Is COVID-19 seasonal? A time series modeling approach

BackgroundDetermining whether SARS-CoV-2 is or will be seasonal like other respiratory viruses is critical for public health planning, including informing vaccine policy regarding the optimal timing for deploying booster doses. To help answer this urgent public health question, we evaluated whether COVID-19 case rates in the United States and Europe followed a seasonal pattern using time series models. MethodsWe analyzed COVID-19 data from Our World in Data from Mar 2020 through Apr 2022 for the United States (and Census Region) and five European countries (Italy, France, Germany, Spain, and the United Kingdom). For each, anomalies were identified using Twitters decomposition method and Generalized Extreme Studentized Deviate tests. We performed sensitivity analyses to determine the impact of data source (i.e., using US Centers for Disease Control and Prevention [CDC] data instead of OWID) and whether findings were similar after adjusting for multiple covariates. Finally, we determined whether our time series models accurately predicted seasonal influenza trends using US CDC FluView data. ResultsAnomaly plots detected COVID-19 rates that were higher than expected between November and March each year in the United States and Europe. In the US Southern Census Region, in addition to seasonal peaks in the fall/winter, a second peak in Aug/Sep 2021 was identified as anomalous. Results were robust to sensitivity analyses. ConclusionsOur results support employing annual protective measures against SARS-CoV-2 such as administration of seasonal booster vaccines or other non-pharmaceutical interventions in a similar timeframe as those already in place for influenza prevention. Summary of the Main PointAlthough SARS-CoV-2 continues to cause morbidity and mortality year-round due to its high transmissibility and rapid viral evolution, our results suggest that COVID-19 activity in the United States and Europe peaks during the traditional winter viral respiratory season.