Forecasting Climate-driven Dengue Incidence in Penang, Malaysia

Dengue continues to pose a major public health challenge in Malaysia, with no definitive cure currently available. Although the Ministry of Health of Malaysia has implemented various measures to control outbreaks, the number of cases keeps rising and is likely to worsen due to the impacts of climate change. Hence, early detection and prediction of dengue outbreaks are vital for the implementation of risk mitigation measures. This study applied and assessed the performance of a coupled ARIMAX and SI-SIR model for forecasting dengue incidence in Penang, Malaysia. Data from 2014 to 2020, including reported dengue cases and climate variables (rainfall and average temperature), were used. Previous research has demonstrated a strong correlation between climate factors and dengue transmission. Granger causality tests also indicated that the time series of rainfall and average temperature are significant predictors of the mosquito biting rate, which is closely linked to dengue transmission. Therefore, these climate variables were incorporated into the coupled model to enhance its forecasting performance. Through multiple simulation rounds with a four-week forecasting period, the coupled model achieved an average forecasting accuracy of around 80% in predicting dengue cases in Penang.