Evaluating the impact of keeping indoor dining closed on COVID-19 rates among large US cities: a quasi-experimental design

Introduction

The COVID-19 pandemic has directly caused more than 560,000 deaths and 31 million cases through April 2021 in the US alone. Key sources of transmission include crowded and poorly ventilated indoor spaces, including indoor dining areas¹. Understanding which Non-Pharmaceutical Interventions (NPIs) reduce transmission is crucial for balancing public health and economic and social cost from COVID-19. Previous research suggests that mask mandates², sick leave policies³, and shelter in place orders reduce the spread of COVID-19⁴, although the quality of studies concerning the causal evaluation of NPIs is mixed⁵. Many analyses evaluate state-level policies, masking substantial city-and county-level heterogeneity.

During the 2020 summer and fall, states and local governments in the US were left to negotiate closing and re-opening policies without federal regulation or guidance related to activities such as indoor dining. In the early months of the pandemic, most cities and states limited indoor dining, while allowing delivery, take-out, and curb-side pickup services, and sometimes, outdoor dining. As the pandemic progressed some cities and states began reopening indoor dining while others remained closed⁶. Other cities attempted to keep indoor dining closed but were prohibited from doing so, and forced to re-open, by their state.

The federal system in the US delegates much of the authority to protect health to states, under the police powers, and states can then further delegate authority to local governments⁷. The authority to prohibit or limit a lower level of government’s power to enact legislation, is a legislative doctrine known as government preemption⁷. States have preempted local government on laws including minimum wage, bans on removing confederate statues, anti-discrimination polices, and anti-gun legislation⁷. Recently, preemption laws have been adopted by predominantly white and conservative legislatures to limit progressive policies in democratic and majority non-white cities⁸. For example, some states have preempted cities from keeping indoor dining closed, forcing cities to re-open indoor dining during the COVID-19 pandemic⁹.

State preemption of city law around indoor dining created substantial heterogeneity between cities in the scope of local governments’ authority to regulate indoor dining and provides an opportunity to test the association between keeping indoor dining closed and COVID-19 rates. The objective of this study is to examine the association between keeping indoor dining closed and COVID-19 rates, leveraging the variation in local government power to enact these orders. This variation creates a counterfactual contrast that we exploit, by comparing cities that kept indoor dining closed to cities that were forced re-open indoor dining by their respective state governments.

Methods

Study Setting

The study population includes cities that are members of the Big Cities Health Coalition (BCHC), an organization comprised of America’s largest metropolitan health departments. We collected information on indoor dining re-opening policies from searching and reviewing city and state orders available on state and city websites.

Following a policy trial emulation framework¹⁰ we define units of analysis and exposure, causal contrasts, outcomes and time-zero. Table 1 and Figure 1 contain details on these design specifications, and a summary follows. The main unit of analysis were cities that explicitly expressed an intention to keep indoor dining closed even after their respective states allowed them to reopen indoor dining. Cities were then divided into treatment and comparison groups, according to whether their respective state governments actually allowed them to continue keeping indoor dining closed. The treatment group includes cities that kept indoor dining closed, while the state allowed indoor dining to reopen. The comparison group includes cities that intended to keep indoor dining closed but were preempted from doing so by the state and so reopened indoor dining. By identifying comparison cities that “would have” remained closed absent the preemption, we provide a stronger counterfactual comparison group than comparing all geographies that remained closed to those that re-opened, or studies that compare policy adoption over time¹¹, as selection into the treatment group may be biased by other city characteristics.

View this table:

• View inline
• View popup
• Download powerpoint

Table 1: Treatment and Comparison Group Definitions

• Download figure
• Open in new tab

Figure 1: Treatment and Comparison group selection flowchart

^aSee Appendix B1 for evidence that comparison cities meet the counterfactual condition that they “would have stayed closed” absent the state preemption.

Figure 1 shows a flowchart of treatment and comparison group selection. We examined the spring/summer period of re-openings after initial stay-at-home orders, and so limited the sample of cities to those that re-opened or were allowed to re-open between March 30, 2020 and October 1, 2020. We excluded cities that did not demonstrate intention to stay closed and then divided cities into treatment and comparison groups. Appendix B provides evidence that comparison cities would have stayed closed absent the preemption order, along with evidence on treatment decisions to remain closed and details on subsequent re-openings in treatment cities. We also excluded treatment cities that re-opened dining less than 3-weeks after the state allowed them to re-open, to ensure a sufficient period of continued closure to impact COVID-19 rates. To increase the number of cities, we included cities that are not in the BCHC but that are among the largest cities in the country, and otherwise met inclusion criteria (opened or allowed to re-open 3/2/20-10/1/20, expressed desire to re-open, and if in treatment, re-opened more than 3-weeks after allowed).

