Early research suggests that fitness trackers can predict COVID-19 by tracking changes in a person’s activity.
One of the keys to preventing the spread of SARS-CoV-2 – the virus responsible for the COVID-19 illness – is quick identification, tracing and isolation of cases before it can be spread further to others. This has been particularly challenging in part due to the lack of quick and reliable testing. In the United States, current screening measures for COVID-19 consist of survey questions about travel history and symptoms, in addition to temperature measurements. However, the current screening methods may not be reliable enough as many individuals with COVID-19 may be asymptomatic or pre-symptomatic (which make up approximately 40 to 45% of COVID-19-infected individuals), but are still infectious. Further, a high temperature measurement (above 37.8°C or 100°F) is not as commonly seen with COVID-19 illness as many believe. Research indicates that only 12% of COVID-19 positively tested individuals have a high temperature and only 31% of hospitalized patients with COVID-19 at admission have a high temperature.
As such, researchers in the United States have begun to explore the role of wearable sensor data (e.g. from smartwatches or activity trackers) to understand how fitness trackers can predict COVID-19. The researchers developed an app-based research platform and database (DETECT) where wearable sensor data, self-reported symptoms, diagnoses and electronic health information can be shared by users. The goal is to use the shared health data to better identify and track viral illness levels, such as COVID-19, in individuals and the population. In the DETECT (Digital Engagement and Tracking for Early Control and Treatment) study, researchers investigated whether wearable sensor data can be used in addition to self-reported symptom data to improve the identification of COVID-19 positive cases versus COVID-19 negative cases in participating users. The results of the study were published in Nature Medicine.
As of June 7, 2020, more than 30,000 participants had enrolled in the study, with individuals represented all over the United States and connected to various fitness tracker devices, such as Fitbit, Apple HealthKit, or Google Fit. Of the 3,811 participants who self-reported symptoms, 54 had tested positive for COVID-19 and 279 had tested negative. Analysis of the sensor and health data from those who reported symptoms and had tested for COVID-19 revealed that decreased activity and increased sleep (compared to the individual’s normal baseline levels) were significant factors in predicting a positive COVID-19 case. The team of researchers were able to use the data from the fitness trackers to predict with 80% accuracy whether an individual self-reporting symptoms was likely to have an infection with the SARS-CoV-2 coronavirus.
The early results of this research show that changes in physiological activity captured by fitness trackers can potentially be used to facilitate a more efficient and cost-saving testing strategy and therefore help public health officials control the spread of the disease. The team of researchers are now recruiting more participants to further their research with hopes of improving their model of using fitness trackers to predict COVID-19 and other viral illnesses.
Written by Maggie Leung, PharmD.
Quer, G., Radin, J.M., Gadaleta, M. et al. Wearable sensor data and self-reported symptoms for COVID-19 detection. Nat Med (2020). https://doi.org/10.1038/s41591-020-1123-x
Early results from DETECT study suggest fitness trackers can predict COVID-19 infections. (2020, October 29). Retrieved from https://www.eurekalert.org/pub_releases/2020-10/sri-erf102820.php
Image by Steve Buissinne from Pixabay