Methods and Challenges for Forecasting Mosquito-borne Disease Incidence Using a Meteorologically Driven, Coupled Entomological-epidemiological Model (Invited Presentation)

Mosquito-borne diseases are re-emerging as major public health threats in both tropical and temperate regions of the Americas. Dengue virus transmission has increased in magnitude and range over the past few decades and Chikungunya virus (2013) and Zika virus (2015) were recently introduced into the Americas. Multiple factors contribute to the introduction and spread of these viruses including increased travel, socioeconomic conditions, and the environment. Weather and climate influence mosquito population dynamics through temperature regulated survival, development, and reproduction rates and precipitation driven habitat formation. The incubation period of the viruses is also largely regulated by temperature. The relationship between weather and climate and incidence of mosquito-borne diseases is well studied, however, most of these studies focus on replicating past epidemics with few studies testing predictive tools. Advances in weather and seasonal climate forecasts provide an opportunity to forecast mosquito-borne disease incidence. Satellite enhanced weather forecasts could be used to drive mosquito-borne disease transmission models to generate weekly predictions of disease incidence. Additionally, seasonal climate forecasts could be used to make predictions 1-3 months into the future. Using weather and climate forecasts to drive disease models introduces several challenges. For example, there is uncertainty in both the climate and weather forecasts and the disease models. Additionally, model parameter values may need to change throughout the transmission season. This study explores these challenges and their potential solutions by conducting seasonal forecasts of dengue in Rio de Janeiro, Brazil and weekly forecasts of Zika virus in several Caribbean countries using a coupled entomological-epidemiological model. The results have been mixed. While climate and weather forecasts provide the opportunity to predict mosquito-borne disease incidence, the quality of the forecasts is extremely important. Additionally, up to date disease incidence data can improve predictions by allowing continual updates of model parameter values. Consequently, improvements of climate and weather forecasts and increased availability of epidemiological data will be vital to improving the quality of mosquito-borne disease prediction.