Ayşe Ercümen, Benjamin F. Arnold, Emily Kumpel, Zachary Burt, Isha Ray, Kara Nelson, & John M. Colford, Jr
PLoS Medicine, 12(10): e1001892.
Publication year: 2015



Intermittent delivery of piped water can lead to waterborne illness through contamination in the pipelines or during household storage, use of unsafe water sources during intermittencies, and limited water availability for hygiene. We assessed the association between continuous versus intermittent water supply and waterborne diseases, child mortality, and weight for age in Hubli-Dharwad, India.

Methods and Findings

We conducted a matched cohort study with multivariate matching to identify intermittent and continuous supply areas with comparable characteristics in Hubli-Dharwad. We followed 3,922 households in 16 neighborhoods with children <5 y old, with four longitudinal visits over 15 mo (Nov 2010–Feb 2012) to record caregiver-reported health outcomes (diarrhea, highly credible gastrointestinal illness, bloody diarrhea, typhoid fever, cholera, hepatitis, and deaths of children <2 y old) and, at the final visit, to measure weight for age for children <5 y old. We also collected caregiver-reported data on negative control outcomes (cough/cold and scrapes/bruises) to assess potential bias from residual confounding or differential measurement error.

Continuous supply had no significant overall association with diarrhea (prevalence ratio [PR] = 0.93, 95% confidence interval [CI]: 0.83–1.04, p = 0.19), bloody diarrhea (PR = 0.78, 95% CI: 0.60–1.01, p = 0.06), or weight-for-age z-scores (Δz = 0.01, 95% CI: −0.07–0.09, p = 0.79) in children <5 y old. In prespecified subgroup analyses by socioeconomic status, children <5 y old in lower-income continuous supply households had 37% lower prevalence of bloody diarrhea (PR = 0.63, 95% CI: 0.46–0.87, p-value for interaction = 0.03) than lower-income intermittent supply households; in higher-income households, there was no significant association between continuous versus intermittent supply and child diarrheal illnesses. Continuous supply areas also had 42% fewer households with ≥1 reported case of typhoid fever (cumulative incidence ratio [CIR] = 0.58, 95% CI: 0.41–0.78, p = 0.001) than intermittent supply areas. There was no significant association with hepatitis, cholera, or mortality of children <2 y old; however, our results were indicative of lower mortality of children <2 y old (CIR = 0.51, 95% CI: 0.22–1.07, p = 0.10) in continuous supply areas. The major limitations of our study were the potential for unmeasured confounding given the observational design and measurement bias from differential reporting of health symptoms given the nonblinded treatment. However, there was no significant difference in the prevalence of the negative control outcomes between study groups that would suggest undetected confounding or measurement bias.


Continuous water supply had no significant overall association with diarrheal disease or ponderal growth in children <5 y old in Hubli-Dharwad; this might be due to point-of-use water contamination from continuing household storage and exposure to diarrheagenic pathogens through nonwaterborne routes. Continuous supply was associated with lower prevalence of dysentery in children in low-income households and lower typhoid fever incidence, suggesting that intermittently operated piped water systems are a significant transmission mechanism for Salmonella typhi and dysentery-causing pathogens in this urban population, despite centralized water treatment. Continuous supply was associated with reduced transmission, especially in the poorer higher-risk segments of the population.