Real-time dengue forecast for outbreak alerts in Southern Taiwan

Area of Science:

• Epidemiology
• Environmental Health
• Biostatistics

Background:

• Dengue fever is a growing global public health concern, with significant outbreaks in Taiwan during 2014-2015.
• Existing statistical models may be insufficient for predicting large-scale dengue outbreaks.

Purpose of the Study:

• To develop a predictive model for dengue fever occurrence to enable timely public health warnings.
• To assess the association between daily weather variability and dengue case numbers in Kaohsiung, Taiwan.

Main Methods:

• Application of a novel autoregressive (AR) model incorporating lagged weather variables.
• Development of 5-day-ahead and 15-day-ahead dengue case prediction models.
• Analysis of daily weather data and dengue case counts in Kaohsiung from 2014-2015.

Main Results:

• Dengue case numbers in Kaohsiung were significantly associated with humidity and the mosquito biting rate (BR).
• The developed AR model demonstrated good performance using only meteorological factors.
• The model is simple, intuitive, and capable of real-time updates.

Conclusions:

• The proposed real-time dengue forecast model can aid health agencies in mitigating epidemic impacts.
• Meteorological factors, particularly humidity and biting rates, are crucial predictors for dengue outbreaks.
• The AR model provides a practical tool for proactive public health interventions against dengue fever.