Participatory prevalence estimation: A pilot survey in Kenya

There is a need to find locally and globally relevant surveillance tools to measure disease prevalence and
inform control programs. Traditional veterinary knowledge has a contested role in surveillance systems.
We examined participatory epidemiology (PE) surveys for agreement between community perceptions
of prevalence of cattle conditions compared to clinical and laboratory results. Community meetings to
understand traditional knowledge of endemic disease were convened in randomly selected communities
in Western Province, Kenya. Community perceived prevalence for common cattle diseases and five focus
conditions (anemia, helminthosis, fascioliasis, trypanosomiasis and theilieriosis) was determined using
PE tools. Within selected communities, all cattle (>1 month old) were clinically examined and blood and
stool specimens collected. Definitive diagnosis was made by clinical findings, microscopy and molecular
techniques. Agreement was determined for prevalence obtained from community perceptions compared
to laboratory and clinical analysis. To assess the ability of individual farmers to identify ill animals and
accurately diagnose disease, the perceived health status from animals on randomly selected farms was
obtained by owner interview. A clinical examination was then performed with blood and stool specimens
analyzed for a battery of other endemic diseases. Agreement between farmer assessment of health status and
diagnosis for selected conditions against results from laboratory and clinical examination was determined.
This study highlights the accuracy and limitations of traditional knowledge. The usefulness of PE for
preliminary prevalence estimation was demonstrated by the ability of non-pastoralist farmers to estimate
herd level prevalence and individual animal health status. The study shows that PE can serve to combine
local knowledge inquiry with scientific study at a cost lower than laboratory and clinical surveys.