SPATIO-TEMPORAL STATISTICAL MODELING OF MALARIA INCIDENCE IN GHANA: A GEOSTATISTICAL APPROACH TO REGIONAL DISPARITIES

This study addresses the persistent regional disparities in malaria incidence across Ghana from 2020 to 2024 by employing a spatio-temporal statistical modeling approach. The justification for this research stems from the inadequacy of national surveillance systems to capture spatial heterogeneities and temporal dynamics that influence malaria transmission, especially in high-burden regions such as the Upper East and Northern zones. The main objective was to develop a geostatistical framework that integrates climatic (rainfall, temperature, humidity), socio-economic (poverty, housing quality), and spatial factors to predict and map malaria incidence. Utilizing a quantitative observational design, data were collected from national agencies and analyzed using kriging, Bayesian hierarchical modeling, ARIMA forecasting, and spatial autocorrelation (Moran’s I). The study revealed that rainfall (r = 0.85), humidity (r = 0.80), and temperature (r = 0.75) are strongly correlated with malaria incidence, with a combined multiple regression model yielding an R² of 0.81. Spatial autocorrelation indicated significant clustering in districts like Bolga and Bawku West (Moran’s I > 0.30; Z-scores > 2.0), confirming hotspot persistence. Forecasts from the ARIMA model project incidence to increase from 305 to 315 per 1,000 between 2025 and 2027. The findings underscore the critical role of environmental and socio-economic variables in shaping malaria patterns and emphasize the need for climate-sensitive, geo-targeted interventions. Policymakers are urged to adopt dynamic, district-specific strategies, prioritize intervention during rainy seasons, and integrate spatial intelligence into malaria control programs. Future research should enhance early-warning capabilities using remote sensing and machine learning techniques.