More Than Merely Academic Interest

Opening with a quote from New York Gov. Andrew Cuomo:

“We were thinking that maybe we were going to find a higher percentage of essential employees who were getting sick because they were going to work—that these may be nurses, doctors, transit workers. That’s not the case,” he said. “They’re not working, they’re not traveling, they’re predominantly downstate, predominantly minority, predominantly older.”

the editors of the Wall Street Journal continue with some pertinent observations:

More investigation is needed, but the virus may be spreading mostly within multigenerational households or public housing. Collecting more information about infected individuals’ habits would have been especially useful early in the pandemic before most businesses were shut down. This data could reveal patterns that suggest the most likely venues for transmission.

New York Mayor Bill de Blasio said the city will start collecting employment and demographic information as part of its antibody tests to pin down how the virus has spread. Mr. Cuomo this month released final results from a random antibody test of 15,000 people across the state, which raised some questions that need to be further investigated.

For instance, estimated infections are 10 percentage points higher in the Bronx than Manhattan, which is more dense. Latinos were also more than twice as likely to have antibodies than Asians. Low-income folks are more likely to work in jobs interfacing with customers, but other behavioral differences may account for their antibody disparities.

Surprisingly, millennials who are known to crowd bars and clubs were about as likely to carry antibodies as baby boomers. Even more curious, 12% of health-care workers tested positive for antibodies compared to 20% of the general population. This is good news since it suggests that the virus isn’t mainly being spread via the health-care system. Surveying more people through antibody and diagnostic tests could help experts better identify the major transmission vehicles.

Scientists often use such observational studies to identify risk factors for diseases when they can’t do randomized experiments. Regression analysis can help discern the variables that most influence an outcome while controlling for others. While studies can’t control for all confounding variables, they can still show important patterns.

Incorporating such surveys in random population tests could also let states reopen more safely. People could be asked how often each week they take mass transit, dine out, use a gym or visit a salon. Those who test positive could be compared to those who test negative, controlling for other factors.

This is of much more than merely academic interest. The measures that have been taken to slow the increase in COVID-19 cases including social distancing, “stay at home” directives, closing of public facilities like parks, and face mask directives, are all guesses. Sometimes they’re informed guesses; sometimes less so. It’s important to know whether sitting in a restaurant with other diners puts you at risk. It’s important to instill confidence in the public and it’s important to the restaurant—it may determine whether the enterprise remains viable. Theoretically, there may be a risk but in practice the risk may be so low as to be negligible. As Yogi Berra put it, in theory there’s no difference between practice and theory but in practice there is.