B0438
Title: Bayesian space-time SIR modeling of Covid19 in SC, USA
Authors: Andrew Lawson - Medical University of South Carolina (United States) [presenting]
Joanne Kim - Medical University of South Carolina (United States)
Abstract: The Covid19 pandemic has spread across the world during much of 2020. Many regions have experienced its effects. South Carolina in the USA has seen cases since early March 2020 and a primary peak in early April 2020. A lockdown was imposed on April 6th, but the lifting of restrictions started on April 24th. The daily case and death data reported by NCHS (deaths) and state health department (cases) via the New York Times GitHUB repository have been analyzed, and approaches to modeling the data are presented. Spatially-referenced Bayesian susceptible/infected/removed (SIR) models with different assumptions concerning transmission and county-neighborhood relations are examined. Prediction is also considered, and the role of asymptomatic transmission is assessed as a latent unobserved effect. Both crude daily and smoothed counts for a single time period are examined, and one step prediction is provided.
