The accuracy of various activity trackers in estimating steps taken and energy expenditure

Purpose: Activity monitors are designed for helping to monitor the quantity and intensity of exercise training. While there is a reasonable experience with step counters, there are a rapidly expanding number of devices available on the marketplace, many purporting to be able to accurately measure energy expenditure (EE). This study was designed to test the accuracy of step count and energy expenditure in several commercially available activity monitors. Methods: Healthy, physically-active volunteers (N=20) performed treadmill walking, treadmill running, elliptical (arm + leg) exercise and an agility drill that included some basketball actions while wearing step counters/activity monitors (Jawbone UP, Nike Fuelband, Fitbit Ultra, NL-2000i, Adidas MiCoach, Body Media FIT Core). Criterion step counts were measured by direct visual observation and EE was measured by a portable metabolic system. Results: During normal ambulation and elliptical exercise, most of the measured step counts were reasonably accurate (+10% of the criterion measure). During the agility drills, the errors in step counts were systematically less than the criterion measure. Although some of the devices were reasonably accurate for measuring EE during normal ambulation, the errors in measuring EE were, in general, unacceptably large and became larger with non-standard ambulation. However, measured across the entire range of activities, both step counts and EE were broadly accurate if not precise. Conclusions: The results suggest that step count accuracy during normal ambulation can be measured accurately by a number of devices. However, during non-standard ambulation (particularly for measurement of EE), these contemporary activity monitoring devices require improved algorithms.; ;