Effects of point-of-collection water treatment program on chlorination rates and under-5 mortality in rural Kenya

About this event

Treating household drinking water with chlorine prevents childhood deaths, but requires constant dosing. We conduct a long-run evaluation of community-level health impact of a water chlorination programs implemented in the same rural Kenyan communities: the water arms of WASH-Benefits (2014-2016), a cluster-randomised trial that included bottled dilute chlorine and chlorine dispensers at water points, and a subsequent large-scale NGO program that continued dispenser provision in the original trial water clusters. We collected data on over 80,000 live births and 1,806 deaths in under-5 children in the period up to seven years after treatment roll-out. We also measured the presence of chlorine in stored drinking water. Measured chlorination rate (total chlorine residual, TCR >= 0.1 mg/L) at the time of the surveys was low overall: 7% in treatment areas; even among households within 100 m of a dispenser it was only 13%. Consistent with limited exposure, analyses by original assignment found no significant effects on study’s primary objective of all-cause under-5 mortality among households not previously enrolled into WASH-Benefits (intent-to-treat HR = 0.99, 95% CI 0.90, 1.09). Taken together, the results indicate that failures of sustained take-up and incomplete coverage can dominate the health impacts of point-of-use water treatment at scale. Water interventions that rely on household behavior require stronger delivery, monitoring, and staged evaluation to yield long-run health benefits.