Background: The prevalence of epilepsy was estimated in two villages of 3134 inhabitants, in Benin, in April and May 1997 using the capture-recapture method.
Methods: Information was obtained from (i) a door-to-door cross-sectional study, (ii) a non-medical source consisting of key informants (traditional practitioners, teachers, village leaders, and religious representatives) and (iii) a medical source through evaluation of medical records in health centres. In all the three situations, the diagnosis of epilepsy was confirmed by a neurologist.
Results: The door-to-door survey found 50 epileptics, i.e. a prevalence of 15.9 per 1000. The non-medical source found 26 patients. The medical source found only four patients. In total, 66 epileptics were found by combining the three sources, giving a prevalence of 21.1 per 1000. After application of the capture-recapture method, the estimated number of cases from the door-to-door survey and non-medical source was 105, and 110 cases when the medical source was considered as well. The respective prevalences were 33.5 per 1000, and 35.1 per 1000.
Conclusions: The door-to-door survey has been usefully improved by using key informants. The epilepsy prevalence estimate found by capture-recapture is clearly higher than that found by traditional cross-sectional methods, and could better depict the frequency of epilepsy in Africa.