A0712
Title: Multi-scale spatiotemporal Covid-19 modeling: The mortality example
Authors: Andrew Lawson - Medical University of South Carolina (United States) [presenting]
Abstract: Data resources are now available that can more fully characterize the dynamics and progression of the COVID-19 pandemic around the world. In the US, county-level weekly data is now available from the Center for Disease Control (CDC), including incident case and mortality counts (as well as cumulative counts) for 173 weeks during the pandemic period. As both state-level and county-level data are now present, multiscale models for the dynamics of the pandemic process are considered. The linkage is described between state-level variation and the dynamics of county-level mortality. A variety of models of varying complexity are evaluated. An initial example of a single state (South Carolina) is demonstrated, whereby a time series model for the state level is linked to county-level weekly space-time variation. Goodness-of-fit (WAIC, MSE, DIC) is evaluated, and model extensions are reported to include deprivation measures and predictive metrics (out-of-data MSPE).
