Abstract
The coronavirus disease (COVID-19) that appeared in 2019 gave rise to a major global health crisis that is topping global health, socioeconomic and intervention programme agendas in 2020. Although the outbreak of COVID-19 has substantial and devastating impacts on developed countries, the countries of Global South share a higher proportion of the epidemic’s effects as shown particularly in morbidity and mortality rates in low-income countries . Globally, as at 13th June 2020, the total number of mortality cases was 428,337 of which 9% were in Asia (38,915) and 13.5% in South America (57,896) while 1.4% were in Africa (6080). The number of infections and deaths is still increasing rapidly at the time of writing. Modelling the effects of underlying factors and disease mortality is essential to plan effective control strategies for disease transmission and risks. The relationship between COVID-19 mortality rates and socio-demographic and health determinants can highlight various epidemic fatality risks. In this research, Geographic Information Systems (GIS) and an Artificial Neural Network (ANN) Multilayer Perceptron (MLP) were adopted to model and examine variations in COVID-19 mortality rates in the Global South. The model’s performance was tested using statistical measures of Mean Square Error (MSE), Root Mean Square Error (RMSE), Mean Bias Error (MBE), and the determination coefficient R2. The findings of this study indicated that the most important variable in explaining spatial mortality rate variations was the size of the elderly (65 and above) population . This was followed first by accessibility to handwashing facilities and second by hospital beds per 1000 population. Mapping the explanatory variables and estimated mortality rates and determining the importance of each variable in explaining the spatial variation of COVID-19 death rates across countries of the Global South can shed light on how public healthcare and demographic structures can offer policymakers invaluable guidelines to planning effective intervention strategies.