|Barreira, Tiago - Pennington Biomedical Research Center|
|Tudor-locke, Catrine - Pennington Biomedical Research Center|
|Champagne, Catherine - Pennington Biomedical Research Center|
|Broyles, Stephanie - Pennington Biomedical Research Center|
|Harsha, David - Pennington Biomedical Research Center|
|Kennedy, Betty - Pennington Biomedical Research Center|
|Johnson, William - Pennington Biomedical Research Center|
|Allen, Ray - Pennington Biomedical Research Center|
|Katzmarzyk, Peter - Pennington Biomedical Research Center|
Submitted to: Medicine and Science in Sports and Exercise
Publication Type: Abstract Only
Publication Acceptance Date: 2/3/2011
Publication Date: 5/3/2011
Citation: Barreira, T., Tudor-Locke, C., Champagne, C.M., Broyles, S.T., Harsha, D., Kennedy, B.M., Johnson, W.D., Allen, R., Katzmarzyk, P.T. 2011. Comparison of Yamax pedometer and GT3X accelerometer steps in a free-living sample [abstract]. Medicine and Science in Sports and Exercise. 43(5 Supplement):S480.
Technical Abstract: Our objective was to compare steps detected by the Yamax pedometer (PEDO) versus the GT3X accelerometer (ACCEL) in free-living adults. Daily PEDO and ACCEL steps were collected from a sample of 23 overweight and obese participants (18 females; mean +/- sd: age = 52.6 +/- 8.4 yr.; body mass index = 31.0 +/- 3.7 m.kg-2). Participants were instructed to wear both devices for 7 consecutive days, removing them only at the end of each day and during any water activities. The analysis sample included 93 valid days based on wear time = 10 hours for ACCEL and available data for both instruments. Because a pedometer is more likely to be used in a community-based intervention program, it was used as the standard for comparison. Percent difference (PD) and absolute percent difference (APD) were calculated to examine agreement between the PEDO (the standard measure) and the ACCEL. In addition, days were categorized based on PD: a) under-counted (> -10 PD), b) corresponded (-10 to 10 PD), and c) over-counted (> 10 PD). The PEDO and ACCEL detected 8,025 +/- 3,967 and 7,131 +/- 3,066 steps/day, respectively, and the outputs were highly correlated (r = 0.87). Average PD was -3.1% +/- 30.7% and average APD was 23.9% +/- 19.4%. There was a moderate correlation between PEDO steps and the PD between PEDO-steps and ACCEL-steps (r = -0.65), indicating a bias of ACCEL over-counting at lower PEDO steps/day and under-counting at higher PEDO steps/day. Relative to the PEDO, a total of 53% of the days detected by the ACCEL were classified as under-counted, 25% as corresponding, and 23% as over-counted. Despite the high correlation between PEDO and ACCEL steps observed in this adult overweight/obese community-based sample, the mean values for ACCEL were lower than for PEDO steps, and there was a high level of disagreement. Caution is needed when directly comparing the output of these two instruments or using their output interchangeably.