Secondary Logo

Share this article on:

Accuracy Of Wrist-worn Activity Monitors At Three Walking Speeds On The Treadmill: 2783 Board #306 June 3, 930 AM - 1100 AM

Smith, John D.

Medicine & Science in Sports & Exercise: May 2016 - Volume 48 - Issue 5S - p 783
doi: 10.1249/01.mss.0000487352.64658.a5
E-40 Free Communication/Poster - Research Methodology Friday, June 3, 2016, 7:30 AM - 12:30 PM Room: Exhibit Hall A/B

Texas A&M University-San Antonio, San Antonio, TX.


(No relationships reported)

Wrist-worn activity monitors are becoming increasingly popular with many manufacturers offering such devices.

PURPOSE: To assess the accuracy of wrist-worn activity monitors at three walking speeds on a treadmill.

METHODS: Seventy-seven participants (age = 24.9±3.5 yrs, ht = 170.0±8.9 cm, wt = 78.6±19.9 kg) were fitted with a Fitbit Flex (FF) on the right wrist, a Movband Model 2 (MV), Lifetrak C200 (LT), Garmin Vivofit (GV), Polar Loop (PL) on the left wrist, and an Omron HJ-113 (HJ) on the right waist. Each participant walked for two minutes at 53.6 m/min (2.0 mph), 80.5 m/min (3.0 mph), and 107.3 m/min (4.0 mph). At the end of each trial monitor counts and actual step counts from a hand tally (AC) were recorded. A repeated measures ANOVA was used to determine significant differences between the counts. Single measure intraclass correlation (ICC) from a two-way random effects ANOVA was used to assess the agreement between AC and monitor counts. Pedometer error was calculated as [(monitor steps-actual steps)/actual steps] * 100.

RESULTS: For all speeds, monitor counts were significantly lower than AC (p<.05) except for GV and HJ at 53.6 m/min and 80.5 m/min, and for LT and HJ at 107.3 m/min, none of which were significantly different than AC (p>.05). Agreement according to ICC was highest with HJ for all speeds (0.68, 0.97, and 0.82 during 53.6 m/min, 80.5 m/min, and 107.3 m/min, respectively). Agreement among the wrist worn monitors was highest with GV (0.71) and lowest with FF (0.18) at the slowest speed, highest with GV (0.67) and lowest with LT (0.03) at 80.7 m/min, and highest with PL (0.28) and lowest with LT (0.0) at the fastest speed. Error was greatest with LT at 40.6% while walking at 53.6 m/min, 19.6% with PL at 80.5 m/min, and 19.0% with PL at 107.3 m/min. Error was least with GV at 3.2% while walking at 53.6 m/min. Although error was least with HJ during 53.6 m/min and 107.3 m/min (1.2% and 1.2%, respectively), GV was 3.2% and LT was 7.7% during 80.5 m/min and 107.3 m/min, respectively.

CONCLUSION: The hip-worn pedometer (HJ) provides the most accurate step count across all speeds. Among the wrist-worn monitors, GV seems to provide the most accurate measure of step count in this laboratory-controlled study.

© 2016 American College of Sports Medicine