How wearing a Fitbit on your wrist could detect COVID-19
Your smartwatch could detect you've contracted COVID-19 a day or even a week before symptoms begin using new technology developed by Fitbit that its co-founder said could "help stop the spread of the virus".
The wearable technology firm has unveiled some results from its study of 100,000 Fitbit users, including 900 who contracted coronavirus, revealing new software using its health data could detect half of all cases with a high degree of confidence.
The algorithm is currently undergoing medical review with plans to release it "as quickly as possible," and Australian medical researchers said the development could be "incredibly useful" in identifying when people should self-isolate.
Fitbit co-founder Eric Friedman revealed details of the company's latest research at the launch of new wearable devices overnight, including a Sense smartwatch that will add sensors to measure users' skin temperature, stress levels, and oxygen intake.
Mr Friedman said Fitbit had been using its health data with Scripps Research to detect influenza outbreaks prior to the pandemic, but had now "accelerated" work with organisations including Stanford University to detect coronavirus infections.
"Today we're working to help develop a means to screen for COVID-19 before symptoms start," he said.
"Our goal is to be able to alert people that they may be sick before they notice symptoms so they can isolate and get tested, helping to stop the spread of the virus."
Stanford University genetics department chair Dr Mike Snyder said researchers had so far looked at whether coronavirus infections could be detected by finding "abnormalities" in users' resting heart rates "so they're not spreading the illness all around".
And Mr Friedman said based on this research, Fitbit launched a study of 100,000 users in the US and Canada to "build an algorithm that detects potential signs of COVID-19".
Users' resting heart rates, heart-rate variability, and breathing rates were assessed - including the details of 900 participants diagnosed with coronavirus during the study - and used to create software that could pinpoint when users were falling ill before symptoms appeared.
Mr Friedman said the technology could be used to "detect nearly 50 per cent of COVID-19 cases a day before the onset of symptoms with 75 per cent specificity," with signals apparent days before that, and was now undergoing a medical peer review.
"The takeaway from these (findings) is that it's possible to know nearly a week before symptoms that you could be falling ill," he said.
"Advance warnings means you can isolate and help stop the spread of this virus."
Fitbit product vice-president Larry Yang told News Corp it was too early to say what Fitbit devices could use the software or what a coronavirus-detecting app could look like, but the research was being fast-tracked.
"The finding that there are correlations between the biometrics we collect and early indicators is super encouraging so the next step is (working out) are there stronger signals and what sort of user experience can we build around it?" Mr Yang said.
"We're hoping to bring it to our consumers as soon as possible."
Griffith University infectious diseases and immunology program director Nigel McMillan said the technology would not replace clinical COVID-19 tests but, if it proved to be accurate, could help potentially infected users to isolate themselves.
Professor McMillan said it could be particularly useful in asymptomatic or pre-symptomatic coronavirus cases, where people did not realise they needed to stay away from others.
"Say you get the alert that you might have an infection and you were going to go out to a party - you might decide not to," he said. "Or you might wait a day and see if it becomes more serious. That might stop you giving it to your friends or family members."
A study of Fitbit data published in The Lancet in January this year found assessing that information could help public health organisations identify influenza outbreaks and "prevent further transmission" of the virus.
Originally published as How wearing a Fitbit on your wrist could detect COVID-19