Begum, S., Ahmed, M., and Sen, B., 2013. Impact of water and sanitation interventions on childhood diarrhea: evidence from Bangladesh, 3ie Grantee Final Report. New Delhi: International Initiative for Impact Evaluation (3ie)
Link to Source
Despite progress in recent years, diarrhoea remains a major cause of infant mortality, morbidity, and acute child malnutrition in Bangladesh. Conventional wisdom promotes policies that expand access to improved water sources or to improved sanitation facilities, but recent evidence regarding the impact of each policy on diarrhoea incidence is mixed. This study examines the impact of improved water sources, improved sanitation facilities, and the two combined on diarrhoea incidence in Bangladesh.
The analysis includes two large, nationally representative household surveys in Bangladesh, the Bangladesh Demographic and Health Survey (BDHS) and the Multiple Indicator Cluster Survey (MICS) conducted jointly by the Bangladesh Bureau of Statistics (BBS) and UNICEF. BDHS data from 2007 form the basis for the propensity score matching (PSM) analysis, whereas MICS data from 2006 provide descriptive statistics on diarrhoea incidence. Estimates of the average effect of treatment on the treated (ATT) of improved water, sanitation, or water and sanitation services are obtained using kernel-matching PSM analysis using common support and a caliper-bound of 0.05. The estimates are obtained by calculating three propensity scores (for access to improved water, sanitation, and both combined), estimating the ATT effect, and then performing a “balancing test” to assess matching quality.
Three notable correlates of the propensity score for the binary outcome variable for access to both improved water and sanitation were wealth (negative correlation), education (with only secondary education showing significant positive correlation), and household size (negative correlation). Social characteristics such as religion and gender of household head were not important correlates. The balancing test found that the treatment and control groups generated by the PSM analysis were not statistically different in terms of observable characteristics.
The ATT estimates for the impact of access to safe water on diarrhoea incidence were insignificant. The authors posit that this may be due to low quality of nondrinking water, but also do not find impacts on use of improved water for nondrinking sources. The ATT estimates for the impact of improved sanitation on diarrhoea were likewise insignificant. However, combined access to both improved water and improved sanitation yielded a significant negative impact of 5 percentage points on diarrhoea incidence. This represents a 41.8-percent decrease in the treatment group relative to the control group (p = 0.05). These findings suggest the importance of access to improved water and sanitation facilities in tandem, and that one without the other may be insufficient for combatting diarrhoea incidence.