Preemption also helped us to identify “time-zero”¹². For treatment cities, time-zero is the date the city kept indoor dining closed when the state allowed re-opening, and for comparison cities, this is the date the state mandated (via preemption) the city re-open indoor dining, or when the city, “would have” stayed closed (see Appendix Table A2 for city and state dates). Incorrectly identifying the time when the comparison group “could have” enacted the policy can bias estimates by inappropriately conditioning on or selecting on post-treatment variables, and the preemption allows for easy identification of time-zero in our comparison group¹⁰. To isolate the immediate impacts of keeping dining closed we limited the study period to 10 weeks total: the 4 weeks before and 6 weeks after “time zero”. Throughout the article, “cities” refers to the city itself, or the county in which the majority of the city resides, depending on data availability (see Appendix Table A2).

Results

We included 11 cities in our analysis, 4 treatment and 7 comparison, across 8 states. The 11 cities contributed a total of 781 city-day observations and accounted for a total of 88,589 cases in the 10-week study period (4 weeks before and 6 weeks after “time-zero”). Table 1 displays descriptive statistics on treatment and comparison cities: groups were comparable in terms of population size, age distributions, socioeconomic conditions (poverty rates and education levels), percent non-Hispanic white, foreign born, housing crowding, service workers, masking levels, and political leaning. Percentage Hispanic was larger, and percentage non-Hispanic Black smaller, in comparison cities, and public transit use was more common in treatment cities. Indoor dining was closed for an average of 12.8 and 5.6 weeks before cities were allowed to re-open dining (in treatment cities) and before having to open (in comparison cities), respectively. Treatment cities remained closed for a mean of 5.1 weeks after they were allowed to re-open. Appendix Figure A1 shows the pattern of COVID-19 cases and timing of indoor dining closing and re-opening from April to October; for treatment cities, the figure shows the date states allowed re-opening and the date the city eventually re-opened. In cities that reopened we saw an increase in cases approximately two weeks after indoor dining re-opened, while case rates decreased or continued to decrease in cities that maintained indoor dining closures.

On average, treatment and comparison cities had 11 and 4 daily cases per 100,000, respectively, before reopening indoor dining or keeping indoor dining closed (Table 1). But, as shown in Figure 2 pre-treatment case rate trends were approximately parallel. Results from the models show a strong association between COVID-19 rates and keeping indoor dining closed. Keeping indoor dining closed reduced COVID-19 incidence by 43% (IRR=0.57, 95% CI 0.46 to 0.69) compared with cities that reopened indoor dining, in the fully adjusted model (Table 2).

• Download figure
• Open in new tab

Figure 2: Trends in Case Rates Comparing Treatment and Comparison Groups

Vertical black line represents a 2-week lag after re-opening (comparison cities) or the date the state allowed the city to re-open but the city stayed closed (treatment cities). Comparison cities include Atlanta, Austin, Charleston, Dallas, Houston, Phoenix, and San Antonio; treatment cities include Indianapolis, Milwaukee, Philadelphia, and San Francisco.

Keeping indoor dining closed averted 91 (95% Confidence Interval [CI]: −168, −15) daily cases in the average city. Had all cities kept indoor dining closed (and holding all other NPIs at their actual levels), we estimate that approximately 27,914 (CI: −51253, −4575), or 133 per 100,000, cases would have been averted over the four-week period. These estimates should be viewed cautiously and as general approximations.

Discussion

In this study of 11 large cities in the US with a total of over twenty-two million inhabitants, we found that keeping indoor dining closed significantly reduced COVID-19 incidence as compared to re-opening. We approached this question leveraging the heterogeneity in state and local policies, specifically comparing cities that kept indoor dining closed to cities that were preempted from doing so but would have kept indoor dining closed if allowed to. Our sensitivity analyses all agreed in the direction of association, enforcing the robustness of these findings. We estimate that, after adjusting for other NPIs, COVID-19 rates decreased by 43% over four weeks in cities that kept indoor dining open compared to cities that re-opened.

In a review of 20 articles using different methods, Bilal et al. found that reopening hospitality venues (bars, restaurants, and night clubs) posed a high-risk for increases in COVID-19 incidence, that closing them was among the most effective COVID-19 mitigation strategies, and that these venues were frequent places for superspreading²². However, most reviewed studies did not differentiate between indoor and outdoor dining, so could not isolate impacts of indoor dining on COVID-19 rates. For example, a recent CDC quasi-experimental analysis found that re-opening “on-premise” dining was associated with an increase in cases 41-100 days after allowing indoor dining¹¹, and another CDC analysis found that people who tested positive were twice as likely to have reported eating at a restaurant²³. Another US study using fine-grained mobility data found that compared with reopening other non-residential locations, reopening full-service restaurants was associated with the largest predicted impact on infections¹. A quasi-experimental study found that closures of restaurant dining areas, in addition to entertainment venues, gyms, and bars, had impacts similar in magnitude to shelter-in-place orders⁴. And, a contact tracing study from Hong Kong found that social venues, including restaurants, accounted for 33% of all traced transmissions, and present an elevated risk for large outbreaks²⁴.

