Fitness trackers help catch worsening COVID-19 symptoms early

Monitoring estimated oxygen variation (EOV) using a wearable fitness tracker may facilitate early detection of coronavirus disease 2019 (COVID-19) symptom exacerbation and prevent untimely death, according to a new study.

“The high EOV signal by Fitbit demonstrated a favourable sensitivity and high negative predictive value (NPV) for COVID-19 symptom exacerbations before their onset,” the researchers said. “High sensitivity and NPV of the device and signal used in this study are of particular importance for screening/early diagnosis of COVID-19 exacerbation that could accurately identify all possible cases that could warrant closer inpatient or outpatient monitoring.”

Twenty-three COVID-19 patients participated in the study and were asked to wear a Fitbit for 30 days, during which EOV and resting heart rate (RHR) were measured. High EOV signal was defined as a significant dip and recovery in the patient’s oxygen level within the index period; high RHR signal was defined as daily RHR >5 beats/day relative to the minimum RHR of the patient during the study period.

The primary outcome was a composite of all-cause death, exacerbation of COVID-19 symptoms, oxygenation, and the use of mechanical ventilation and extracorporeal membrane oxygenation. When either RHR or EOV signal occurred within 2 days of the primary outcome, the researchers considered that as successful prediction.

There were 48 high EOV signals during the 30-day study duration, with each patient showing a median of 1 (interquartile range, 1–3) high EOV signal. The primary outcome occurred seven times in six patients: two patients were put on high-flow nasal cannula, two experienced exacerbation of coughs, and one each had worsening dysosmia and fever/general malaise. [JMIR Form Res 2021;doi:10.2196/30819]

After excluding one patient due to obstructive sleep apnoea, the researchers found that four of five patients with symptom exacerbation experienced high EOV signals within 2 days of such events. This corresponded to a sensitivity of 80 percent and a specificity of 90 percent. Moreover, while the positive predictive value (PPV) was only 9.3 percent, NPV was excellent at 99.7 percent.

In comparison, RHR only had sensitivity, specificity, NPV, and PPV values of 50 percent, 88 percent, 99.1 percent, and 6.7 percent, respectively.

Despite its promising results, several limitations are worth taking note of, the researchers said. In particular, the small sample size and the very low number of primary outcome episodes limit the generalizability of the findings. In addition, Fitbit has yet to be approved as a medical device.

Further studies with larger samples, longer wear times, and more thorough controls for background characteristics will help to validate the current results.

“In future, we hope to integrate EOV and other physiological metrics such as RHR, respiratory rates, or sleep data to improve the prediction accuracy of COVID-19 symptom exacerbations in advance,” the researchers said.