JMIR Publications

ABSTRACT

Background:

Wearable devices hold great promise particularly for data generation for cutting-edge health research, and demand has risen dramatically in recent times. However, there is a shortage in aggregated insights for how wearables have been used in health research.

Objective:

We aim to broadly overview and categorize the current research conducted with consumer-grade wearable devices.

Methods:

We performed a scoping review to understand the uses of consumer-grade wearables for health research from a population-health perspective using PRISMA-Scoping Review framework. We found a total of 7 499 articles in four medical databases (PubMed, Ovid, Web of Science and CINAHL). Studies were eligible if they used non-invasive wearables; i) worn on the wrist, arm, hip and chest, ii) measured vital signs, and iii) analyse collected data quantitatively. We excluded studies not using wearables for outcome assessment and those that used prototypes, devices that cost >500€ or obtrusive smart clothing.

Results:

We found 179 studies involving 10 835 733 participants. Most studies were observational (n=128, 72%), conducted in 2020 (n=56, 31%) and in North America (n=94, 53%), and 93% (10 104 217) of participants featured in global health studies. Typical wearable measures comprised fitness trackers (n=86, 46%) and accelerometer wearables (measures movement) (n=49, 26%) that counted steps (n=95, 53%), heart rate (n=55, 31%), as well as sleep duration (n=51, 28%). Rarely used devices typically measured blood pressure (n=4, 2%), skin temperature (n=3, 2%), oximetry (n=3, 2%), or respiratory rate (n=2, 1%). The most popular wearables were; i) worn on the wrist (n=138, 73%), ii) cost <200€ (n=120, 63%), and iii) used accelerometery (n=145, 83%). Reviewing aims and approaches of all 179 studies revealed six prominent uses for wearables, comprising: i) correlations - wearable and other physiological data (n=40, 22%), ii) method evaluations (with subgroups) (n=40, 22%), iii) population-based research (n=31, 17%), iv) experimental outcome assessment (n=30, 17%), v) prognostic, forecasting (n=28, 16%), and vi) explorative analysis of big datasets (n=10, 6%). The most frequent strengths of consumer-grade wearables were validation, accuracy and clinical certification (n=104, 58%).

Conclusions:

The potential and application of consumer-grade wearables used in research increased in the last years. Although most studies perceived wearables as advantageous for research, some studies experienced shortcomings like inaccuracy or technical issues. There is a lack of research conducted of wearable devices in low-resource contexts. Fuelled by the COVID-19 pandemic, we see a shift to more large-sized, online studies where wearables were used to increase knowledge of the developing pandemic, including forecasting models and effects of the pandemic on different populations regarding age, nationality and morbidity. Big data extracted from wearables may potentially transform understanding of populations, health trends and forecasts, as some studies in this field adumbrate. Wearables – although often piloted in the studies – also showed an increasingly diverse field of application and included also rare and possibly underrepresented populations.