Expanding drinking water chlorination could substantially reduce the burden of disease in low- and middle-income countries, but the taste of chlorinated water often impedes adoption. We developed a Monte Carlo simulation to estimate the effect of people's choice to accept or reject drinking water based on chlorine taste and their subsequent exposure to E. coli and trihalomethanes, a class of disinfection byproduct (DBP). The simulation used empirical data from Dhaka, Bangladesh, a megacity with endemic waterborne disease. We drew on published taste acceptability thresholds from Dhaka residents, measured residual chlorine and thermotolerant E. coli inactivation following the addition of six chlorine doses (0.25-3.0 mg/L as Cl2) to untreated piped water samples from 100 locations, and analyzed trihalomethane formation in 54 samples. A dose of 0.5 mg/L, 75% lower than the 2 mg/L dose typically recommended for household chlorination of low-turbidity waters, minimized overall exposure to E. coli. Doses of 1-2 mg/L maximized overall exposure to trihalomethanes. Accounting for chlorine taste aversion indicates that microbiological exposure increases and DBP exposure decreases above certain doses as a higher proportion of people reject chlorinated water in favor of untreated water. Taken together with findings from other modeling analyses, empirical studies, and field trials, our results suggest that taste acceptability should be a critical consideration in establishing chlorination dosing guidelines. Particularly when chlorination is first implemented in water supplies with low chlorine demand, lower doses than those generally recommended for household water treatment can help avoid taste-related objections while still meaningfully reducing contaminant exposure.
Keywords: Bangladesh; Chlorine; Disinfection byproducts (DBPs); Drinking water; Monte Carlo simulation; Taste.
Copyright © 2021. Published by Elsevier Ltd.