Forecasting Vector-Borne Disease in the United States

• Activity Number: 358 - Contributed Poster Presentations: Section on Statistics in Epidemiology
• Type: Contributed
• Date/Time: Tuesday, July 30, 2019 : 10:30 AM to 12:20 PM
• Sponsor: Section on Statistics in Epidemiology
• Author(s): Maddy St. Ville* and Christopher McMahan and Stella Self
• Companies: Clemson University and Clemson University and Clemson University

Abstract

A Bayesian spatio-temporal model was developed to forecast the seroprevalence of ehrlichiosis in the United States from eight factors deemed to be significant predictors. The eight factors consist of both static and dynamic predictors: elevation, percent water coverage, percent forest coverage, population density, median household income, temperature, relative humidity, and precipitation. In order to construct the 2019 forecast, dynamic factors were extrapolated to 2019 levels by various methods while the static ones were taken to be their most recent value. These factors were combined with test results that were collected on the county level across the contiguous United States from 2012-2018 and were analyzed using a Bayesian spatio-temporal model. The fitted model was then utilized to forecast the seroprevalence of ehrlichiosis for the 2019 calenda