Sensor-measured versus reported latrine use to characterize sanitation intervention uptake in a randomized controlled trial among households in rural Bangladesh

Sanitation programs typically measure latrine access, which does not equate to use. We aimed to objectively measure latrine use with sensors among households enrolled in the sanitation and control groups of a randomized controlled trial (WASH Benefits) in Bangladesh. The intervention provided upgraded latrines and behavioral promotion. We recorded self-reported latrine use quarterly 1-3.5 years after intervention initiation. We installed motion sensors in household latrines in two annual waves (1.5-2.5 and 2.5-3.5 years after intervention initiation). We used zero-inflated negative binomial regression to compare sensor-measured daily latrine use events/person between (1) sanitation and control groups, and (2) households with different levels of self-reported latrine use. Households receiving the sanitation intervention had more sensor-measured daily latrine use events/person than controls in the first wave of sensor observations (ratio: 1.18, 1.06-1.32) but not in the second wave (ratio: 0.95, 0.86-1.05). In the sanitation group, households reporting exclusive latrine use (individuals >3 years always defecating in latrine) had a similar number of sensor-measured latrine use events as those not reporting exclusive use (ratio: 0.97, 0.86-1.09). In the control group, households reporting exclusive latrine use truly had more sensor-measured latrine use events than households not reporting exclusive use (ratio: 1.19, 1.03-1.37). We objectively demonstrate higher latrine use among sanitation intervention recipients than controls up to 2.5 but not 3.5 years after intervention initiation, indicating reduced uptake over time. Self-reported latrine use appears inflated among intervention recipients but not controls. Our findings underscore the importance of longitudinal follow-up and objective measurements in sanitation program assessments.