Diarrhoea continues to be a principle cause of child mortality and morbidity in the developing world. In the Philippines it was responsible for 0.6 infant deaths per 1,000 live births in 2004, spurring the Philippine government to prioritize access to safe drinking water and basic sanitation facilities in advancing toward the Millennium Development Goal of reducing under-5 mortality. Though interventions promoting access to safe water and sanitation have been shown to reduce the incidence of diarrhoea, the full health impacts and cost-effectiveness of this method vary across countries and have not been fully quantified. In this paper, the authors assess the impact of access to piped water and flush toilets in the rural Philippines on the incidence of child diarrhoea.
The authors use propensity score matching methods to control for selection bias on observable characteristics, and estimate the average treatment effect on the treated with a data set from the National Demographic and Health Survey (NDHS) taken in 1993, 1998, 2003, and 2008. Examining under-5 children in rural households yields a sample of approximately 5,000 in 1993 and 1998, and about 3,800 in 2003 and 2008. The outcome of interest is a binary health outcome variable indicating that a child did or did not have diarrhoea during the reference period of the survey. The authors used propensity score matching of parental characteristics, economic status of household, ethnic background, religion, and administrative region. The analysis uses five-nearest neighbour matching, using caliper sizes of 0.001, 0.01, 0.02, and 0.03; the authors then test the robustness of the impact estimates using kernel matching with caliper sizes of 0.03 and 0.05.
The propensity square matching estimates indicate that access to piped water prevented 1.8'4.5 percent of diarrhoea in under-5 children in 2008 and up to 3.2 percent in 2003, and that flush toilets prevented 3.4'10 percent of diarrhoea in 2008 and 2.5'3 percent in 2003. All results are significant at the 5-percent level. Piped water had a 2-percent impact in 1993 at p = .1, but results were otherwise insignificant for both piped water and flush toilets in 1993 and 1998.
Overall, access to piped water and flush toilets appears to lower the prevalence of child diarrhoea slightly but inconsistently across years. The authors suggest that one possible explanation for the disappointing impacts may be water contamination at the point of use. The data set also may not sufficiently control for covariates, which could be remedied by further analysis using difference-in-differences techniques or randomised controlled trials. The authors recommend that the Philippine government re-examines which water and sanitation facilities are considered safe and appropriate, and monitor quality as well as quantity while expanding access to water and sanitation facilities across the country.