Bayesian Hierarchical Modeling of Mpox in the African Region (2022–2024): Addressing Zero-Inflation and Spatial Autocorrelation

Abstract Mpox remains a significant public health challenge in endemic regions of Africa. Understanding spatial risk patterns and key drivers in high-risk countries is crucial for identifying vulnerable populations and guiding targeted interventions. This study employs a Zero-Inflated Poisson (ZIP) model with spatial autocorrelation to estimate adjusted relative risk of Mpox incidence across 24 African countries, stratified by Human Development Index (HDI) levels. The model accounts for overdispersion and excess zeros by incorporating spatial random effects and socio-environmental covariates. Spatial analysis of unadjusted and decomposition models revealed substantial heterogeneity in Mpox incidence, with elevated risk in the Democratic Republic of Congo (DRC) and Central African Republic (CAR) persisting after covariate adjustment (p < 0.001). Higher HDI levels correlated with reduced Mpox risk, with HDI quintile Q4 (very high HDI) showing a significant risk reduction (aRR: 0.009, 95% CrI: [0.0002, 0.45]). Protective factors in low-risk areas included increased life expectancy at birth (aRR: 0.092, 95% CrI: [0.864, 0.983]), educational attainment (aRR: 0.717, 95% CrI: [0.413, 0.605]), nonlinear increases in gross national income (GNI) per capita, and higher density of skilled health workers (aRR: 0.531, 95% CrI: [0.324, 0.603]). In contrast, urban density increased the risk of Mpox, highlighting the influence of dense population areas on disease transmission. Notably, Mpox risk in DRC and CAR could not be fully explained by assessed risk factors, suggesting spatial clustering of environmental or infectious exposures such as climate conditions, proximity to animal reservoirs, and human-animal interactions. Further research is essential to refine Mpox epidemiology and public health responses.