Effect of community-led total sanitation on open defecation in Uganda: A propensity score-matched analysis

A substantial proportion (7%) of people in Uganda practice open defecation. A Community-Led Total Sanitation (CLTS) project was started in 2011 to reduce indiscriminate disposal of excreta but the effect has not been rigorously evaluated. We, therefore, evaluated the effect of CLTS on reducing open defecation in the program intervention districts in Uganda. We used the 2016 Uganda Demographic and Health Survey (UDHS) data to conduct a quasi-experimental study using a propensity score matching (PSM) approach. The intervention group consisted of districts that implemented the CLTS and the comparison group were districts that did not implement the CLTS. We matched the intervention and comparison groups on several covariates in a 1:1 ratio within a caliper of 20% of the standard deviation of the propensity score. We confirmed balance in covariates using standardized mean difference (SMD) being <0.1 or 10%. We applied a conditional logit regression model on the matched dataset adjusting for matched pairs to estimate the intervention effect, reported as odds ratio (OR) and 95% confidence interval (CI). We assessed the robustness of the estimate using Mantel-Haenszel sensitivity analysis. There were 17,415 participants (4,373 intervention vs. 13,042 comparison) before PSM and 8,470 participants (4,235 comparison vs. 4,235 intervention) after PSM. In the unmatched weighted dataset, 5.4% of participants had open defecation (7.9% intervention vs. 4.6% comparison). In the matched weighted data, open defecation was 9.3% and 7.6% in the comparison and intervention groups, respectively. Results show that CLTS reduced the prevalence of open defecation by 37% (OR 0.63, 95% CI 0-54-0.74) and this finding is robust to unmeasured confounders (Gamma = 1.35, p = 0.052). We recommend the scale-up of CLTS to the remaining districts by the Government of Uganda to reduce open defecation.