Climate, Spatial Clustering and Hotspots of Non-Communicable Disease Mortality in Sub-Saharan Africa: A Bayesian Spatial Epidemiology Study, 2000–2019

Background Non-communicable diseases (NCDs) now account for a growing share of premature mortality in sub-Saharan Africa (SSA), yet little is known about how climate and geography shape spatial inequalities in NCD deaths. Methods We assembled a country–year panel for forty-one SSA countries from 2000–2019, combining World Health Organization mortality estimates for four major NCD groups namely cardiovascular diseases, diabetes mellitus, malignant neoplasms and chronic respiratory diseases, with mid-year population denominators, climate surfaces (mean temperature and precipitation from WorldClim v2.1) and macro-socioeconomic covariates. Expected deaths were derived from age-standardised mortality rates and used as offsets in disease-specific Bayesian Poisson spatio-temporal models with Besag–York–Mollié 2 (BYM2) spatial random effects, first-order random walk temporal effects, and country–year interaction terms. Models were fitted in INLA with penalised complexity priors. We mapped climate-adjusted spatial random effects, identified multi-disease hotspots using posterior relative risks averaged over 2015–2019, and quantified cross-disease spatial correlations at SSA and regional-bloc level. Predictive performance was assessed using conditional predictive ordinates, PIT histograms and observed-versus-fitted plots. Results After adjustment for socioeconomic, health-system and HIV indicators, substantial residual spatial clustering remained, with elevated NCD mortality in parts of southern and eastern Africa and lower risk in several Sahelian countries. Countries classified as hotspots for one NCD often exhibited raised risks for others. Spatial correlations were positive between cardiovascular disease and diabetes, but negative between cardiovascular and respiratory mortality. Higher long-term temperature was associated with lower diabetes and respiratory mortality, while precipitation effects were generally weak. Conclusions NCD mortality in SSA displays marked, climate-adjusted spatial heterogeneity and partially shared hotspot patterns across causes. These findings support geographically targeted, climate-sensitive NCD prevention and health-system strengthening strategies. Trial registration: Not applicable.