Developing a dengue prediction model based on climate in Tawau, Malaysia

Area of Science:

• Environmental Science
• Epidemiology
• Public Health

Background:

• Dengue incidence has surged in Malaysia, posing a significant public health challenge.
• Current control strategies rely heavily on vector control due to vaccine uncertainties and lack of antiviral drugs.

Purpose of the Study:

• To analyze the relationship between weather variables and dengue incidence in Tawau, Malaysia.
• To develop a predictive model for forecasting dengue outbreaks in the region.

Main Methods:

• Collected 12 years of monthly dengue incidence data and weather data (temperature, humidity, rainfall) from Tawau.
• Employed cross-correlation analysis to identify significant weather predictors and their lagged effects.
• Developed and compared three forecasting models: Poisson regression, SARIMA, and SARIMA with external regressors.

Main Results:

• Statistically significant correlations were found between lagged weather variables (maximum temperature, relative humidity, rainfall) and dengue incidence (4-6 month lags).
• The SARIMA model with external regressors, incorporating lagged temperature, humidity, and rainfall, proved to be the best fit.
• The selected SARIMA model demonstrated high accuracy in predicting dengue outbreaks one month in advance (MAE: 7.032, MSE: 83.977).

Conclusions:

• Weather variables, particularly temperature and humidity, are significant predictors of dengue incidence magnitude.
• The developed forecasting model offers a novel, evidence-based approach for predicting dengue outbreaks in Tawau, enabling proactive public health interventions.
• The model can forecast potential dengue outbreaks 1-4 months ahead, aiding resource allocation and decision-making, though further data is needed for refinement.