Predicting quality and quantity of water used by urban households based on tap water service

Abstract

Despite significant progress in improving access to safe water globally, inadequate access remains a major public health concern in low- and middle-income countries. We collected data on the bacterial quality of stored drinking water and the quantity of water used domestically from 416 households in Uvira, Democratic Republic of the Congo. An indicator of tap water availability was constructed using invoices from 3685 georeferenced piped water connections. We examined how well this indicator predicts the probability that a household’s stored drinking water is contaminated with Escherichia coli, and the total amount of water used at home daily, accounting for distance from alternative surface water sources. Probability of drinking water contamination is predicted with good discrimination overall, and very good discrimination for poorer households. More than 80% of the households are predicted to store contaminated drinking water in areas closest to the rivers and with the worst tap water service, where river water is also the most likely reported source of drinking water. A model including household composition predicts nearly two-thirds of the variability in the reported quantity of water used daily at home. Households located near surface water and with a poor tap water service indicator are more likely to use water directly at the source. Our results provide valuable information that supports an ongoing large-scale investment in water supply infrastructure in Uvira designed to reduce the high burden of cholera and other diarrhoeal diseases. This approach may be useful in other urban settings with limited water supply access.