Geospatial Mapping of Pediatric Febrile Illness in Uganda to Inform Precision Public Health Interventions.

Background: Febrile illness is a leading cause of morbidity and mortality among children in low- and middle-income countries, yet the spatial distribution and environmental drivers of pediatric fever in Uganda remain poorly characterized. Methods: We analyzed data from the 2016 Uganda Demographic and Health Survey to estimate the prevalence of febrile illness among children under 5 years of age. Using a geostatistical binomial model, we evaluated associations between fever prevalence and environmental, nutritional, and sociodemographic covariates. Spatial prediction and model calibration were conducted using the PrevMap package in R, and model performance was assessed using nonrandomized probability integral transform (nrPIT) and theoretical variograms. Results: Among 14 195 children from 685 clusters, 4990 (35.1%) were reported to have had fever in the prior 2 weeks. Predicted fever prevalence varied substantially by region and month, with highest rates in the eastern and northeastern regions and in the period following the rainy season. Covariates including poverty, anemia, rainfall (2-month lag), enhanced vegetation index (1-month lag), and seasonality significantly improved model performance and reduced spatial uncertainty. Conclusions: Our findings reveal pronounced geographic and temporal heterogeneity in pediatric febrile illness in Uganda. Environmental and nutritional factors significantly contribute to this variation. These results support targeted, region-specific public health interventions and inform future research into the etiologic drivers of pediatric fever.