Diarrheal illness remains a leading cause of morbidity and mortality in children < 5 years in developing countries, and contaminated water contributes to diarrhea risk. To address this problem, a novel hollow fiber ultrafilter (HFU) was developed for household water treatment. To test its impact on water quality and infant health, we conducted a cluster-randomized longitudinal evaluation in 10 intervention and 10 comparison villages in Kenya, attempting to enroll all households with infants (< 12 months old). We conducted a baseline survey, distributed HFUs to intervention households, made biweekly home visits for 1 year to assess water treatment practices and diarrhea in infants, and tested water samples from both groups every 2 months for Escherichia coli. We enrolled 92 infants from intervention households and 74 from comparison households. During the 1-year study period, 45.7% of intervention households and 97.3% of comparison households had at least one stored water sample test positive for E. coli. Compared with comparison households, the odds of E. coli contamination in stored water was lower for intervention households (odds ratio [OR]: 0.42, 95% CI: 0.24, 0.74), but there was no difference in the odds of reported diarrhea in infants, adjusting for covariates (OR: 1.19, 95% CI: 0.74, 1.90). Although nearly all water samples obtained from unprotected sources and filtered by the HFU were free of E. coli contamination, HFUs alone were not effective at reducing diarrhea in infants.