Global Forecasting Models for Dengue Outbreaks in Endemic Regions: A Systematic Review

Abstract

Background. Dengue is a rapidly spreading mosquito-borne disease, posing significant global health challenges, particularly in endemic regions. Recent years have witnessed an increase in the frequency and intensity of dengue outbreaks, necessitating robust forecasting models for early intervention. This systematic review aims to synthesize recent literature on dengue forecasting models, evaluate their predictive performance, and identify the most effective approaches. Materials and methods. A comprehensive search in Scopus, PubMed, ScienceDirect, and Springer databases was conducted following PRISMA guidelines. Studies were selected based on strict inclusion and exclusion criteria, and the quality of the research was evaluated using TRIPOD criteria. Out of 1,366 identified studies, 13 met the eligibility criteria. Data were extracted and analyzed to assess the accuracy and validity of the forecasting models employed. Results. The findings indicate that machine learning-based models, particularly random forest, outperform conventional statistical models such as ARIMA and Poisson regression. Additionally, climate data — especially temperature and rainfall play a critical role in forecasting dengue incidence. Conclusion. The present study corroborates the superior efficacy of machine learning-based forecasting models, particularly random forest, in forecasting dengue cases compared to conventional statistical methods. This finding provides a foundation for the development of an enhanced early warning system to address future outbreaks of dengue.