Assessing the interactions between COVID-19 and influenza in the United States

Casey M Zipfel; Shweta Bansal

Our study has several limitations. Our use of ILI data relies on healthcare-seeking behavior which is known to be heterogeneous spatially and across age groups [15] . Additionally, we selected seasons antigenically similar to the 2019-2020 season for comparison, but the current season displayed unusually high levels of influenza B activity. Our regression analysis is limited by known factors or available data. The humidity data assumes that each US state is represented by a random weather station. We use information seeking behavior measured by Google Trends as representative of COVID-19 and flu risk perception; while past work has shown such data to be a useful proxy [18] , they are limited in aggregate and have not been fully validated. Lastly, our epidemiological model, while it captures local and global contact structures relevant to influenza transmission, models interventions and seasonality simplistically.