Can a school-based water, sanitation and hygiene intervention catalyze changes in household behaviors and environment? Evidence from a randomized trial in western Kenya

Background. Improvements in school water, sanitation and hygiene (WASH) are intended to improve the health, education and wellbeing of school aged children. However the overall impact in low-income settings may depend on whether they trigger changes in household behaviors as well.

Methods. A randomized cluster trial was conducted in western Kenya, with 135 school-communities randomized into three arms: control; water treatment and handwashing; and that plus sanitation improvement. A baseline household survey was conducted in early 2007 and a follow up in mid-2008. The analysis focuses on self-reported water treatment and residual chlorine in drinking water. Attributable changes in water treatment were compared using cluster-level double-difference analysis. Household determinants of use were assessed with cross-sectional multivariate regression analysis.

Results. Confirmed use of chlorine products for water increased in both the intervention and control arms (ranging from 3-5% at baseline to 13-16% at follow-up). Difference-in-difference analysis showed a statistically significant increase attributable to intervention (p<0.05). The impact of the intervention represented a doubling of the percent of households with residual chlorine. Multivariate analysis of household predictors of uptake showed that economic status, maternal education, and other factors were significant determinants of household water treatment.

Conclusion. School WASH improvement resulted in increased household water treatment. While statistically significant, the change was small compared to the secular change in all communities. More deliberate strategies designed to increase behavior diffusion to households may increase diffusion. The results point out the importance of rigorous study designs to control for background changes in behavior.