Infectious Disease
Forecasting Human West Nile Virus Incidence Using Mosquito Pool Infection Data in High-Risk Land Use Categories Jordan Robinson* Jordan Robinson Robinson University of Arizona
West Nile Virus (WNV) is one of the most prevalent vector-borne diseases in the United States. As the scale and distribution of WNV and its vector, the Culex mosquito, changes, there is a growing need to predict when and where WNV outbreaks will occur. Arizona suffers from above average WNV incidence. Previous work has highlighted an association between viral presence in mosquito pools and a higher number of reported human cases for the season. While previous forecasting has focused on peak mosquito activity, there has been little evaluation of early season activity and its use in predicting seasonal case burden. This project aims to measure if early season disease rates in mosquito pools can predict total human WNV cases for the season. Further, this work aims to determine if efficiency in forecasting can be developed by narrowing input data to mosquito pool information from specific land-use categories. Mosquito surveillance data from Maricopa County Vector Control was used to calculate average abundance and vector index at different intervals, which were then tested using negative binomial regression to identify the earliest seasonal interval at which mosquito indicators were associated with human WNV incidence. Surveillance data were further stratified by land-use categories, according to land-use composition within trap flight-range buffers, to determine which land-use categories were most impactful for predicting human cases. Results from the forecast model will inform WNV spread and relevant mapping can be used to focus vector surveillance and control efforts, and shape priorities in public education.