Apple / WIRED
In the middle of March, Luca Foschini began to feel unwell. He had a high fever, cough, a light head and a lot of bone fatigue. His wife had just returned from London which was experiencing a spike in Covid-19 cases at the time, and the couple feared that they both might have contracted the disease too.
Foschini rang his local doctor in Santa Barbara, California, who told him there were no tests available and that the pair should stay home, monitor their symptoms and call him back if things got worse. “I was very frustrated at first, then anxiety kicked in,” says Foschini.
The only things that made him feel better, though, were the readings he was getting from his Apple Watch. They showed that even before he felt ill his resting heart rate jumped to over 60 beats per minute, up from its average of 50. Although another person might have been alarmed, Foschini was excited. It was another indicator towards the hypothesis that his company — alongside roughly 15 other private and public research institutions — has spent the past couple of months trying to prove: that biodata from wearable devices can provide an early warning system for Covid-19.
Wearables are everywhere, says Foschini, co-founder and chief data scientist at the California-based tech startup Evidation Health. It’s estimated that one in every five people living in the US and one in ten in the UK wears a fitness tracker or smartwatch. Depending on the device, users can monitor their sleep and activity patterns, heart rate, blood oxygen saturation levels and other biomarkers.
Every individual has a different baseline level of what’s normal for them. The idea is that the wearable devices could monitor deviations from that normal rate and raise a flag that someone might be coming down with Covid-19. “Wearables could give you an early warning to isolate yourself and get a confirmatory test,” says Foschini.
“This is a strategy that is very inexpensive — it’s just about data and analytics,” says Eric Topol, a cardiologist at the Californian biomedical non-profit Scripps Research, which is conducting one of the largest studies of wearables and Covid-19 involving 40,000 participants. One problem with current swab tests for Covid-19 is that they have to be administered at the right time for the virus to be detected. Plus the tests are relatively expensive, says Topol, and “you can’t do it every day or frequently enough in 330 million Americans.”
“But digital assessment is a continuous process, so it’s very complementary,” he says. In addition, wearables make for an attractive screening tool because they’re non-invasive and can be done from the comfort of one’s own home. Even more than that, fitness trackers and smartwatches yield a potential trump card in the fight against Covid-19 — they can possibly detect pre-symptomatic individuals, who can be contagious up to three days before they start to feel sick.
“They are the real spreaders, based on modelling and what we know about when people were infected by members of their families,” says Amos Folarin at King’s College London, who is leading the Covid-Collab study. “Presymptomatic people are never going to get caught by test and trace,” says Folarin. “So having something that can continuously read your vitals and identify when you have a respiratory infection really gets into a window where you just cannot reach by any other means.”
Some researchers believe wearables could even help detect asymptomatic spreaders — those with the virus yet who never feel ill themselves. An estimated 40 per cent of all infections fall into this category. “But when these people go through a chest scan, half or more have abnormalities,” he says. “So we know that the virus takes a toll on the body even if you don’t have any symptoms.”
Researchers aren’t exactly sure what the best signals for a Covid-19 infection are, but so far the most useful metric appears to be resting heart rate, which Topol says “goes up a day or even a couple of days” before symptoms appear. Your heart works harder when you’re fighting off an infection, and this in turns translates to an elevated heart rate — which is why nearly all studies involving wearables and Covid-19 factor it into their algorithms.
But to boost prediction accuracy from wearables, we need to rely on more than just one signal. “It’s being able to marry [heart rate] together in an intelligent way with some of the other measures,” says Folarin.
Most commonly, studies also observe users’ activity levels and sleep patterns as a proxy for possible infection. Some studies — such as University of California, San Francisco’s TemPredict and the German Robert Koch Institute’s Corona-Datenspende — also measure skin temperature. The metric, however, is less popular partly due to the fact that most wearables aren’t able to measure it (the Oura ring is one of the few that can) and that fever isn’t as prevalent in infected patients as once believed (roughly 30 per cent are febrile).
