I’m sort of surprised this NYT article describing an unreviewed manuscript about gym use and COVID-19 spread in Oslo, Norway hasn’t gotten more play. Here’s a synopsis (boldface mine):
Like many countries, Norway ordered all gyms to close in March to prevent the spread of the coronavirus. But unlike any other nation, Norway also funded a rigorous study to determine whether the closings were really necessary.
It is apparently the first and only randomized trial to test whether people who work out at gyms with modest restrictions are at greater risk of infection from the coronavirus than those who do not. The tentative answer after two weeks: no.
So this week, responding to the study it funded, Norway reopened all of its gyms, with the same safeguards in place that were used in the study.
Is there hope for gymgoers in other parts of the world?
…The trial, begun on May 22, included five gyms in Oslo with 3,764 members, ages 18 to 64, who did not have underlying medical conditions. Half of the members — 1,896 people — were invited to go back to their gyms and work out.
They were required to wash their hands and to maintain social distancing: three feet apart for floor exercises, and six feet apart in high-intensity classes. The subjects could use the lockers, but not the saunas or the showers. They were not asked to wear masks.
Another 1,868 gym members served as a comparison group; they were not permitted to return to their gyms.
During the two weeks of the study, 79.5 percent of the members invited to use their gyms went at least once, while 38.4 percent went more than six times…
The results? The researchers found only one coronavirus case, in a person who had not used the gym before he was tested; it was traced to his workplace. Some participants visited hospitals, but for diseases other than Covid-19, the illness caused by the coronavirus.
However, there’s a big caveat:
Some experts felt the results demonstrated that returning to the gym was relatively safe — but only in places where there were few infections.
“This shows us that low-prevalence environments are safe for gyms and probably just about everything else,” said Dr. Gordon Guyatt, a professor of medicine at McMaster University in Canada. “It is very unlikely you will get infected.”
“If you were in a different environment where there is a substantially higher prevalence, we don’t know what will happen,” he added.
And “substantially higher prevalence” doesn’t have to mean eleventy gajillion percent: it can be seemingly small shifts that make activities risky.
Let’s compare D.C. and Oslo. Recently, I calculated that 0.1% of D.C. potentially was able to infect someone with COVID-19 (though the rate among wards differed seven-fold, and the rate has declined somewhat since then). Between the NYT article and the manuscript*, I find three different numbers of the total number of Oslo-ians (sic? Oslots? Osloids? I’ll stop…) who were infected over the two weeks of the study (which is the same proxy I use), but it seems like 0.02% is a fair estimate. Keep in mind, my estimated percent of currently uninfected people in D.C. and Oslo respectively is 0.999% and 0.9998%–which, at first glance are pretty similar.
Suppose, when you go to the gym, you encounter twenty people. If the gym members are representative of the population as a whole (I’ll get to caveats later), then the odds that one or more fellow attendees have COVID-19 are 0.4% and 2.0% for Oslo and D.C. respectively per visit. Let’s say, over that two week period you hit the gym eight times, then the chance that you will have had one or more gym visits with one or more infected people is 3.1% and 14.8% for Oslo and D.C. respectively.
One criticism raised in the NYT article is that the study didn’t run long enough to see any harmful effects. If we expand our toy calculations out to three months (six two week periods), then the chance that you will have had one or more gym visits with one or more infected people is 17.5% and 61.7% for Oslo and D.C. respectively.
Now, to flip the old statistical saw around, my toy example is precise, not (necessarily) accurate. Obviously, there’s a lot more that we would have to consider, everything from the age of attendees (young people who bar hop four nights a week versus older gymgoers who stay at home more often), the type of exercise (i.e., sustained deep inhalation and exhalation, like for bike sprints), to the facilities (my gym’s weightroom is a large, high ceiling room with lots of external ventilation, while the bike studio is a small enclosed space). And, of course, being in the gym with an infected person doesn’t necessarily mean you’ll become infected.
But what these back of the envelope calculations do show is that crushing the curve makes it much less likely to be around someone who is infected with COVID-19. Relatively small differences matter, and, at the state level, the only U.S. state that has accomplished said crushing appears to be Hawaii.
If people want things to ‘return to normal’, or some semblance of normal, then we need to dramatically lower the prevalence of infected people. When COVID-19 is very rare, one is far more likely to survive higher risk activities. That means masks and physical distancing for an extended period of time–the first, sort of shutdown didn’t cut it, and we left far too early.
Once we get it to Oslo levels (and we could, if we had a competent and coordinated response), life will be far more like normal.
*One problem with some pre-prints is that they can be unclear–they really need editing and reviewers who don’t understand the writing because it’s unclear.