Using a model-based geostatistical approach to design and analyse the prevalence of schistosomiasis in Kenya

Background: Infections caused by both Schistosoma mansoni and Schistosoma haematobium are endemic in Kenya, with over six million children at risk. A national school-based deworming programme was launched in 2012 with the goal of eliminating parasitic worms as a public health problem. This study used a model-based geostatistical (MBG) approach to design and analyse the impact of the programme and inform treatment strategy changes for schistosomiasis (SCH). Methods: A cross-sectional survey of 200 schools across 27 counties of Kenya was utilised. The study design, selection of the schools, and analysis followed the MBG approach, which incorporated historical data on treatment, morbidity, and environmental covariates. Results: The overall SCH prevalence was 5.0% (95% CI 4.9%–5.2%) and was estimated, with a high predictive probability of 0.999, to be between 1% and< 10%. The predictive probabilities at county level revealed county heterogeneity, with that of four counties estimated to be between 0% and< 1%, that of 20 counties estimated to be between 1% and< 10%, that of two counties estimated to be between 10% and< 20%, and that of one county estimated to be between 20% and< 50%. Conclusion: SCH treatment requirements can now be confidently refined based on the World Health Organization’s guidelines. The four counties with prevalences of between 0% and< 1% may consider suspending treatment only in areas (i.e., sub-counties and wards) where the prevalence is< 1%