Abstract / Summary
Abstract Malaria remains a major public health challenge in Uganda, where climatic variability and persistent transmission cycles continue to shape disease dynamics. While previous forecasting studies have emphasized either environmental drivers or purely statistical time-series structures, limited attention has been given to integrating climatic, socioeconomic, and intervention-related determinants within a unified framework. This study develops an ARIMAX (Autoregressive Integrated Moving Average with exogenous variables) model to forecast malaria incidence in Uganda from 2000 to 2023, combining precipitation, temperature, GDP, undernourishment, insecticide-treated nets (ITN), and indoor residual spraying (IRS). After conducting multicollinearity diagnostics and stationarity tests, the model was estimated and validated using time-series cross-validation. Results indicate that precipitation and temporal dependence (AR(1)) are the most statistically robust predictors of malaria incidence. Precipitation shows a stable positive association with malaria cases, confirming its central role in vector breeding and transmission cycles. The significant autoregressive coefficient highlights strong persistence in malaria dynamics. In contrast, socioeconomic and intervention-related variables exhibit plausible but weaker and statistically less robust effects. Forecasts for 2024–2028 project a continued gradual decline in malaria incidence, with low prediction errors (RMSE = 65.51; MAE = 24.63), demonstrating satisfactory predictive performance. However, the study is limited by the use of aggregated national-level annual data, potential measurement inconsistencies, and the linear assumptions of the ARIMAX framework. Despite these constraints, the findings underscore the dominant influence of precipitation and temporal dynamics while emphasizing the complementary role of socioeconomic factors in strengthening malaria forecasting and informing resource allocation strategies.