used in less developed countries (LDCs) since the 1980s. The use of PA by veterinarians in LDCs has been restricted to mainly small-scale community-based animal health projects. Adoption of PA by veterinarians, particularly those working for government, was limited because of concerns about the reliability and validity of the methods. Three studies were conducted with pastoralist and agropastoralist communities in East Africa to vaiidate PA, by comparison of data derived from PA with conventional veterinary investigation and epidemiological information. In southern Sudan, research was conducted on a chronic wasting syndrome in adult cattle in Dinka and Nuer communities; in Kenya, research was conducted on bovine trypanosomiasis in Orma communities; and in Tanzania research was conducted on possible association between a chronic heat intolerance syndrome (HI) and foot and mouth disease (FMD). Participatory appraisal methods, called matrix scoring, seasonal calendars and proportional piling, were standardised and repeated to generate quantitative data. The level of agreement between informant groups was assessed using the Kendal coefficient of concordance (Jf).. Matrix scoring was adapted for use by veterinarians to enable comparison of veterinarian's perceptions of disease signs and causes, with those of pastoralist informants. The data were compared using direct visual assessment, hierarchical cluster analysis and multidimensional scaling. Matrix scoring, seasonal calendars and proportional piling were judged to have good validity and reliability. In Tanzania, adaptation of proportional piling enabled calculation of the relative risk of HI cases being observed in cattle herds with previous history of FMD, and demonstrated significant association between HI and FMD. This finding was confirmed by detection of antibody to non-structural proteins to FMP in herds with and without HI. It was concluded that PA methods were reliable and valid methods for veterinary epidemiology when used by trained PA practitioners in agropastoral and pastoral settings. The methods were valuable for data collection and analysis, and for enabling greater involvement of livestock keepers in veterinary service development and research. Participatory appraisal could be further adapted to improve the design of primary veterinary services and disease surveillance systems. In veterinary research, PA was considered to be particularly useful during the exploratory phase of research and for generating research hypotheses. It was also concluded that institutional changes were required for the widespread adoption of PA by veterinarians in Africa.