Could Your Fitness Tracker Really Detect COVID-19?

Originally published in GQ on July 14, 2020

When professional golfer Nick Watney woke up on Friday, June 19, after playing the first round of a PGA tournament in South Carolina, he felt physically fine. But when he checked the WHOOP fitness tracker he’s worn on his wrist for the past year, he was startled to see a spike in his breathing rate while he was sleeping.

He had heard that could be a sign of COVID-19, so just to be safe, despite showing no other symptoms, he got a test. To Watney’s surprise, it came back positive. Over the next ten days, while he self-isolated in South Carolina, he never developed a fever, cough, or shortness of breath—though he did end up losing his sense of smell for a while.

A week after Watney tested positive, the PGA Tour announced it would be distributing WHOOP bands to all players and caddies, in the hopes that they too might be able to identify potential coronavirus infections early. The straps have since been credited with early detection for other PGA Tour golfers who have gone on to contract the virus. This comes on the heels of the NBA’s announcement earlier in June that it would be purchasing more than 2,000 Oura Rings, a similar fitness tracker, to help detect cases of the virus when the league re-starts in Orlando.

Whether it’s WHOOP and Oura Ring, or other fitness trackers like Fitbit and the Apple Watch, there’s increasing enthusiasm around fitness trackers among researchers as a first line of defense in the fight against coronavirus. While the science is still in its earliest days, the hope is these devices could alert individuals to changes in their health they might not otherwise notice. Early detection is particularly crucial for combating the coronavirus, which has the unusual characteristic of spreading “silently” from people who are not feeling symptoms. These devices use sensors to track a range of physiological markers—sleep patterns, heart rate, breathing rate, and temperature. For example, if your resting heart rate is normally 62 bpm, but jumps inexplicably to 75 bpm, you likely wouldn’t feel any different, but that jump could prompt your tracker to issue an alert to get tested. Of course, no researcher expects the tech to replace real diagnostic testing. At least right now, the best anyone can say about this right now is that, when combined with other coronavirus safety protocols, it doesn’t hurt.

Researchers also see potential to use wearable tech to study how the virus moves through large populations. If a cluster of people living in the same area all start to notice similar changes in their heart rate or temperature, that could help officials better respond to outbreaks and mitigate their spread. “If you and many of your neighbors are showing similar reactions—that’s when it becomes a signal rather than noise,” said Eric Topol, the director of the Scripps Research Translational Institute, which is leading one of the major ongoing fitness tracking studies. Topol’s interest in using fitness trackers to predict disease predates the current pandemic. In January he and his colleagues published a study that found by using de-identified heart rate data from Fitbit users, researchers were able to significantly improve their predictions of influenza-like illness when compared to using CDC data.

A built in advantage for fitness trackers is that millions of people already have them. In 2019 Gallup reported nearly 20 percent of Americans currently use one, and consumer analysts say the devices represent one of the fastest-growing sectors in global technology, especially smartwatches. Another benefit is that they work passively; they don’t require manually entering symptoms into an app, or sticking a thermometer under your armpit on a daily basis.

While researchers are optimistic about the potential, they caution that the existing knowledge base is very, very scant. There’s been almost no research, for example, on how wearables could be turned into reliable clinical tools. “Your average doctor doesn’t want to see your wearable data because they don’t know what to do with it or how to make sense of it any more than anyone else,” said Benjamin Smarr, a data science and bioengineering professor who is leading a University of California at San Francisco study on Oura ring data and COVID-19.

Smarr is among a handful of researchers who have been studying wearable tech data for the past decade. He said the overall assumption is that having lots of information health professionals wouldn’t otherwise have access to will prove to be useful in some way, even if they don’t know exactly how yet.

And while Harvard health policy professor Thomas Tsai supports the ongoing research—“it’s fascinating and important,” he says—from a public perspective, Tsai worries about the mixed messages these fitness trackers could be sending during the pandemic. “We’re working really, really hard to break the message that only symptomatic individuals should get tested,” he said. “That was true back in March and April, when we had a huge shortage of tests, but that’s not true anymore.”

Tsai also raised concerns that someone’s smartwatch could give them a false sense of security about their health. “The danger is that someone who needs to get tested may not want to because their wearable device says they are showing no symptoms,” he says.

Right now there’s at least four large-scale fitness tracker studies underway; the one at UCSF, the one at Scripps, a third at West Virginia University, and a fourth at Stanford. In late May the West Virginia researchers announced preliminary results from a study of 600 healthcare professionals and first-responders, which found that, using Oura Ring data and artificial intelligence models, researchers could predict COVID-19 symptoms three days in advance of their onset, with over 90 percent accuracy. That’s not a huge sample size, but they’re now scaling up their next phase of research to roughly 10,000 people. (Oura is not funding the West Virginia study, but is funding the trial at UCSF.)

Topol says he hopes that when (and if) we have high-quality coronavirus tests that people can administer from home, the fitness trackers could pair well with them, allowing infected people to self-isolate earlier than otherwise would have. “There are sixteen companies working on at-home COVID tests and we’re in discussions with five of them,” he said.

“These fitness tracker studies are going to lead to a really, really large amount of comparative physiological data,” Smarr added. “We don’t know yet what we’ll find— nobody has ever done something like this before.”