While research consistently suggests increases in COVID-19 due to indoor dining, our article is the first to isolate indoor dining independent of outdoor dining using a quasi-experimental design. Additionally, our comparison group satisfies several conditions suggesting a strong counterfactual and our use of preemption improves upon prior analysis that cannot control for differences influencing the likelihood that a geography implements an NPI (selection into the comparison group). The analysis also highlights differences in state and local policy decision powers and authorities, a key policy pandemic question as local government often enacted, or attempted to enact, stricter NPI policies than state governors, producing heterogeneous within-state COVID-19 patterns²⁵. Indoor dining has not only been associated with increased COVID-19 incidence, but also has been shown to potentially be responsible for some of the stark racial/ethnic disparities in COVID-19 incidence and mortality¹. Higher death rates in Black and Hispanic populations are driven by higher rates of infection and exposure, in large part thought to be due to higher rates of occupational exposure²⁶. Analyses of occupation and COVID-19 deaths found that food preparation and serving workers²⁷, and food and agricultural workers²⁸, had among the highest rates of COVID-19 deaths of all workers, and among food preparation and service workers, Hispanic workers had 8 times higher death rates as white workers. Given disproportionate representation among restaurant workers - Hispanic/Latinos account for 27% of food and restaurant workers but only 17% of the population²⁹ - reopening may increase risk of exposure, particularly among Hispanic/Latino populations. While our study could not estimate impacts of indoor dining on inequities in COVID-19, due to lack of longitudinal data on neighborhood level or race specific COVID-19 rates across cities, future studies should extend this work to examine disparate impacts of indoor dining policies.

Restaurant groups and lobbyists sent letters to governors and mayors advocating for restaurant re-opening, and in some cases sued (successfully and unsuccessfully) governors and mayors³⁰, citing the lack of research on the relationship between indoor dining and COVID rates and superspreading events³¹. Our study cannot speak to impacts of indoor dining on superspreading events, but we do provide robust evidence that indoor dining bans are effective mitigation measures, evidence that can help US and other country courts make determinations about the public health purpose of indoor dining bans. Nearly all cities and states across the US have re-opened indoor dining in the Spring, despite increasing cases in the Northeast and Michigan. These re-opening decisions involve difficult tradeoffs between protecting public health and preventing further layoffs and restaurant closures but have been made more difficult because of limited strong evidence on the association between indoor dining and COVID-19 spread.

Our study has several limitations. The cities are similar on multiple potential confounders that may affect COVID incidence and NPI implementation and compliance, including political leaning, age structure, socioeconomic and housing factors, service workers percentage and mask compliance. However, only a small number of cities met study inclusion criteria, and treatment cities were smaller, had larger, Black and smaller Hispanic populations, and used transit more frequently than comparison cities. These differences should not bias our difference-in-difference analysis unless they were changing significantly during the study period. However, our treatment and comparison groups may differ on other factors related to COVID-19 rates, and our approach does not capture relevant unmeasured or unobservable time-varying between-city confounders, that may have increased COVID-19 transmission, such as mobility data or longitudinal test count, which is not available at the city/county level. Re-opening indoor dining timing in some states coincided with other re-openings such as museums, malls, and theaters. Though contact tracing data suggests that such activities pose relatively limited risk for COVID-19 transmission³², if re-opening other non-essential leisure activities also contributed to increased infections then our results may slightly overestimate the association attributable to indoor dining. Importantly, and given transmission dynamics of an infectious disease like COVID-19, the difference between groups in pre-trend levels of COVID-19 incidence may impact the subsequent rate of change²¹, and people may have been less cautious in cities with lower baseline rates. Our model measures associations with indoor dining policies, not the act of engaging in indoor dining specifically, so also captures cases due to policy-related behavior change, and policies re-opening dining may signal reduced risk to residents, leading to more risky activities. Last, we only measured a binary indicator of re-opening (open/closed), though different capacity levels (e.g., 50% vs 25%) or restrictions may differentially impact COVID spread¹ and future research should examine the impacts of these forms of policy implementation.

Public Health Implications

Evidence from the 1918 influenza pandemic showed that cities with more stringent public health measures had better economic performance following the pandemic³³. The closure of indoor dining has devastated the restaurant industry and employees³¹. Providing economic support to the restaurant industry while indoor dining remains closed, while continuing to allow lower-risk activities such as take-out, delivery, and outdoor dining, can achieve COVID-19 mitigation goals without sacrificing economic security. Moreover, if we hope to re-open or keep open schools, a priority for the Biden administration³⁴, COVID-19 incidence rates should be as low as possible³⁵. Our study shows that one way to achieve this is through keeping indoor dining closed.

Keeping indoor dining closed is one tool to prevent further spread of COVID-19. Cities and states have re-opened indoor dining in the spring, despite increasing cases on the East Coast and Michigan, and increasing predominance of more contagious strains. Though vaccination rates are rapidly increasing in the US, large vulnerable populations remain unvaccinated, and other countries have far lower vaccination coverage. This study suggests that keeping indoor dining closed can prevent thousands of COVID-19 cases.