The Effect of Shelter-in-Place Orders on Social Distancing and the Spread of the COVID-19 Pandemic: A Study of Texas

Abstract

Objectives: We study how the state-wide shelter-in-place order affected social distancing and the number of cases and deaths in Texas.Methods: We use daily data at the county level. The COVID-19 cases and fatalities data are from the New York Times. Social distancing measures are from SafeGraph. Both data are retrieved from the Unfolded Studio website. The county-level COVID-related policy responses are from the National Association of Counties. We use an event-study design and regression analysis to estimate the effect of the state-wide shelter-in-place order on social distancing and the number of cases and deaths.Results: We find that the growth rate of cases and deaths is significantly lower during the policy period when the percentage of the population that stays at home is highest. The crucial question is whether the policy has a causal impact on the sheltering percentages. The fact that some counties in Texas adopted local restrictive policies well before the state-wide policy helps us address this question. We do not find evidence that this top-down restrictive policy increased the percentage of the population that exercised social distancing.Discussion: Shelter-in-place policies are more effective at the local level and should go along with efforts to inform and update the public about the potential consequences of the disease and its current state in their localities.

Validation of the COVID-19 Transmission Misinformation Scale and Conditional Indirect Negative Effects on Wearing a Mask in Public

“…This rhetoric was found to be associated with lower perceived COVID-19 threat [25]. COVID-19 conspiratorial beliefs associated with less cooperation to safety guidelines [26]. By not participating in proven safety measures, such as wearing a mask, many people unnecessarily contracted the disease.…”

COVID-19 Research: Challenges to Interpret Numbers and Propose Solutions

“…This is only an example of the caution that needs to be exerted before drawing inferences from open data that are of variable quality and sometimes wrongly interpreted. On the other hand, epidemiologic studies reported in this journal have progressed our understanding of the risk factors associated with SARS-CoV-2 transmission and COVID-19 fatal outcomes (13)(14)(15)(16)(17)(18)(19). Likewise, modeling studies have shed light on effective control measures and likely trajectory of the pandemic in various settings (20)(21)(22)(23).…”

“…More detailed epidemiologic investigations revealed important determinants of epidemic spread at the local level by using diverse statistical models (17,18). De Ridder et al used a spatiotemporal cluster detection algorithm to monitor SARS-CoV-2 transmission dynamics in neighborhoods of Geneva until 30 April 2020 (17).…”

“…By using survival analysis and Cox model adjusted for population density, they found a dose-response relationship between level of socio-economic deprivation and prolonged duration of virus transmission within local clusters, highlighting the need for inequality mitigation measures as part of COVID-19 risk mitigation strategies. Castaneda and Saygili examined the county-level proportion of residents staying at home as measured by mobile device location data and COVID-19 daily case increase rates in Texas during February-May 2020 (18). They found that the growth rate of COVID-19 cases decreased when a larger proportion of the local population stayed at home.…”

“…They found that the growth rate of COVID-19 cases decreased when a larger proportion of the local population stayed at home. Interestingly, county emergency policies coincided more closely with the increase of people at home than the later State-wide order to "shelter in place, " suggesting that to reach out with an alert to the local population may be a more effective communication strategy in that setting (18). Bönisch et al assessed the effect of confinement in Germany by using an Interrupted Time-Series analysis linking actively collected population mobility data and weekly estimates of COVID-19 reproduction number before, during and after the lockdown from January to May 2020 (19).…”