Comparison Between Wrist-Worn and Waist-Worn Accelerometry

Objective: To use the most recent ActiGraph model (GT9X) to compare counts per minute (CPM) estimates between wrist-worn and waist-worn attachment sites. Methods: Participants completed 2 conditions (laboratory [N = 13] and free-living conditions [N = 9]), in which during both of these conditions they wore 2 ActiGraph GT9X accelerometers on their nondominant wrist (side-by-side) and 2 ActiGraph GT9X accelerometers on their right hip in line with the midaxillary line (side-by-side). During the laboratory visit, participants completed 5 treadmill-based trials all lasting 5 min: walk at 3 mph, 3.5 mph, 4 mph, and a jog at 6 mph and 6.5 mph. During the free-living setting, participants wore the monitors for 8 hours. Paired t test, Pearson correlation and Bland-Altman analyses were employed to evaluate agreement of CPM between the attachment sites. Results: Across all intensity levels and setting (laboratory and free-living), CPM were statistically significantly and substantively different between waist- and wrist-mounted accelerometry. Conclusion: Attachment site drastically influences CPM. As such, extreme caution should be exercised when comparing CPM estimates among studies employing different attachment site methodologies, particularly waist versus wrist.