Enacting national social distancing policies corresponds with dramatic reduction in COVID19 infection rates

Abstract

The outbreak the SARS-CoV-2 (CoV-2) virus has resulted in over 6.5 million cases of COVID19, greatly stressing global healthcare infrastructure. Lacking medical prophylactic measures to combat disease spread, many nations have adopted social distancing policies in order to mitigate transmission of CoV-2. While mathematical models have suggested the efficacy of social distancing to curb the spread of CoV-2, there is a lack of systematic studies to quantify the real-world efficacy of these approaches. Here, we first demonstrate that implementation of social distancing policies in US states corresponded with a reduction in COVID19 spread rates, and that the reduction in spread rate is proportional to the average change in mobility. We validate this observation on a worldwide scale by analyzing COVID19 spread rate in 134 nations with varying social distancing policies. Globally, we find that social distancing policies significantly reduced the COVID19 spread rate, with resulting in an estimated 65% reduction (95% CI = 39-80%) in new COVID19 cases over a two week time period. These data suggest that social distancing policies may be a powerful tool to prevent spread of COVID19 in real-world scenarios.

Fig 1. Social distancing policies corresponded with inhibited spread of COVID19 in the United States.

(A) Fitting of COVID19 initial spread rate and spread rate following social distancing implementation in Idaho. (B) Fitting of COVID19 initial spread rate and a second time interval matched to post-social distancing in Nebraska where no policy was implemented. (C) Heat map showing change in COVID19 spread rate after implementing statewide social distancing policy, or a time-matched interval in states without social distancing policies (Nebraska, Wyoming, and South Dakota). (D) Bar graph showing values in (C), with states that did not implement social distancing policies shown in red. Error bars indicate 95% confidence interval. Rank-sum test.

Fig 2. Social distancing policies correspond with reduced average mobility in the United States.

(A) Average mobility from Google mobility reports prior to implementation of social distancing and after implementation of social distancing in US states that enacted social distancing policies. Signed rank test. (B) Change in average mobility determined from Google mobility reports following implementation of social distancing policies (blue), or time-equated periods in states with no social distancing policies (red). Median with interquartile range. Rank-sum test. (C) Correlation of change in COVID19 spread rate and change in average mobility. Spearman correlation coefficient.

Fig 3. Global reduction of COVID19 spread reduced mobility through social distancing policies.

(A) Change in COVID19 spread rate following implementation of social distancing policy, or time matched interval in countries with no policy. Kruskal-Wallis test with Dunn’s posthoc. (B) Change in average mobility from Google mobility reports following implementation of social distancing policy, or time matched interval in countries with no policy. Kruskal-Wallis test with Dunn’s posthoc. (C) Correlation of change in COVID19 spread rate and change in average mobility. N = 84 countries. Spearman correlation coefficient. (D) Analysis of change in COVID19 spread rates per (A) only in nations with overlapping changes in average mobility, defined as from -56 to -22. Rank-sum test.

Fig 4. Prevention of new COVID19 cases by social distancing policies.

(A) Change in COVID19 spread rates following implementation of social distancing policies (blue), or time-equated periods in countries with no social distancing policies (red). Data were fit using a generalized linear mixed-effects model taking each country as a random effect. Bars represent change in regression coefficient ± standard error. (B) Modeling of new COVID19 cases per million inhabitants in countries with social distancing policies using either model fit directly to data following implementation of social distancing policies (blue), or after correcting for the observed reduction in COVID19 spread rates associated with social distancing policies (red). Points represent model COVID19 cases per million ± standard error. (C) Estimation of total new cases over a two-week period if countries had not implemented social distancing policies (red), with implementation of social distancing policies (light blue), and actual new cases (dark blue). Bars represent number of COVID19 cases in the countries with social distancing policies ± standard error.