Non-pharmaceutical interventions and covid-19 burden in the United States: retrospective, observational cohort study.

Objective: To evaluate the adoption and discontinuation of four broadly used non-pharmaceutical interventions on shifts in the covid-19 burden among US states. Design: Retrospective, observational cohort study. Setting: US state data on covid-19 between 19 January 2020 and 7 March 2021. Participants: US population with a diagnosis of covid-19. Main outcome measures: Empirically derived breakpoints in case and mortality velocities (ie, rate of change) were used to identify periods of stable, decreasing, or increasing covid-19 burden. Associations between adoption of non-pharmaceutical interventions and subsequent decreases in case or death rates were estimated by use of generalised linear models accounting for weekly variability across US states. State level case and mortality counts per day were obtained from the Covid-19 Tracking Project. State level policies on non-pharmaceutical interventions included stay-at-home orders, indoor public gathering bans (mild >10 or severe ≤10 people), indoor restaurant dining bans, and public mask mandates. National policies were not included in statistical models. Results: 28 602 830 cases and 511 899 deaths were recorded during the study. Odds of a reduction in covid-19 case velocity increased for stay-at-home orders (odds ratio 2.02, 95% confidence interval 1.63 to 2.52), indoor dining bans (1.62, 1.25 to 2.10), public mask mandates (2.18, 1.47 to 3.23), and severe indoor public gathering bans (1.68, 1.31 to 2.16) in univariate analysis. In mutually adjusted models, odds remained elevated for orders to stay at home (adjusted odds ratio 1.47, 95% confidence interval 1.04 to 2.07) and public mask mandates (2.27, 1.51 to 3.41). Stay-at-home orders (odds ratio 2.00, 95% confidence interval 1.53 to 2.62; adjusted odds ratio 1.89, 95% confidence interval 1.25 to 2.87) was also associated with a greater likelihood of decrease in death velocity in unadjusted and adjusted models. Conclusions: State level non-pharmaceutical interventions used in the US during the covid-19 pandemic, in particular stay-at-home orders, were associated with a decreased covid-19 burden.