Modeling COVID-19 latent prevalence to assess a public health intervention at a state and regional scale

Philip J. Turk; Shih-Hsiung Chou; Marc A. Kowalkowski; Pooja P. Palmer; Jennifer S. Priem; Melanie D. Spencer; Yhenneko J. Taylor; Andrew D. McWilliams

Using regional and state data, we demonstrate how epidemiological modeling based on local context is critically important to informing pandemic preparedness for health systems and policy leaders. The results highlight the importance for such models to be created using local data, as opposed to running a simulation which makes many assumptions about the truth of parameter values. All models should be continuously re-calibrated, and adapted to the rapid, continuously changing situations inherent to a pandemic. A one size fits all approach to the underpinning forecasting model or reliance on data that does not incorporate local context, sets the stage for misguided forecasting. Additionally, our study shows that while a classic SIR model may perform well in the early days of the pandemic, it begins to lose relevance with the emergence of additional influences like social distancing and enhanced awareness of personal hygiene.