Effects of Brief Intermittent Walking Bouts on Step Count Accuracy of Wearable Devices

• DOI: 10.1123/jmpb.2018-0050
• Publication Date: 2019-02-13T04:59:51Z
• Authors: Lindsay P. Toth; Susan Park; Whitney L. Pittman; Damla Sarisaltik; Paul R. Hibbing; Alvin L. Morton; Cary M. Springer; Scott E. Crouter; David R. Bassett
• Keywords: Rest (music); Activity monitor; Pedometer

Purpose: To examine the effect of brief, intermittent stepping bouts on step counts from 10 physical activity monitors (PAMs). Methods: Adults (N = 21; M ± SD, 26 9.0 yr) wore four PAMs wrist (Garmin Vivofit 2, Fitbit Charge, Withings Pulse Ox, and ActiGraph wGT3X-BT [AG]), hip (Yamax Digi-Walker SW-200 [YX], Zip, Omron HJ-322U, AG), two ankle (StepWatch [SW] with default modified settings). AG data were processed without low frequency extension (AGL) Moving Average Vector Magnitude algorithm. In Part 1 (five trials), walking varied (4–12 steps) rest intervals held constant (10 s). 2 (six (4 (1–10 Percent hand-counted steps mean absolute percentage error calculated. One sample t-test was used to compare percent 100%. Results: Parts SWdefault, SWmodified, YX, AGLhip captured within 10% across nearly all conditions. 1, estimates most methods improved as number per bout increased. decreased duration Conclusion: Most required >6–10 consecutive record steps. Rest 1–2 seconds sufficient break up in many methods. The requirement for several some causes an underestimation bouts.