Table 2 Accuracy problems in studies using physiological measurements by wearables
Measure | Number tested | Wearables | Patients | Finding | Study |
|---|---|---|---|---|---|
Total energy expenditure | 12 devices | Withings Pulse Jawbone Garmin Vivofit Suzuken Lifecorder EX Panasonic Actimaker Epson Pulsense Tanita-AM-160 Fitbit Flex Misfit Shine Omron Active Style Pro Omron CaloriScan | 19 healthy adults, not obese. | “absolute values differed widely among products and varied significantly from the gold standard measures” | Murakami et al. (2016) |
Step count | 3 pedometers | Yamax Digiwalker Fitbit | 14 young health participants walking at 3 speeds | “all the evaluated devices had high error rates at 1 km/h” (slow walking speeds). | Beevi et al. (2016) |
Step count | 10 activity trackers | Polar Loop Garmin Vivosmart Fitbit Charge HR Apple Watch Sport Pebble Smartwatch Samsung Gear S Misfit Flash Jawbone Up Move Flyfit Moves | 31 healthy participants on a treadmill | “Test–retest validity depends on walking speed”; “consumer activity trackers perform better at average and vigorous walking speed…” | Fokkema et al. (2017) |
Sleep | 2 monitors | Withings Pulse Jawbone Up | 36, including 22 with obstructive sleep apnea (OSA) | “confirmed… in patients suffering from OSA, the limited performance of wearable sleep monitors” | Gruwez et al. (2019) |
Sleep | 2 monitors | Fitbit Change 2 Neuroon | 25 students | Underestimate light sleep and overestimate deep sleep. “Reasonably satisfactory for general purpose and non-clinical use” | Liang and Martell (2018) |