Dengue: recent past and future threats.

Area of Science:

• Epidemiology
• Climate modeling
• Disease ecology

Background:

• Vector-borne diseases pose significant global health challenges.
• Accurate predictive models are crucial for understanding disease spread under climate change.
• Dengue, transmitted by Aedes mosquitoes, is a prime example of a climate-sensitive disease.

Purpose of the Study:

• To investigate key factors influencing statistical models for vector-borne disease prediction.
• To assess the impact of variable number, spatial resolution, and vector distribution on model accuracy.
• To identify critical predictor variables for present and future disease distribution.

Main Methods:

• Developed and evaluated statistical models for dengue prediction using varying numbers of meteorological variables (1, 5, 10) and spatial resolutions (1/6 to 2 degrees).
• Incorporated vector risk maps (Aedes aegypti, Aedes albopictus) into models alongside climate data.
• Utilized HadCM3 global circulation model predictions for future climate scenarios (2020s, 2040s, 2080s).
• Applied the Garthwaite-Koch corr-max transformation to quantify variable contributions.

Main Results:

• Model accuracy increased with a higher number of descriptor variables.
• Spatial resolution of climate data did not significantly impact model accuracy.
• Including vector distribution maps improved dengue model predictions.
• The Garthwaite-Koch transformation effectively demonstrated the relative importance of predictor variables.

Conclusions:

• Optimizing the number of predictor variables is key for robust vector-borne disease models.
• Future disease distribution modeling benefits from integrating both climate projections and vector ecology.
• The Garthwaite-Koch transformation offers a valuable method for interpreting complex predictive models.