B0179
Title: Predicting the risk of novel pathogen introductions from disease surveillance data
Authors: Nelson Walker - Kansas State University (United States) [presenting]
Trevor Hefley - Kansas State University - Department of Statistics (United States)
Daniel Walsh - United States Geological Survey - National Wildlife Health Center (United States)
Ian McGahan - University of Wisconsin - Department of Statistics (United States)
Daniel Skinner - Illinois Department of Natural Resources (United States)
Daniel Storm - Wisconsin Department of Natural Resources (United States)
Abstract: In the course of an infectious disease outbreak, researchers often must estimate or infer the source of the causative pathogen, the risk factors associated with the spread and growth of the pathogen, and risk factors that may be associated with new outbreaks. Because the exact time and location of introduction for the pathogen is usually unobserved, these questions must be addressed using incomplete or indirect data, such as spatio-temporal disease surveillance data. We introduce a Bayesian hierarchical mixture model for spatio-temporal, binary disease surveillance data that accounts for the dynamic process of the pathogen diffusing and multiplying through a population from multiple sources. Our framework provides approximate posterior estimates for the number, locations, and times of introduction of the pathogen in a population, as well as posterior inference on parameters associated with pathogen growth and diffusion. We also obtain posterior inference on the generative spatial process that produced the pathogen introductions. We demonstrate this framework using disease surveillance data for chronic wasting disease in white-tailed deer from Wisconsin and Illinois in the USA.
