Abstract
BACKGROUND AND AIMS: The Sustainable Development Goal is targeted to end all types of malnutrition including underweight by 2030. However, the reduction rate is not as expected to meet the target. Thus, we aimed to investigate the spatiotemporal distributions and drivers of underweight among children aged below 5 years in Ethiopia. METHODS: Geostatistical analysis using the Bayesian framework was conducted to map the spatial and Spatiotemporal distributions of underweight. Data for the primary outcome was obtained from the Ethiopian Demographic and Health Survey 2000-2019. Covariate data were accessed from different credible online sources at high resolutions. Spatial binomial regression was fitted to identify drivers of underweight using the Bayesian approach. RESULTS: The overall national prevalence of underweight was 44.7, 37.7, 35.4, 25.5, and 23.8% in 2000, 2005, 2011, 2016, and 2019, respectively, with a total reduction rate of 46.8%. Significant spatial clustering of underweight was observed in Northern, Northwestern, Southeastern, Eastern borders, and the border between Oromia and SNNPR regions. Mean annual temperature (mean regression coefficient (β): -0.39; 95% credible interval (95% CrI): -0.63, -0.14), altitude (β:-0.30; 95% CrI: 0.57, -0.05), population density (β:-0.03; 95% CrI: -0.03, -0.02), and distance to water bodies (β:-0.03; 95% CrI: -0.05, -0.004) were negatively associated with being underweight. However, travel time to the nearest cities in minutes (β: 0.09; 95% CrI: 0.03, 0.14) was positively associated with being underweight. CONCLUSION: The national prevalence of underweight is reduced slower than expected in Ethiopia, with significant spatial variations across subnational and local levels. Temperature, altitude, population density, and distance to water bodies were negatively associated with underweight, whereas travel time to the nearest cities was positively associated with underweight in Ethiopia. Improving child nutrition through creating awareness and providing clean water should be strengthened.