Spatially heterogeneous and neighborhood impacts of deep tubewells on childhood diarrhea in rural Bangladesh

Deep tubewells that procure groundwater from deep arsenic-free aquifers are crucial sources of safe drinking-water in rural Bangladesh. However, their health benefits may be heterogeneous and context dependent: inequitable distribution, long collection distances and continued use of unsafe alternatives may increase microbial contamination and diarrheal disease. We examined whether the association between household deep tubewell use and under-five diarrhea varies across space and how neighborhood context helps explain this variation. Prospective diarrhea surveillance was conducted across 19,203 households with under-five children in Matlab, Bangladesh from March 2018 to October 2019. We fitted hierarchical logistic regression models accounting for spatially varying relationships and neighborhood-level measures of deep tubewell use. In non-spatial models, household deep tubewell use is associated with lower childhood diarrhea (Odds Ratio (OR) 0.81, 95% Uncertainty Interval (UI) 0.68-0.96). Accounting for space suggests spatial confounding and that the effect of household deep tubewell use on diarrhea may change based on location. Additionally, there is evidence of neighborhood effects: households in areas where more than 60% of neighbors used deep tubewells had lower odds of childhood diarrhea than those in neighborhoods with lower deep tubewell use (OR 0.72, 95% UI 0.57-0.92), with strongest protection among households that used deep tubewells. Results suggest that despite positive effects, benefits of deep tubewells are unequally distributed, and increasing deep tubewell density can further reduce childhood diarrheal disease burden in rural Bangladesh. This study highlights the importance of incorporating geographic context in the design and evaluation of drinking-water and other population health interventions.