(Talk) Novel modeling tools. Estimating effective reproductive number and forecasting hospitalizations.

Abstract

This presentation introduces innovative modeling tools designed to enhance the public health response to COVID-19 by leveraging wastewater-based epidemiology. We present a sequential method for monitoring COVID-19 infections through test positivity rates derived from SARS-CoV-2 RNA concentrations in wastewater. Our adaptive model, which adjusts parameters in real time, captures significant temporal changes in the positivity rate and provides estimates of the effective reproductive number. This approach enables more accurate and timely tracking of viral transmission dynamics. Additionally, we explore a linear mixed-effects model to forecast COVID-19 hospitalizations based on wastewater data. This model accounts for the non-independence of longitudinal data by incorporating random effects, and evaluates various predictors, including confirmed cases, test positivity rates, and wastewater concentrations. Our results show that wastewater data can effectively predict hospitalization trends and that the relationship between wastewater concentrations and hospitalizations varies across counties and infection waves.