So far, preliminary results have been encouraging. In a study involving more than 600 frontline workers wearing Oura Rings, researchers at West Virginia University in the US reported being able to forecast Covid-19 three days before symptoms appeared, with an accuracy of more than 90 per cent. In August, Fitbit announced that in their study of over 1,000 coronavirus patients, they were able to detect more than half of the positive cases a day before symptom onset with 70 per cent specificity. Similarly, both Evidation and Scripps have published results suggesting that wearable data can forewarn of a Covid-19 infection.
But experts caution it’s early days yet. None of the publications have been peer-reviewed and the number of patients involved needs to be significantly higher — in the scale of hundreds of thousands, even millions — to improve the accuracy of the algorithms and push down the rates of false positives.
Then there are other issues to contend with, such as the uneven distribution of wearable devices in the general population. “They’re not something that everybody has,” says Albert Titus, a biomedical engineer at the University of Buffalo in New York. “Is it going to be that only the people who can afford them are going to be the ones who are tracked?”
Part of the appeal of using wearables is that public health officials could be alerted to a cluster of signals and be able to home in early on hotspots. But ownership of wearables tend to be skewed in any given population. Excluding lower-income populations who can’t afford wearables will “just exacerbate the existing inequalities that we’re seeing with Covid,” says Brent Mittelstadt, a data ethicist at the University of Oxford.
Then there’s the issue of privacy. “We are increasingly comfortable in letting our everyday lives be surveilled,” says James Gilmore, a communications professor at Clemson University in South Carolina. “We have to ask: are those three days [of early detection] worth the tradeoff of wearing technology day in and day out, allowing it to extract and automatically analyse our biometric data and have it stored on a company’s servers for an unspecified amount of time?”
But of all the barriers that remain, the really big question is this: does Covid-19 have “a sufficiently unique and reliable fingerprint” to be able to distinguish it from other coronaviruses and respiratory infections, asks Mittelstadt.
For now, the answer remains unclear. “Certainly wearables are offering us insight into trends that evaded us before,” says Colin Drummond, a biomedical engineer at Case Western Reserve University in Ohio. “But we haven’t quite parsed out the mosaic and the signs are incomplete.”
Foschini’s Apple Watch, for example, may have detected an unusual spike in his resting heart rate. But, as he later discovered, that was because of a respiratory infection and not Covid-19.
Topol says that pinpointing the exact cause of a red flag will be difficult. “We have a very good validated signal and have made some progress…but I think once we get to the area of other infections, that’s where our specificity is going to be weak.” The real test, he says, will be when flu season rolls around this winter.
Still, Topol is optimistic for a future with wearables on the pandemic’s front line. “My hope is that we will become more embracing because it’s low cost,” he says. “Just like how the pandemic ushered in telemedicine, hopefully we’ll see something similar here with digital surveillance of infections like this.”
Eric Topol will be one of the speakers at WIRED Health:Tech on September 22 – an event exploring the health trends accelerated by Covid-19. Speakers include Heidi Larson, director and founder of The Vaccine Confidence Project, and CRISPR co-inventor Jennifer Doudna. Tickets start at only £40 + VAT. Book your tickets here.
More great stories from WIRED
🐾 A liver disease is putting the Skye Terrier’s existence at risk. Doggy DNA banks could help save it
🔞 As AI technology gets cheaper and easier to use, deepfake porn is going mainstream
🏡 Back at work? So are burglars. Here’s the tech you need to keep your home safe
🔊 Listen to The WIRED Podcast, the week in science, technology and culture, delivered every Friday
👉 Follow WIRED on Twitter, Instagram, Facebook and LinkedIn
Get The Email from WIRED, your no-nonsense briefing on all the biggest stories in technology, business and science. In your inbox every weekday at 12pm sharp.
Thank You. You have successfully subscribed to our newsletter. You will hear from us shortly.
Sorry, you have entered an invalid email. Please refresh and try again.