Article Title



J.K. Coley, A.X. Bruns, B.E. Perin, D.P. Heil, FACSM,

Montana State University, Bozeman, MT USA

Wrist-worn electronic devices are common for everyday monitoring of many biophysical metrics (e.g., heart rate, SpO2, physical activity, energy expenditure (EE), and sleep quality). Without exception, these devices always include an accelerometry-based physical activity monitor (AM) for indirectly sensing whole-body motion and EE metrics. Due to the need for FDA clearance, however, AMs for clinical assessments have had less attention than their non-clinical commercial counterparts. PURPOSE: This study’s purpose is to determine the accuracy of a clinical wrist-worn AM to predict energy expenditure during steady-state walking in both lab (treadmill, TM) and free-living (overground, OG) settings. METHODS: Twenty college-aged women (Mean±SD: 21±3 yrs; 22.3±2.3 kg/m2 BMI; 21.9±5.6 %BF; n=11) and men (19±1 yrs; 24.2±1.9 kg/m2 BMI; 14.3±7.2 %BF; n=9) were recruited to participate in a one-hour visit to the MSU Human Performance Lab. For each subject, direct measures of EE (portable indirect calorimetry) and measures of predicted EE from a wrist-worn AM designed for clinical environments were collected simultaneously during 10-mins of quiet sitting (RMR), 5 mins each of self-selected slow, medium, and fast-paced TM walking, and 5 mins each of the same speeds during outdoor OG walking, paced with a pacer using GPS. After collection, predicted measures of activity EE (PAEE, kcals/min) from the AM were then compared directly with calculated measures of AEE (EE for walking activity - EE for RMR; kcals/min) using paired t-tests for all 6 TM and OG walking conditions (𝞪=0.01 after Bonferroni adjustment). RESULTS: Predicted mean AEE for TM walking was higher than mean AEE for all TM speeds (Mean difference = +0.9 to +1.4 kcals/min; P=0.01-0.03), but this difference was only significant at the fast TM speed (P=0.01). Additionally, both the slope and y-intercept values for the regression of TM PAEE on AEE differed significantly from those of the line of identity (P<0.001). In contrast, PAEE was statistically similar to AEE at all speeds for OG walking (Mean differences = +0.3 to -0.5 kcals/min; P=+0.34 to +0.53). CONCLUSION: Our findings indicate that this clinical wrist-worn AM was accurate for OG walking, but tended to overpredict AEE for TM walking. This AM should predict AEE relatively well for free-living walking activities.

This document is currently not available here.