Modelling the spatial varying relationships between socioeconomic inequalities and COVID-19 mortality in the African subregion

Infectious diseases continue to be a global health menace to society. Admittedly, there have been improvements in public health infrastructure and structural response to these diseases, however, a lot more needs to be done to have a firm grip. This study applied the geographically weighted poisson regression (GWPR) model to study the spatial variations in the effects of socioeconomic factors on COVID-19 Mortality in the African subregion. The performance of the GWPR model was compared to the conventional Poisson regression model. The results suggest that accounting for the spatial heterogeneity through the GWPR model (RMSE = 100.604, MAE = 70.582, Pseudo R² = 0.773) improves the modelling performance compared to the conventional Poisson regression model (RMSE = 118.574, MAE = 83.054, Pseudo R² = 0.530). In addition, the effects of all the socioeconomic variables considered (i.e., educational index, unemployment rates, income index, gross domestic product, and forest area) on COVID-19 mortality were found to be spatial non-stational across the subregion. These results highlight the need to draw preventive and response policies applicable to infectious diseases with greater consideration of the different geographical points on the continent.