Comparison of key informant and survey methods for ascertainment of childhood epilepsy in West Bengal, India.

BACKGROUND This study aimed to compare efficacy and cost of key informants and survey for ascertainment of childhood epilepsy within a treatment context in rural India. METHODS The study was set in a non-governmental, community programme for the functional and socioeconomic rehabilitation of children with disabilities in rural West Bengal, India. Ascertainment was by two methods: house-to-house survey of 15000 households and also by 430 key informants including village leaders, health workers and 670 schoolchildren. Methods were compared for positive predictive value, and sensitivity by capture-recapture technique. Ninety four children were enrolled into treatment. Predictors of treatment success were determined by multiple logistic regression analysis, giving adjusted odds ratios for remission. The costs of identifying one case and one treatment success were measured by costing personnel, materials and overheads. RESULTS The survey was four times as sensitive as key informants although the positive predictive values were similar (36%, 40%). The survey had an absolute sensitivity of only 59%. Identification by key informants strongly predicted successful treatment outcome (odds ratio [OR] = 4.74, 95% confidence interval [CI] : 1.19-18.85). The cost of finding one case was US$11 and US$14, and of finding one successful treatment outcome US$35 and US$67 for informants and survey respectively. Key informants were essential in attaining longer term programme objectives. CONCLUSIONS In the context of a treatment programme, key informants were the more cost-effective method, but community involvement was traded against low sensitivity in the short term. Overall ascertainment costs were significant in the context of primary health care in India.