Objective: We sought to estimate visceral leishmaniasis (VL) burden in Bangladesh, India, and Nepal and document care-seeking behaviour for VL to provide baseline information for monitoring the VL elimination program and identify options for improved case finding and management.
Design: A cross-sectional study using cluster sampling (clusters being villages) of 4 VL endemic districts was used in order to document all current and existing VL cases over the preceding 12 mo. Extended (in-depth) interviews were conducted in a subsample of households to explore (a) VL-related knowledge, attitudes, and practices of the population; (b) use of VL care by patients; and (c) delay between onset of symptoms, diagnosis, and start of treatment, as well as treatment interruption. Findings were discussed with national program managers and policy makers to develop improved strategies.
Results: Screening for VL was done in 18,933 households (106,425 inhabitants). The estimated annual incidence of VL in the endemic districts was on average 22 times higher than the elimination target of less than one case per 10,000 inhabitants in 2015. This incidence varied widely between study sites, from 9.0 to 29.8 per 10,000 inhabitants. The percentage of newly detected cases through the household screening was high in the districts least covered by health-care services (particularly Rajshahi, Bangladesh, 49%; and to a lesser extent Vaishali in Bihar, India, 32.5%), and much lower in districts with greater availability of VL care (Muzaffarpur, India, 3.8%). On average 267 houses had to be visited, i.e., at least three to four working days per health worker, to identify a new VL (ranging from 1,432 houses in Muzaffarpur, India to only 166 houses in Rajshahi, Bangladesh). Knowledge of the disease and its vectors was good in India and Nepal but poor in Bangladesh (Rajshahi) where very little attention has been given to VL over the last decades. Although all socio-demographic indicators showed high levels of poverty, people in India preferred private medical practitioners for the treatment of VL, whereas in Nepal, and even more in Bangladesh, the public health-care sector was preferred. Delays between onset of symptoms and diagnosis as well as start of treatment was high. Reported non-adherence to treatment was particularly high in the more under-served districts and was mainly due to lack of resources.
Discussion: The findings suggest that (a) house-to-house screening may be useful in highly endemic districts with a poor passive case detection system, but further evidence on case finding adapted to local conditions has to be collected; (b) strengthening the quality of the public health sector is imperative in the three countries, especially in India, with its largely unregulated private-sector provision of VL care.