Water Quality Monitoring

The use of sanitary inspections combined with periodic water quality testing have been recommended by the WHO as cost-effective methods of identifying bacterial contamination at drinking water points. In some cases, sanitary surveys are recommended as screening tools for faecal contamination. Other studies have found poor statistical correlation between sanitary inspections and microbial contamination. However, the discriminatory performance of the sanitary inspection as a method to screen water sources for potential faecal contamination has not been well characterized.

We conducted sanitary inspections and tested for thermotolerant coliforms (TTC), a faecal indicator bacteria, among 7,317 unique water sources in West Bengal India. Our results indicate that the sanitary inspection score has poor ability to identify TTC contaminated sources. For deep handpumps, shallow handpumps, and gravity fed piped supplies, ROC analysis revealed that the sanitary inspection score was only marginally better than chance in predicting whether water points were TTC contaminated. Among unprotected springs and unprotected dug wells, the sanitary inspection score performed worse than chance at predicting TTC contamination. Aggregating over all source types, the sensitivity (true positive rate) of a high/very high sanitary inspection score for TTC contamination (TTC > 0 CFU/100ml) was 29.4% and the specificity (true negative rate) was 77.9%, resulting in substantial misclassification of the sites when using the established risk categories.

Our results were largely invariant to increasing the threshold for defining sites as contaminated. These findings suggest that sanitary surveys are inappropriate screening tools for identifying TTC contamination at water points.

Sanitation Monitoring

Measuring use of sanitation services in developing countries, including India and Bangladesh where open defecation is a cultural norm in rural areas, has historically been challenging. Numerous studies have shown a respondent bias, and structured observations, previously the gold standard approach, have now been demonstrated to be highly reactive. Therefore, improved, objective utilization methods are required. In the past several years, my contribution has been the supply and analysis of data generated from electronic movement sensors installed in latrines, across five controlled health impact studies in India and Bangladesh.

In one recent study in Bangladesh, our instruments demonstrated a significant over-reporting of latrine utilization. Across 1207 households randomly selected from 52 village committees in Bangladesh, the mean 4-day self reported latrine use was 32.8, while our sensor analysis suggested 21.7 uses on average, indicating more than 50% exaggeration of latrine use. At the low end of the regression model, the intercept suggests that many households report using latrines when in fact no use is detected (Delea et al., 2015).

This exaggerated self-reported use raises serious questions about the accuracy of self-reported data often used for policy and programmatic decision-making. Critically, the metrics used by program funders and implementers must at minimum narrow the gap between inputs and impact. While use may not be a sufficient measure, it is clear that it is coverage alone is insufficient. Electronic sensors may improve the objectivity of latrine use measurement, and enable more continuous monitoring.

Electronic sensors have demonstrated the inadequacy of previously standard methods for evaluating sanitation service utilization. Likewise, latrine coverage has been demonstrated as an insufficient measurement of sanitation impact. The data collected to-date across thousands of households in at least five rigorously designed epidemiological investigations may enable a robust meta-analysis of user utilization of sanitation services.


London School of Hygiene and Tropical Medicine

Texas A&M University

DelAgua Health