Using data from a wearable fertility tracking device, researchers developed an algorithm that successfully identified 68% of COVID-19 cases before symptom onset.
These findings were based on the Ava bracelet, an FDA-approved fertility tracker that also monitors heart and breathing rates, skin temperature, blood flow, and sleep quality and amount.
Researchers believe this may be applicable to other similar activity tracking devices and smartwatches, and sought to find out whether physiological changes monitored by an activity tracker could be used in the development of a machine learning algorithm detecting presymptomatic COVID-19 infection.
To do so, they included 1163 participants between March 2020 and April 2021 from the ongoing GAPP study. Participants were aged younger than 51 with a mean (SD) age of 44 (5.5) years, and 57% were female. All participants wore the Ava bracelet at night during the study period.
The bracelet—which is 90% accurate in detecting women’s most fertile days in real time—saves data every 10 seconds and is synced to a corresponding app on the user’s smartphone. In the app, participants reported any activity that could possibly affect their central nervous system, including consumption of alcohol, medications, or recreational drugs. They also reported symptoms potentially related to COVID-19.
Of the group, 127 (11%) participants contracted COVID-19 during the study period, confirmed through a PCR test. Of this group, 66 (52%) reported wearing the device for at least 29 days before symptom onset.
Data collected through the device and app showed there were significant changes in all 5 physiological indicators—respiratory rate (RR), heart rate (HR), HR variability (HRV), wrist-skin temperature (WST), and skin perfusion—during the incubation, presymptomatic, symptomatic, and recovery periods compared with baseline. The researchers found that symptoms lasted 8.5 days on average.
Using data from between 2 and 10 days before symptom onset, the researchers developed an algorithm to detect COVID-19. It successfully identified 68% of COVID-19 cases up to 2 days before participants developed symptoms.
According to the authors, this machine-learning algorithm could be applied to any sensor device measuring the same 5 physiological indicators.
“Wearable sensor technology is an easy-to-use, low-cost method for enabling individuals to track their health and well-being during a pandemic,” the study authors said. “Our research shows how these devices, partnered with artificial intelligence, can push the boundaries of personalised medicine and detect illnesses prior to [symptom onset], potentially reducing virus transmission in communities.”
They also noted that similar research going forward should focus on how medical-grade wearable sensor technologies can be used to monitor sensor data during the ongoing COVID-19 pandemic.
“We acknowledge that our sensitivity was less than 80%,” the study authors wrote. “We expect to improve the algorithm‘s performance further in a larger cohort within the setting of the COVID-RED study (n=20,000).”