Study: Capacity Limits Reduce Community COVID Transmission

Research conducted at Stanford and Northwestern University, published in the journal Nature, used mobility networks to examine the risk of reopening various points of interest (POI) in the community and resulting inequities in case transmission.

While some media have reported these findings to indicate gyms are a high-risk venue for COVID-19 transmission, the key findings of this study point to a different conclusion:

• gyms are less risky than many other indoor venues studied and
• the types of capacity limits that gyms have implemented can minimize risk.

Using Mobile App Data to Assess Mobility During the Pandemic

Researchers analyzed data from SafeGraph, a company that collects and aggregates anonymous location data from various mobile applications. Google was used to confirm mobility data, and the New York Times COVID-19 dashboard provided the case data. In this study, researchers analyzed data between March 1, 2020, and May 2, 2020, in 10 metropolitan areas: Atlanta, Chicago, Dallas, Houston, Los Angeles, Miami, New York, Philadelphia, San Francisco, and Washington D.C.

The data is able to show how people move between various census block groups (CBG), geographical areas containing 600-3,000 people, and visits to POI in the community, such as restaurants, gyms, and religious organizations. This data can tell researchers how far and often people traveled from their homes and which POI they visited.

The modeling also looked at how different subpopulations from various CBGs interacted as they visited different POI throughout the community and used the area, median visit duration, and time-varying density of infectious people at each POI to determine that POI’s hourly infection rate.

Key Findings: Capacity Limits Work, Socioeconomic Inequalities Persist

Some of the study’s findings are not surprising: many people moved around less between March and April, and super-spreading events drive a majority of COVID-19 cases.

The data shows mobility dropped sharply in the United States between the beginning of March and the beginning of April. In Chicago, visits to POI dropped 54%. The model predicted that had mobility not dropped at all, infections would have increased more than six-fold.

We cannot stress enough that the study looks at a situation in which mobility returned to pre-pandemic levels, which given widespread capacity limits is of limited applicability.

As previous research has suggested, a small number of venues are responsible for an outsized number of cases. In Chicago, 10% of POI were linked to 85% of cases.

The true headline of this study should be that capacity limits are an effective method for controlling COVID-19 transmission while blunting economic damage. In Chicago, the study predicted limiting capacity at 20% of maximum occupancy cut infections—compared to a full reopening—by 80%, while only costing businesses 42% of their daily visits.

Other metropolitan areas saw similar trends. Limiting maximum occupancy was also more effective at reducing infections and mitigating economic fallout than a blanket limitation on visits to each POI.

The section of this study that likely resulted in misguided headlines pertains to the modeling of reopening various POI. For this analysis, researchers modeled a full reopening of each category of POI based on March mobility levels while leaving the others closed.

They found that “on average across metro areas, full-service restaurants, gyms, hotels, cafes, religious organizations, and limited-service restaurants produced the largest predicted increases in infections when reopened.” Full-service restaurants were especially risky, with predicted 596,000 cases by the end of May if reopened at full capacity.

Models are only as good as the assumptions on which they are based. In this case, the assumption is resuming at full maximum occupancy to pre-COVID-19 levels. But no business—including fitness centers—has opened at full capacity or is even advocating for opening at full capacity.

This analysis tells us much of what we already know: if we reopen to pre-pandemic levels of capacity, which often do not facilitate adequate social distancing, then it is likely cases of COVID-19 will spike.

It is also unclear from this study the extent to which researchers factored mask-wearing, social distancing, ventilation, and hygiene practices into the assumptions. For example, it is logistically possible for a fitness center patron to wear a mask 100% of the time they are there—some states like Maryland require it—but this is not possible in a full-service restaurant, bar, or coffee shop.

Another key finding in this study and one the authors assert holds up even in the face of study limitations, is that people in lower-income CBG tend to have higher transmission rates because they are not able to limit mobility—for example, by working from home—and visit more dense POI. The authors recommend a number of policy measures, including occupancy limits, emergency food distribution, and affordable, accessible testing to address socio-economic inequities in COVID-19 infections.

These policies are important short-term measures to control COVID-19 spread. However, long-term approaches to address and prevent comorbidities like obesity and diabetes—such as regular physical activity—should not be forgotten either.