A seven-day cycle in COVID-19 infection, hospitalization, and mortality rates: Do weekend social interactions kill susceptible people?

Seven-day cycles in numbers of COVID-19 new-cases and deaths are markedly evident in most public databases (e.g. Worldometer, ECDC), but it is unclear whether they reflect systematic artifacts of delays in information reporting/gathering, or have a more profound basis. To address this question we located 11 databases of US states that provide date- authenticated information (actual date of symptom onset and/or specimen collection, or actual hospitalization or death date) that reported more than 1,000 deaths each. Numbers of new cases showed a weekly cyclic pattern in 10 out of 11 states, commonly peaking on weekdays, 2-6 days after the weekend, corresponding with a reported median 5-day lag between infection and the manifestation of clinical symptoms. We postulate that this pattern emerges from interactions with different and/or extended social-circles during weekends, including increased inter-generational meetings, which in turn facilitate transfer of COVID-19 from younger people to older vulnerable individuals. Furthermore, we found weekly periodicity in hospitalizations in 2 out of 2 authenticated databases providing this information. Actual death date, which is more difficult to attribute to individual choice, and is expected to occur approximately 2-3 weeks following hospitalization, showed significant 7-day periodicity in 1 out of 11 states, and a trend in 2 additional states. If weekly peaks in new cases can be truncated by physical/social distancing, especially during weekends, the mortality of COVID-19 may be reduced, or at least hospitalization and mortality curves may be flattened.