Apple Watch, Fitbit data can spot if you are sick days before symptoms show up

A Stanford study is showing how wearables have promise in helping to keep a lid on coronavirus infections.
Written by Jo Best, Contributor

Stanford University is hoping to turn wearables like the Apple Watch and Fitbit's health-tracking bands into the latest weapon in the fight against coronavirus.

The Stanford Medicine's Healthcare Innovation Lab launched the Coronavirus Wearables Study earlier this year to research whether wearables can be used to detect if someone has COVID-19 before they start showing any symptoms (if they ever do).

The study, which is currently recruiting for participants, will ask users to give data from their wearables -- such as heart rate, skin temperature, and blood oxygen saturation -- through an app created by Stanford's bioinformatics team. (The system currently works with Apple Watch, Garmin, and Fitbit devices, among others.) Would-be participants will also need to fill in a symptom checklist regularly and can optionally also share details from their medical records. 

SEE: IoT security: A guide for IT leaders (TechRepublic Premium)    

Anyone looking to take part will either need to have been diagnosed with COVID-19 or be at high risk of contracting it -- for example, grocery store or healthcare workers, or anyone who's been on a cruise recently. Participants will also need to be over 18. Anyone thinking of taking part should also be prepared for the long haul: the study is set to last two years.

The coronavirus study isn't the first time Stanford has looked to investigate the potential of wearable devices for improving health. In 2017, the university launched a study with Apple to see if the Watch could identify atrial fibrillation, a heart condition that can lead to strokes and heart attacks.

The current study, which is a collaboration between Stanford Medicine, Scripps Research, and Fitbit, will use data gathered from the wearables to create algorithms that can detect the physiological changes in someone that show they're coming down with an infection, potentially before they even know they're sick. Once the signs of infection -- such as an increase in resting heart rate -- have been detected, the user will be alerted through the app that they may be getting sick, allowing them to self-isolate earlier and so spread the infection to fewer people.

The lab has been investigating the potential of wearable devices to shed light on changes in users' health for some years. Researchers published a study in 2017 that showed devices could pick up changes in physical parameters before the wearer noticed any symptoms. 

The algorithm from that research, known as 'change of heart', detected that changes in heart rate could signal an early infection, and the lab is now building on that research for the current pandemic. "We continued to improve the algorithm, then when the COVID-19 outbreak came, as you might imagine, we started scaling at full force," Michael Snyder, professor and chair of genetics at the Stanford School of Medicine, told ZDNet.

The first phase of the study is designed to prove that wearables can pick up a coronavirus infection by detecting changes in heart rate and other physiological parameters. The results are promising, according to Snyder: the first case the lab investigated detected physiological changes indicative of an infection through their wearable over nine days before symptoms showed up.

"You can't miss the signal. It's very, very clear. That person is running around for nine and a half days ill, asymptomatic or infectious and not knowing it, therefore presumably infecting many others," he said. 

On average, the system is detecting the signs of coronavirus through wearables four days before the symptoms appear in most cases. 

It's not an infallible system, however: not everyone with the virus will be detected. Typically those with infections that aren't getting picked up tend to be people with respiratory conditions, such as severe asthma, which might throw their baseline parameters off. What's more, it's not yet known how specific the system is -- whether it will be able to tell the difference between coronavirus, pneumonia, or a cold, for example (though Snyder does note that it can tell bacterial infections from viral, as bacterial infections tend to give off a stronger signal). 

"But the main thing, certainly when you're in a pandemic, is it'll tell you if you're getting ill. And we think that information is super-useful, because [health authorities] are not going to go around PCR testing everyone every day -- you can't do it and it doesn't make sense. But you can have everybody wearing a smartwatch." 

In order to be able to work out some of the variables that might affect any changes, the Stanford team is asking those who want to to contribute their medical records to enable the algorithm to be adjusted to take into account certain circumstances -- those with diabetes for example. 

The second phase of the project is about to roll out, where the system will start alerting those who have a suspected infection. Anyone affected will receive a text alert to their smartphone notifying them that there have been changes in their normal physiology.

However, study participants will need to view any alert in relation to what's happening in their life and act accordingly. "If you have a reason for [the parameters] being elevated -- maybe you're on a long hike in a desert or you're watching scary movie marathon for 12 hours or something -- you can ignore our alert. But if you're just sitting around on the couch and we tell you your heart rate and things are elevated, you might want to get yourself tested… You'll need to contextualise the information."

The research is also looking into whether the data from wearables might be able to tell more than whether or not an infection is present -- including whether it could even indicate if certain people may have a more severe response to the infection.

The first official results from the study are expected to be published in the coming weeks, Snyder said. 

SEE: Apple Watch Series 5 review: This is the watch I've been waiting for

Once the system is production-ready, the team hopes to release it under an open-source licence: "I would hope that every device manufacturer, basically, uses the apps that we'll have set up for being able to do infectious disease monitoring, and that's just a standard part of the product."

There are other potential uses of such technology too that might warrant investigation. Take people who might have difficulty articulating when they're ill -- very young children, or people with dementia, for example -- health tracking using wearables could give valuable insight to caregivers into whether their charges are well or otherwise. Similarly, the tech could also be teamed up with track-and-trace apps to help identify people who might have come into contact with asymptomatic virus carriers. 

Tech companies have also been working to make their products more useful for tracking suspected COVID-19 infections. Apple and Google announced a joint COVID-19 effort aimed at helping public authorities and health organisations create apps to track and trace people that might have been exposed to the virus. The Exposure Notification framework will enable iPhone and Android users' phones to record the people they've come into contact with and alert them if those people are subsequently diagnosed with coronavirus.

Apple has also put the Watch to work in battling COVID-19. At its WWDC conference, Apple announced that the next Watch OS would detect when users were washing their hands and set a timer for 20 seconds -- the recommended time to prevent the spread of infection -- and vibrate to tell them when their handwashing time is up.

Editorial standards