ROCHESTER, Minn. — In America, wealth can buy you isolation and uncrowded spaces, and that is a fact about inequality that has made COVID-19 disproportionately dangerous to the working poor.
It has also made restaurants, especially those more crowded venues serving low-income communities, to the coronavirus pandemic what the Broad Street pump was to cholera in 1854 London: an epicenter of spread (although hardly as exclusively so).
That's the finding of an ambitious and unique study released this week charting the relationship between COVID-19 outbreaks, economic status, shop density and daily travel patterns for 98 million people at over 500,000 stores within 10 major metropolitan areas at the start of the epidemic.
For eight of the 10 of the cities studied, infection rates were twice as high in low-income areas as high-income areas.
The study, conducted by an interdisciplinary team of computer scientists, preventive medicine specialists and sociologists from Stanford and Northwestern universities in Palo Alto and Chicago, looked at anonymous cell phone mobility (tracking) data, focusing on Atlanta, Chicago, Dallas, Houston, Los Angeles, Miami, New York, Philadelphia and Washington, D.C.
One caveat for the study's usefulness in relation to the outbreaks now occurring in the Midwest, is that traffic patterns and scale of venues visited in major metropolitan areas are oftentimes denser to begin with.
Another caveat is that the study worked from mathematical modeling, which, thanks to some wild early attempts at using equations to predict the spread of the virus, has taken something of a black eye.
But in this case, the teams with pocket protectors had a secret weapon -- anonymous cell phone data that told them the movements of people down to a shop-by-shop basis.
The results have come to the fore during a week in which Minnesota Gov. Tim Walz, citing 71 % of infection linked to bars and small gatherings, has turned back its dials in the direction of targeted mitigation.
On Friday, Minnesota ordered the closing of food and beverage service at 10 p.m. for bars, restaurants and receptions, in addition to limiting all venues to 150-person maximums, halting bartop seating and reducing the size of small gatherings to 10 persons from three households.
The new cell phone study came along in the midst of these changes, as did an update of the restaurants connected to spread within Minnesota, and together they underscore the importance of identifying how and where Minnesotans and especially lower-income Minnesotans who face higher risk of spread gather.
"We need to understand how people come in contact with each other," said Jure Leskovec, an associate professor of computer science at Stanford University and coauthor of "Mobility network models of COVID-19 explain inequities and inform reopening," a study published last week in the journal Nature.
"Our main insight in the exercise was to use anonymized large-scale data coming from cell phones to understand contact patterns between individuals in the population," he explained. "It told us how people from different neighborhoods visited different types of places like parks, grocery stores, schools, churches and so on -- how long they stayed there and what is the area of that place."
"Using this data we can find where and when people were getting infected," co-author Serina Chang said, "We found that 10 percent of points of interest accounted for 80 percent of infections in the city."
The study implicated restaurants, gyms, grocery stores, hotels, cafe's and religious organizations, with restaurants four times more risky than other venues, and low income grocers twice as risky.
It also found that low-income stores were more crowded (they had almost 60% more hourly visitors per square foot), and that visitors stayed 17% longer on average.
The authors said the results argued for COVID-19 policies based on density caps. They said health officials should develop alternate food distribution options for the low-income who cannot avoid crowded options that require long visits.
Minnesota restaurants associated with outbreaks, some commonalities
It's been 10 days since health officials updated the list of restaurants associated with outbreaks in Minnesota, and the new reporting tells how a diverse set of venues were driving spread in the months leading up to the third wave now engulfing the state.
Though health officials have specifically named weddings, funerals and small informal gatherings as the primary source of spread, the state has learned that thousands of new cases were tied to restaurants and bars.
The state identified 117 outbreaks as identified by "seven patrons from at least three different households who report visiting only that establishment during the month before their start of symptoms," according to Doug Schultz of the Minnesota Department of Health.
Of the 2,406 unique cases tied to outbreaks in restaurants, "it’s possible the person could have become infected elsewhere," Schultz says, "but it is also very likely that transmission was occurring at the establishment with seven or more unique cases."
The venues with highest cases since September included both chains and proprietor-owned restaurants, with the latter hosting more cases, but the former hosting outbreaks in more parts of the state.
For October, the highest number of cases was tied to a Buffalo Wild Wings in Mankato, with 26 cases, 18 reporting only having been at the establishment. El Patron in Winona came in second at 24 and 12, respectively.
Maple Tavern and Maynards in the north metro each had 23 cases in October, 12 having only been tied to those venues. The Texas Roadhouse chain had outbreaks in locations across the state, including Coon Rapids, Duluth, Rochester, Waite Park, and Woodbury.
Applebees were tied to outbreaks in seven locations across the state, with all of them outstate, including Baxter, Bemidji, Grand Rapids, Mankato, New Ulm, Roseville and St. Cloud.