B0809
Title: A Bayesian surveillance metric to predict emerging high risk cluster regions of infectious disease
Authors: Andrew Lawson - Medical University of South Carolina (United States)
Joanne Kim - Medical University of South Carolina (United States) [presenting]
Abstract: Detection of the high-risk disease cluster has been the main research area for spatial disease mapping researchers. Numerous clustering methods have been developed to map the high-risk areas of the disease of interest. However, previous spatial disease mapping research focused on identifying the current hotspot of the elevated risk area. Still, it did not provide information about where the next high-risk cluster is likely to occur, given the existing hotspot. We will introduce a novel Bayesian metric to predict the occurrence of new clusters of the elevated risk areas for the infectious disease outbreak. Our novel metric is based on the Bayesian spatio-temporal hierarchical model and extended the posterior exceedance probability, which is commonly used for hotspot clustering. Specifically, we predict the next high-risk neighboring area given the existing hotspot by 1) using both exceedance probability and exceedance level and 2) effective information sharing between the areas' own risk with its risk trend over time and its neighborhood risk. We evaluate the performance of our metric with a simulation study based on different infectious disease outbreak situations and the real data of COVID-19 outbreak. We expect our novel metric would contribute to the public health surveillance of the infectious disease outbreak by providing a novel perspective for the high-risk area cluster prediction.
