Abstract
There is increased awareness of the health benefits of regular exercise, and quantifying daily activity has become popular. Consequently, there are an increasing number of devices for measuring physical activity. Healthcare professionals and the general public should know the accuracy and limitations of these devices to better determine which ones suit their needs.
Ten devices were tested: one ankle-based device, StepWatch™ Activity Monitor (SAM); two wrist-based devices, FitBit Force™ and Nike+ Fuelband SE; seven waist-based devices, Omron HJ-321 Pedometer, Sportline 340 Strider Pedometer, FitBit One™, Samsung Galaxy S4 utilizing the two most popular applications (Runtastic and Noom Walk), and the iPhone 5 utilizing the two most popular applications (Runtastic and ARGUS). Thirty healthy volunteers, mean age 25.6 years (range 20–30) and mean body mass index 23.5 (range 17.3–29.0), completed the following protocol: (1) walk briskly around a 400-M track simulating community ambulation (2) jog around a 400-M track (3) walk slowly for 10-M, approximating household or workplace pace (4) ascend 10 steps, and (5) descend 10 steps. Each subject completed 3 trials for each task. Manual count was the gold standard (Champion Sports Tally Counter). Accuracy and mean percent error were calculated to demonstrate overall performance and any tendencies for over or undercounting. An Aggregate Accuracy Score was calculated using the mean accuracy of each activity and multiplying by a corresponding weighted value for a prototypical person: 400-M walk represents community ambulation, weighted 40%; 10-M walk represents household and workplace ambulation, weighted 30%; 400-M jog represents jogging or running, weighted 20%; Stair Ascent and Descent represent community and household stair use, weighted 5% each.
Device rank based on the Aggregate Accuracy Score was #1 FitBit One™ (98.0%), #2 Omron HJ-321 (97.0%), #3 StepWatch™ Activity Monitor (93.3%), #4 Runtastic Google App (92.7%), #5 Runtastic iPhone App (89.5%), #6 Fitbit Force™ (88.2%), #7 Argus iPhone App (87.2%), #8 Sportline 340 Strider (85.7%), #9 Nike Fuelband (76.1%), #10 Noom Walk Google App (75.9%). The FitBit One™ was 99.5%, 97.8%, 96.7%, 94.3%, and 96.9% accurate in the 400-M walk, 10-M walk, 400-M jog, 10 stair ascent, and 10 stair descent, respectively. The Omron HJ-321 was 99.3%, 94.9%, 97.9%, 92.2%, and 91.3% accurate, respectively. The SAM performed well (>95% accurate) in all activities except one, consistently undercounting the 400-M jog by about 25% (95% CI: −27.2% – −23.9%). The FitBit ForceTM and Nike+ Fuelband SE wrist devices were ≥90% accurate in the 400-M walk and 400-M jog, but ≤83% accurate for all other activities. Three of the 4 smartphone applications were >97% accurate in the 400-M walk, 1 of 4 was 97.3% accurate in the 400-M jog, but all devices performed poorly (≤90% accurate) for all other activities.
Smartphones are very popular, but current technology is less accurate for measuring overall daily activity. The relatively inexpensive FitBit One™ and Omron HJ-321 pedometer are highly accurate for quantifying a variety of activities, including running. The StepWatch™ Activity Monitor performs well in lower cadence, but consistently undercounted jogging. Wrist-based activity devices are not as accurate as waist-based. Next generation technologies, including smartphones, should undergo accuracy testing before recommending them for daily use.
