Response to the Stanford/Northwestern Study on COVID-19 Spread from the Texas Restaurant Association

Nov 11

The recent Stanford/Northwestern study rightfully concludes that disadvantage groups are experiencing worse outcomes from COVID-19 and that policymakers should address this disparate impact. However, the Texas Restaurant Association (TRA) has very serious concerns about the research methodology and the conclusions it drives regarding the risk of reopening different businesses, including restaurants. We’ve detailed some of these concerns below. 

TRA continues to call for health data, not assumptions, to inform our response to COVID-19. When information like the Stanford/Northwestern study’s assumptions is taken out of context and promoted, we confuse the public and miss critical opportunities to implement evidence-based strategies that will bend the curve. We’re eight months into this pandemic, and the stakes are simply too high for our public health and for our economy to continue to let sensational news headlines—rather than health data—drive our collective efforts to combat COVID-19. Restaurants will continue to implement all of the evidence-based strategies doctors advise so we can provide a safe alternative to the unregulated, in-home gatherings that we know are driving the spread of COVID-19. 

Some of our specific concerns with the Stanford/Northwestern study include: 

  • First, the study is a computer model that predicts trends based on location data. This is very different than doctors and epidemiologists studying actual cases of COVID-19. Considering the life-altering impact of COVID-19, actual data from patients is the most relative data that should be studied and should drive policy decisions.  
  • In fact, when you read the fine print of the study, it’s clear key assumptions within the study are based, not on data, but on pre-existing assumptions. The authors explain: “As the reviewer notes, we assume the transmission rate at a [point of interest (POI)] in a given hour depends on two key ingredients: how much time visitors spend there, and the density (number of visitors per sq ft) of the POI in that hour. These assumptions are based on prior expectations that a visit is more dangerous if you spend more time there and/or if the location is more crowded…. We agree that contract tracing data would have been ideal for testing our model’s predictions on the relative riskiness of different POI types; however, as the reviewer notes, this data is hard to come by. We opted instead for a second-best solution, as we searched for independent epidemiological experts who have also ranked POI types by their relative riskiness.” In other words, critical points of the analysis to determine relative risk rely on assumptions. 
  • In addition, the study focuses on what it calls “non-residential location visits,” but we know a significant source of COVID-19 spread is home gatherings. Just this week, the CDC published guidelines stating, “small household gatherings are an important contributor to the rise in COVID-19 cases.” How can a model that discounts this significant source of spread be reliable?  
  • Similarly, the study’s analysis leaves out child care services, elementary and secondary schools, drinking places, hospitals, and airports, all of which are generally thought to contribute significantly to the spread of COVID-19 due to large numbers of people in confined spaces and the length of time they spend in those spaces. Both are the key factors of the model used in this study, and yet, these entities were not included in its detailed analysis. Also critical to our concerns with this study is that the distinctions between restaurants and drinking places, as well as indoor and outdoor dining, are blurred.  
  • Relatedly, because the study assumes transmission rate is dependent on just two factors, it seemingly ignores other factors that scientists tell us are critical to reducing transmission, including compliance with COVID-19 safety protocols like social distancing, wearing a mask, and increased sanitation. Also disregarded are the precautions many restaurants are taking to mitigate risk in enclosed spaces like installing plexiglass barriers, improving air circulation, and serving people outside. Data from the Texas Alcoholic Beverage Commission repeatedly confirms that Texas restaurants have incredibly high rates of compliance with the state’s comprehensive COVID-19 safety protocols, with around 1% of inspections resulting in a warning or suspension.   
  • Finally, the timeline of the study is concerning because it relies on data collected from March 1 - May 2, 2020. Texas required restaurant dining rooms to close completely for most of this period, and since then, much has changed, including statewide mask mandates and stronger compliance from businesses, consumers, and the general public. This time frame is simply not reflective of our state’s COVID-19 journey. 


Add new comment

Use your first and last name as your username.
This question is for testing whether or not you are a human visitor and to prevent automated spam submissions.
5 + 6 =
More information?

* indicates a required field.