Human mobility-based spatial decay models of unsafe and safe water usage in rural Uganda

Over 2 billion people lack access to safe water and are at risk of waterborne pathogens when relying on unsafe open water sites including lakes, rivers, and streams. Yet, there is limited understanding of open water usage and related human mobility for disease control. We collected 10 days of wearable GPS logger data from 452 individuals aged 5-82 years in rural Uganda. Spatial decay models were built to predict individual-level usage of open water sites and taps/boreholes to inform schistosome pathogen re-exposure and optimal tap/-borehole placement. 64.4% and 33.1% of participants visited[≥] 1 open water site and tap/borehole, respectively. Open water site usage occurred within 10 km and tap/borehole usage within 1 km of households despite human mobility up to 100 km. Exponential decays accurately predicted site-specific open water (auROC 0.899) and tap/borehole (auROC 0.945) usage. Mobility and water site usage were non-linearly related. Open water usage was positively correlated with Schistosoma mansoni reinfection one year later. Even with universal tap/-borehole access, simulations showed open water usage would only decline by a maximum of 23%. Our spatial decay models offer scalable tools for identifying high-risk open water sites for pathogen transmission and optimising placement of safe water infrastructure and focal environmental interventions.