Statistical Methods for COVID-19 Mortality Forecasting at the Small-Area Level

Back to the main COVID-19 projects page.

The Centre for Global Health Research has recently developed the Global COVID-19 Assessment of Mortality (GCAM) tool, which is an open, transparent, and continuously updated statistical model that combines actual COVID-19 mortality counts with Bayesian inference to forecast COVID-19 deaths.

Led by Jeffrey Rosenthal and Patrick Brown from the University of Toronto, students will improve the current GCAM tool by creating spatio-temporal predictions of COVID-19 mortality at the small area level with a hierarchical model; and accounting for irregularities in reporting times of non-hospital mortality by treating true event dates as latent variables. By improving this tool, this will allow the team to make the best possible use of COVID-19 mortality data as the pandemic spreads through its first wave, aiding countries in understanding national disease trajectories, and setting the stage for anticipating the course of future waves.

Comments are closed.