The innovative wearable device that helped detect COVID-19 on the PGA Tour
As economies are trying to reopen, there is a strong need for technologies that can help detect possible COVID-positive individuals before they can spread the virus to others. Testing availability is still limited in some locations, and rapid testing is quite rare. One of the most promising ideas is coming from a surprising place. Golf.
One PGA Tour golfer, Nick Watney, tested positive for COVID-19 last week during a tournament. And he credits his WHOOP wearable device with giving him an early indication of symptoms.
What is the WHOOP? It's a wearable device, similar to a Fitbit, but a bit larger. The device is targeted towards the serious athlete, because it lacks some of the mass-market fitness-tracker features like a step counter. And they count some well-known pro athletes as users of their device.
The WHOOP uses a PPG sensor to detect and track your heart rate. There is an associated app on your smartphone that uses this data to analyze sleep quality, recovery, and overall strain on your body.
Sounds interesting, but how could any of that tell you that you may have COVID-19?
The key is your respiratory rate, the number of breaths you take each minute. WHOOP has an interesting writeup on how this is done here, but in simple terms, it relies on the fact that your heart rate increases slightly when you inhale, and decreases slightly when you exhale. By measuring the number of times your heart rate goes through these small cycles each minute, it can compute respiratory rate.
Watney remarked that he noticed a "spike" in his respiratory rate one morning, and this clued him into the possibility of trouble. A COVID-19 test was performed, and came back positive.
So far, the link between respiratory rate and COVID-19 is anecdotal based on Watney and one WHOOP team member. But further research is underway by the Cleveland Clinic and others to study this relationship better.
A quick look at WHOOP's recent patent filings shows they have much more planned in the future. Since the device collects data from many users, there are lots of opportunities for machine learning to enhance the device features. US Patent 10,548,513 ("Activity Recognition") was just issued this year, and includes claims to machine learning algorithms to use accelerometer data to detect types of activity, so the device could know whether you went on a run or went for a bike ride. This patent application (US2019/0110755 - "Applied data quality metrics for physiological measurements") is still pending, but seeks to input the physiological data of multiple users to a machine learning model to improve the quality of the resulting measurements. Finally, there is another pending application (US2019/0099116 -- "Tissue oxygen saturation detection and related apparatus and methods") that attempts to move beyond the traditional pulse oximetry measurement and directly measure the oxygenation of the muscle. This could be a big help to athletes looking to monitor performance.
This is a promising technology for our current pandemic, and one that may yield even more features in the future.