Multiplicative Mixed-Effects Modelling of Dengue Incidence: An Analysis of the 2019 Outbreak in the Dominican Republic

Abstract

Dengue is a vector-borne disease that is endemic to several countries, including the Dominican Republic, which has experienced dengue outbreaks for over four decades. With outbreaks growing in incidence in recent years, it is becoming increasingly important to develop better tools to understand drivers of dengue transmission. Such tools are critical for providing timely information to assist healthcare authorities in preparing human, material, and medical resources for outbreaks. Here, we investigate associations between meteorological variables and dengue transmission in the Dominican Republic in 2019, the year in which the country’s largest outbreak to date ocurred. We apply generalized linear mixed modelling with gamma family and log link to model the weekly dengue incidence rate. Because correlations in lags between climate variables and dengue cases exhibited different behaviour among provinces, a backward-type selection method was executed to find a final model with lags in the explanatory variables. We find that in the best models, meteorological conditions such as temperature and rainfall have an impact with a delay of 2–5 weeks in the development of an outbreak, ensuring breeding conditions for mosquitoes.

Keywords

• dengue
• Dominican Republic
• climate variables
• lags
• generalized linear mixed models