Objective To determine whether the installation of deep tube wells to reduce exposure to groundwater arsenic in rural Bangladesh had an effect on the incidence of childhood diarrhoeal disease. Methods Episodes of diarrhoeal disease in children aged under 5 years that occurred on one specified day each month between 2005 and 2006 were reported to community health workers for six rural villages. A geographical information system containing details of household water use and sanitation in the villages was built using data obtained by a global positioning system survey. The information system also included health, spatial and demographic data. A field survey was carried out to determine whether households obtained drinking water from deep tube wells installed in 2005. The effect of deep tube well use on the incidence of childhood diarrhoea was assessed using a random effects negative binomial regression model. Findings The risk of childhood diarrhoea was 46% lower in the 179 households that used a deep tube well than in the 364 that used a shallow tube well (P = 0.032). Neither socioeconomic status, latrine density, population density nor study year had a significant influence on disease risk. The incidence of childhood diarrhoea declined dramatically between 2005 and 2006, irrespective of water source. Conclusion The introduction of deep tube wells to reduce arsenic in drinking water in rural Bangladesh had the additional benefit of lowering the incidence of diarrhoea among young children.