Exposure measurement error and the characterization of child exposure to fecal contamination in drinking water

Characterizing fecal contamination exposure from drinking water can introduce exposure measurement errors, i.e., differences between the observed and true exposure. These errors can mask the true relationship between fecal contamination exposure and waterborne diseases. We present a framework to quantify the impact of measurement errors on exposure–outcome health effect estimates introduced by variability in measured drinking water fecal contamination levels and household versus community sampling strategies. We matched fecal indicator bacteria (FIB) data for >37,000 drinking water samples to children aged 0–72 months from 19 studies in low- and middle-income countries and took two complementary analytical approaches. We found that household-level exposure assessments may attenuate effect estimates of FIB concentrations in drinking water on diarrhea, and single water samples may attenuate health effect estimates of FIB concentrations on linear growth. To understand the health effects of fecal contamination exposure, measurement error frameworks can be used to estimate more biologically relevant exposures.