Philip Gigliotti1, Erika G Martin. 1. Department of Public Administration and Policy, Rockefeller College of Public Affairs and Policy, University at Albany, State University of New York (SUNY), Albany, New York.
Abstract
OBJECTIVE: To evaluate predictors of stay-at-home order adoption among US states, as well as associations between order enactment and residents' mobility. DESIGN: We assess associations between state characteristics and adoption timing. We also assess associations between enactment and aggregate state-level measures of residents' mobility (Google COVID-19 Community Mobility Reports). SETTING: The United States. PARTICIPANTS: Adoption population: 50 US states and District of Columbia. Mobility population: state residents using devices with GPS tracking accessible by Google. INTERVENTION AND EXPOSURES: State characteristics: COVID-19 diagnoses per capita, 2016 Trump vote share, Republican governor, Medicaid expansion status, hospital beds per capita, public health funding per capita, state and local tax revenue per capita, median household income, population, percent residents 65 years or older, and percent urban residents. Mobility exposure: indicator of order enactment by March 29, 2020 (date of mobility data collection). MAIN OUTCOME MEASURES: Order adoption timing: days since adoption of first order. Mobility: changes in mobility to 6 locations from February 6 to March 29, 2020. RESULTS: In bivariate models, order adoption was associated with COVID-19 diagnoses (hazard ratio [HR] = 1.01; 95% confidence interval [CI], 1.00 to 1.01), Republican governor (HR = 0.24; 95% CI, 0.13 to 0.44), Medicaid expansion (HR = 2.50; 95% CI, 1.40 to 4.48), and hospital capacity (HR = 0.43; 95% CI, 0.26 to 0.70), consistent with findings in the multivariate models. Order enactment was positively associated with time at home (beta (B) = 1.31; 95% CI, 0.35 to 2.28) and negatively associated with time at retail and recreation (B = -7.17; 95% CI, -10.89 to -3.46) and grocery and pharmacy (B = -8.28; 95% CI, -11.97 to -4.59) locations. Trump vote share was associated with increased mobility for 4 of 6 mobility measures. CONCLUSIONS AND RELEVANCE: While politics influenced order adoption, public health considerations were equally influential. While orders were associated with decreased mobility, political ideology was associated with increased mobility under social distancing policies.
OBJECTIVE: To evaluate predictors of stay-at-home order adoption among US states, as well as associations between order enactment and residents' mobility. DESIGN: We assess associations between state characteristics and adoption timing. We also assess associations between enactment and aggregate state-level measures of residents' mobility (Google COVID-19 Community Mobility Reports). SETTING: The United States. PARTICIPANTS: Adoption population: 50 US states and District of Columbia. Mobility population: state residents using devices with GPS tracking accessible by Google. INTERVENTION AND EXPOSURES: State characteristics: COVID-19 diagnoses per capita, 2016 Trump vote share, Republican governor, Medicaid expansion status, hospital beds per capita, public health funding per capita, state and local tax revenue per capita, median household income, population, percent residents 65 years or older, and percent urban residents. Mobility exposure: indicator of order enactment by March 29, 2020 (date of mobility data collection). MAIN OUTCOME MEASURES: Order adoption timing: days since adoption of first order. Mobility: changes in mobility to 6 locations from February 6 to March 29, 2020. RESULTS: In bivariate models, order adoption was associated with COVID-19 diagnoses (hazard ratio [HR] = 1.01; 95% confidence interval [CI], 1.00 to 1.01), Republican governor (HR = 0.24; 95% CI, 0.13 to 0.44), Medicaid expansion (HR = 2.50; 95% CI, 1.40 to 4.48), and hospital capacity (HR = 0.43; 95% CI, 0.26 to 0.70), consistent with findings in the multivariate models. Order enactment was positively associated with time at home (beta (B) = 1.31; 95% CI, 0.35 to 2.28) and negatively associated with time at retail and recreation (B = -7.17; 95% CI, -10.89 to -3.46) and grocery and pharmacy (B = -8.28; 95% CI, -11.97 to -4.59) locations. Trump vote share was associated with increased mobility for 4 of 6 mobility measures. CONCLUSIONS AND RELEVANCE: While politics influenced order adoption, public health considerations were equally influential. While orders were associated with decreased mobility, political ideology was associated with increased mobility under social distancing policies.