Spatial impact of mobility changes and non-pharmaceutical interventions on COVID-19: A county level study in the contiguous United States.

Abstract:

In early February 2020 the novel coronavirus (COVID-19) began to spread via community transmission in the USA. From individuals to institutions, the response to this growing pandemic became evident through changes in mobility patterns and varying restrictions and non-pharmaceutical interventions (NPIs). This study analyzes the impact of mobility changes and NPI measures on the spread of COVID-19 in the US. We examine different factors of mobility and NPIs and specify a spatial panel to avoid biased and inefficient estimation. This study covers 1968 counties in the contiguous United States for 62 days (March 1^st, 2020 – May 1^st, 2020). We utilize daily county level COVID-19 cases per 100,000 people along with a total of six NPI measures and fixe mobility classes. We also determine and include spatial effects from the model estimation. Our investigation suggests that under the model assumptions during the study period, reductions in workplace related mobility had the greatest impact on reducing the COVID-19 case rate growth, followed by visits to retail places. The direct effects from the model estimates that, a 1% increase in workplace mobility increases the rate of COVID-19 by approximately 2.5%. Similarly, a 1% increase in retail and transit visits will increase COVID-19 growth rates by approximately 1.8% and 1%, respectively. On the policy side, workplace closure was subsequently found to be most effective in decreasing the spread. A one percent increase in retail will increase COVID-19 cases by 1.8% and a one percent increase in transit will lead to a 1% increase in COVID-19 cases. Policies related to cancelation of public events and various public information campaigns were also found to be effective.