Fitbit research suggests wearable biometrics tracking can detect COVID-19
A new research paper published by Fitbit suggests the wearable could be used to detect COVID-19 by monitoring the wearer’s biometrics earlier than is possible from self-reported symptoms, and resources are also in place to improve the system’s accuracy.
The paper on ‘Assessment of physiological signs associated with COVID-19 measured using wearable devices’ is published by npj Digital Medicine, and discusses the early signs of illness that can be detected from measurements of respiration rate, heart rate, and heart rate variability (HRV). The researchers found that between 11 and 12 percent of 2745 subjects who tested positive for COVID-19 with PCR tests were asymptomatic. A total of 21 symptoms were reported, and just over half reported fevers.
Based on self-reported symptoms, the researchers scored an AUC of 0.82 plus or minus 0.017 for predicted hospitalization, and illness on a specific day based on physiological data was predicted with an accuracy of 0.77 plus or minus 0.018 AUC.
Generally, they found that respiration and heart rate are elevated by the illness while HRV is decreased. A neural network was trained to use these metrics as predictive factors.
“Measuring these metrics, taken in conjunction with molecular-based diagnostics, may lead to better early detection and monitoring of COVID-19,” the researchers conclude.
The U.S. Army announced a $2.5 million research grant for Fitbit to work on early COVID-detection just over a month ago.
Lux Research Analyst Danielle Bradnan tells Biometric Update in an email that while the company has been measuring employee biomarkers to detect an immune response since June, it has avoided making specific claims about COVID-19 until publishing the research paper. That may change now, Bradnan suggests.
“The accuracy may appear low at first glance, but the ability to detect a communicable disease with any level of accuracy is revolutionary, and with additional funding granted by the U.S. Army, the accuracy is likely to go up,” Bradnan observes. “Clients should view this research as a call to arms regarding the importance of digital biomarkers not just in the fight against COVID-19 but as a way to enter the healthcare space.”
The research paper goes into some detail about the likelihood of various symptoms and the duration of cases of different severity. The raw data from the test is being shared with Stanford University and Scripps Translational Research Institute, the latter of which has also reportedly been working with Garmin.