Data from your FitBit could help predict COVID

Research suggests wearable devices could help in virus fight

Dr. Michael Snyder is leading Stanford's research into the capacity of wearable devices as an early detection system for infections. Photo: Steve Fisch, Stanford Medicine

Data from wearable health-tracking devices can detect signs of COVID-19 several days before patients notice symptoms, according to preliminary research from Stanford University.

FitBits, Apple watches and newer devices like the BioButton could hold the key to early detection of cases, Michael Snyder, director of the Center for Genomic and Personalized Medicine at California’s Stanford University told the Cayman Compass.

Data from the smart watches of COVID-positive patients, analysed by Snyder and his team, showed clear spikes in heart rate an average of four days before the onset of symptoms and seven days before a positive COVID diagnosis.

Snyder said there was growing evidence that elevated heart rate in particular could provide a key signal for viral illness.

“If it is up for 12 hours, you can give a reasonable alarm that something is wrong,” he said. “It doesn’t necessarily mean it is COVID but it is an early alarm system that something is up.”

- Advertisement -

Snyder, who wears at least five health-tracking devices, including a ring which monitors his blood oxygen levels, has been studying the links between the data they produce and illness for several years.

Analysing his own data, he said, he had been ill four times in the space of two years and on each occasion his devices showed spikes in heart rate and skin temperature prior to him getting sick.

Initially, he was able to generate little interest in his research, he said.

“A lot of people said, ‘That is really neat, but who cares? You get sick, you go to bed and drink chicken noodle soup,’” he said.

Now, he said, knowing when you are about to get sick could be the difference-maker in preventing new outbreaks of COVID-19.

Snyder

Snyder believes wearable devices that provide key health data to the user will become more advanced and far more common over time.

“Everybody is going to be wearing these things in the future,” he said.

“You don’t drive your car around without a dashboard that tells you when you are running out of gas, so why are people running around without a dashboard that gives you early clues to your health state?”

Stanford has developed an algorithm for an ‘alarm system’ that can be synced with any device to alert users to worrying changes in their health metrics.

Differentiating COVID-19

More research is needed, Snyder said, to differentiate COVID from other types of illness, though he believes this is also possible.

He said devices would get more sophisticated and become more ubiquitous post-COVID. And he believes governments, workplaces and any centres where large numbers of people gather could use this kind of data as a tool to evaluate the risk status of citizens, employees or customers.

“You could say ‘Show me your data before you come into work or school today’. That is very possible,” he added.

There is potential for false positives – with similar spikes in data likely for flu or stress-related conditions – but the alarm system would provide a fairly definitive guide that something is wrong and needs to be checked out, said Snyder.

Research ongoing

There are at least five separate studies under way in the US focussed on how smart watches and other health-tracking devices can be used in the fight against COVID-19, said Jennifer Radin, an epidemiologist with California-based Scripps Research.

Jennifer Radin.

Scripps is currently researching data uploaded by 36,000 people from FitBits, Apple watches and other devices.

“Our goal is to look and see if there are detectable changes in users’ wearable data that are associated with a viral illness, such as COVID,” she said.

Early results suggest that this data can help identify people who are not yet symptomatic.

Radin said the capacity to use baseline data, that is specific to an individual, is likely to be useful in early identification of a viral infection.

The key advantage of devices like smart watches and BioButtons, she says, is that they give solid data about each individual, rather than operating from a broad definition of what is ‘normal’.

The average heart rate for any individual, for example, can be anywhere between 50 and 100 beats per minute.

Knowing where an individual typically fits on that scale helps detect spikes that would otherwise go unnoticed, she said.

Radin said there was some evidence that these spikes would also occur in asymptomatic carriers of COVID-19.

The study is in the early stages and the impacts of conditions like asthma and diabetes are still being contemplated.

But she sees room for hope that devices that many people are wearing anyway could help public health authorities manage COVID-19.

“I think it will be a tool that can supplement other traditional methods of screening,” she said.