Prominent Spatiotemporal Waves of COVID-19 Incidence in the United States: Implications for Causality, Forecasting, and Control

Abstract

Better understanding of the spatiotemporal structure of the COVID-19 epidemic in the USA may help inform more effective prevention and control strategies. By analyzing daily COVID-19 case data in the United States, Mexico and Canada, we found four continental-scale epidemic wave patterns, including travelling waves, that spanned multiple state and even international boundaries. These major epidemic patterns co-varied strongly with continental-scale seasonal temperature change patterns. Geo-contiguous states shared similar timing and amplitude of epidemic wave patterns irrespective of similarities or differences in state government political party affiliations. These analyses provide evidence that seasonal factors, probably weather changes, have exerted major effects on local COVID-19 incidence rates. Seasonal wave patterns observed during the first year of the epidemic may become repeated in the subsequent years.

Epidemic tracking and forecasting: Lessons learned from a tumultuous year

“…To reliably forecast the progression of pandemics, where relevant historical data are almost nonexistent, we must have a detailed quantitative understanding of how different, diverse factors affect disease transmissibility. Such an understanding is currently grossly lacking, as evidenced by our collective failure to predict (42) [or even understand post hoc (43)] the high-level temporal and geographic contours of the main pandemic waves in the United States. Yet, this very pandemic, the most instrumented in human history, is also a rare opportunity to attempt this vital scientific and technological goal.…”

Community incidence patterns drive the risk of SARS-CoV-2 outbreaks and alter intervention impacts in a high-risk institutional setting

“…In addition, the timing and magnitude of peak incidence at the state or county level has fluctuated over the course of these waves. There have also been regional outbreaks outside of these waves ( 14 ). To assess the risk of an outbreak in high-risk settings, it is therefore important to estimate the importation risk based on the current epidemiological context of the broader community in which the setting is situated.…”