Could Your Fitness Tracker Detect COVID-19?

Most of these smart devices use a mass of tiny sensors to get a picture of how you're doing and where you are. Optical sensors against your skin estimate heart rate and blood oxygen levels; accelerometers and gyroscopes track movement; and galvanic skin response, a change in sweat gland activity, assesses your levels of exertion.

Other sensors measure ambient light and temperature, your location through GPS and barometric pressure, delivering a wealth of information on your health by establishing a baseline.

"Some of these measurements, like heart rate, are very accurate,” says Michael Snyder, M.D., chairman of the genetics department at Stanford University School of Medicine, about 30 miles southeast of San Francisco. “Others like SpO2 [blood oxygen levels] are not, but it is the change from baseline that counts.”

The programs first establish normal readings for a week or two (the so-called baseline), then look for any abnormalities that might indicate illness in future readings. Radin and Snyder say they are focused mostly on heart rate changes and sleep disruptions, but other information, including breaths per minute and pulse oxygenation levels, could help identify a specific illness, such as COVID-19, along with the assistance of data from thousands of patients and the application of artificial intelligence programs.

The more people who participate in such programs, the more accurate the health predictions may be about a specific illness, they say. That also will make more public health data available that officials can use during new outbreaks.

Smart ring might be less intrusive

One of the most ambitious programs aimed at warning the public and health professionals about coronavirus outbreaks is run by the West Virginia University Rockefeller Neuroscience Institute and WVU Medicine in Morgantown, in conjunction with Oura Health. Finland-based Oura makes a smart ring that measures not only heart rate but a person's temperature — one of the few such smart devices to do so.

"The Oura ring has a good form factor, and people can wear it when they are sleeping,” instead of a watch, which can be uncomfortable to wear in bed, explains Ali Rezai, M.D., a neuroscientist and executive chairman of the WVU Rockefeller Neuroscience Institute. By observing resting heart rates, body temperatures and deep-sleep patterns, the group is looking to forecast the illness with a degree of accuracy and predictability previously not possible.

Oura

Using a volunteer group of more than 800 first responders and medical professionals working in hospitals around the country, the researchers are combining the smart ring information taken passively with a smartphone app that asks wearers questions about their health plus possible coronavirus exposure and compares reaction times using a cognitive game to judge fatigue.

Regular nasal tests for the coronavirus and blood tests for antibodies to COVID-19 are also included and combined online in the cloud, to predict how individuals will feel three days later. Rezai says they are seeing about 95 percent accuracy in predicting what a person's temperature will be three days in the future.

For the public, though, eliminating those nasal and blood tests to accurately pinpoint a COVID-19 infection versus the flu will take a lot more data and volunteers. The Rockefeller Neuroscience Institute already has more than 29,000 people using another smartphone app, but it will take many readings to teach a computer what the earliest COVID-19 symptoms look like.

Stanford's Snyder says they've scaled their program to accommodate millions of users. Whether that many people will be comfortable with such health-monitoring devices and software, which U.S. medical privacy laws cover, has yet to be determined. Fitbit is already testing the waters with a Ready for Work questionnaire and a monitoring program for employers to help workers determine if they are healthy and able to return to their jobs.

The technology will evolve

Some sensor readings, such as a person's blood oxygen levels, are trailing indicators, rather than early warnings, Rezai says. But the sophistication and accuracy of sensors will improve.

That could make it more attractive to many looking for some early sign of an impending health problem. For example, he envisions automatic messages that inform the wearer that she is going to be ill in three days when she's scheduled to fly, indicating it might be better to cancel the trip and seek medical attention. But none of the programs being tested right now can do that or offer alerts to the wearer.

"We don't want to alarm people. It's still a research study,” Radin says about DETECT.

What's more, none of the more than a half dozen researchers contacted would offer a date on when such a feature might be available. But if these research programs are successful, similar apps for smartwatches and fitness devices could be on the market as soon as a few months from now.