A Coupled Statistical and Deterministic Model for Forecasting Climate-Driven Dengue Incidence in Selangor, Malaysia

• Published in: Bulletin of mathematical biology Vol. 86; no. 7; p. 81
• Main Authors: Lu, Xinyi, Teh, Su Yean, Koh, Hock Lye, Fam, Pei Shan, Tay, Chai Jian
• Format: Journal Article
• Language: English
• Published: New York Springer US 01.07.2024; Springer Nature B.V
• Subjects: Aedes - virology; Animals; Cell Biology; Climate; Climate Change - statistics & numerical data; Computer Simulation; Dengue - epidemiology; Dengue - transmission; Dengue fever; Disease Outbreaks - statistics & numerical data; Epidemics; Epidemiological Models; Epidemiology; Forecasting - methods; Humans; Humidity; Incidence; Life Sciences; Malaysia - epidemiology; Mathematical and Computational Biology; Mathematical Concepts; Mathematical models; Mathematics; Mathematics and Statistics; Methods; Models, Biological; Models, Statistical; Mosquito Vectors - virology; Mosquitoes; Outbreaks; Public health; Rain; Rainfall; Temperature; Malaysia; Selangor Malaysia; Rainfall; Temperature; ARIMAX; SIR model; Humidity
• DOI: 10.1007/s11538-024-01303-2
• ISSN: 0092-8240; 1522-9602

The mosquito-borne dengue virus remains a major public health concern in Malaysia. Despite various control efforts and measures introduced by the Malaysian Government to combat dengue, the increasing trend of dengue cases persists and shows no sign of decreasing. Currently, early detection and vector control are the main methods employed to curb dengue outbreaks. In this study, a coupled model consisting of the statistical ARIMAX model and the deterministic SI-SIR model was developed and validated using the weekly reported dengue data from year 2014 to 2019 for Selangor, Malaysia. Previous studies have shown that climate variables, especially temperature, humidity, and precipitation, were able to influence dengue incidence and transmission dynamics through their effect on the vector. In this coupled model, climate is linked to dengue disease through mosquito biting rate, allowing real-time forecast of dengue cases using climate variables, namely temperature, rainfall and humidity. For the period chosen for model validation, the coupled model can forecast 1–2 weeks in advance with an average error of less than 6%, three weeks in advance with an average error of 7.06% and four weeks in advance with an average error of 8.01%. Further model simulation analysis suggests that the coupled model generally provides better forecast than the stand-alone ARIMAX model, especially at the onset of the outbreak. Moreover, the coupled model is more robust in the sense that it can be further adapted for investigating the effectiveness of various dengue mitigation measures subject to the changing climate.