There may be insufficient evidence to support the broad use of consumer wearable devices for outpatient health monitoring in patients with post–acute sequelae of SARS-CoV-2 infection and related postinfection syndromes, although heart rate monitoring and atrial fibrillation detection showed the strongest support in a recent review.
In a preregistered systematic umbrella review, investigators evaluated the accuracy, validity, reliability, and clinical utility of consumer wearable devices for biometrics relevant to post–acute sequelae of SARS-CoV-2 infection (PASC) and related syndromes. Following a search of 3,988 records, they identified 42 eligible reviews encompassing more than 30 wearable brands and more than 150 device series. The included reviews evaluated wearable-derived measures such as heart rate, heart rate variability, atrial fibrillation detection, arrhythmia detection, blood pressure, physical activity metrics, energy expenditure, sleep measures, oxygen saturation, electrodermal activity, and skin temperature.
Among the reviewed biomarkers, heart rate monitoring had the largest evidence base. Fifteen reviews assessed heart rate accuracy across 42 devices. Overall, wearable-derived heart rate measurements demonstrated variable but generally favorable performance when compared with clinical-grade reference standards and other consumer devices. Apple Watch devices received the most consistent evaluations of heart rate across the included reviews. The investigators also found that measurement accuracy varied according to activity level, health status, and patient characteristics. Several reviews reported reduced accuracy during high-intensity exercise and in patients with darker skin tones, although the devices using red-spectrum light appeared less affected.
Six reviews evaluated atrial fibrillation detection across 16 devices, most of which demonstrated moderate-to-high accuracy when benchmarked against clinical-grade electrocardiography. Across the reviews, Samsung devices achieved reported sensitivity as high as 97% and specificity as high as 98%. Apple Watch devices generally performed well but reported sensitivity varied substantially among the reviews.
The evidence was less consistent for other wearable-derived measures. Sixteen reviews assessed step-count accuracy and found substantial variation across devices, populations, and settings. Accuracy was reduced in patients with slower gait speeds, a finding the investigators noted may be particularly relevant in patients with postinfection syndromes. Fourteen reviews evaluated energy expenditure estimates and generally found low or inconsistent accuracy. Some reviewed studies reported substantial measurement error, with performance varying by sex, body mass index, activity type, and device placement.
Five reviews examined sleep-related measures, which included total sleep time, sleep efficiency, sleep onset latency, and wake after sleep onset. The results varied according to the device, population, and comparator. Fitbit and Jawbone devices frequently overestimated some sleep measures when compared with polysomnography, while the findings differed when compared with actigraphy or sleep logs.
The investigators also identified substantial gaps in the evidence base. They noted little or no review-level evidence for oxygen saturation, skin temperature, electrodermal activity, flights of stairs climbed, and arrhythmias other than atrial fibrillation. Evidence supporting the devices' clinical utility was similarly limited. Although several reviews found that wearable use may increase physical activity, the investigators identified no review-level evidence demonstrating that wearable-based monitoring improves morbidity, self-management, or health care utilization in patients with PASC or related postinfection syndromes.
Several limitations may affect the interpretation of the findings. The review quality was highly variable. On average, included reviews met fewer than half of the applicable AMSTAR 2.0 quality criteria, and only 16 of the 42 reviews conducted formal risk-of-bias assessments. Most of the available evidence was derived from medically healthy populations rather than patients with PASC or other chronic postinfection conditions. Device performance frequently varied according to age, skin tone, gait speed, health status, and setting of use. The investigators found no review-level evidence addressing adverse effects associated with wearable use.
The findings suggested that evidence supporting consumer wearable use in PASC remains limited, despite growing interest in these technologies. Heart rate monitoring and atrial fibrillation detection currently have the strongest supporting evidence, whereas evidence for many other biomarkers remains sparse or inconsistent.
"There is not sufficient evidence to recommend the use of consumer wearables for self-monitoring to the millions of individuals living with PASC and other post–acute infection syndromes," wrote lead study author Deanna M. Kaplan, of the Department of Family and Preventive Medicine at the Emory University School of Medicine and the Department of Spiritual Health at Emory Healthcare, and colleagues.
Full disclosures of the study authors can be found in the study.
Source: PLOS Digital Health