Laboratory calibration and validation of the Biotrainer and Actitrac activity monitors

Abstract

Purpose: The Biotrainer and Actitrac activity monitors (IM Systems) offer potential research advantages over existing accelerometry-based activity monitors, but they have not been tested under controlled conditions. The purpose of this study was to develop and test laboratory-based prediction equations for both monitors to estimate energy expenditure (EE) for walking/running movements. Methods: Participants in the study wore a Biotrainer and Actitrac monitor on both hips and completed three paced bouts on the treadmill (3, 4, and 6 mph for 6 min each). Metabolic data collected using an indirect calorimetry system were used as the criterion measure. Multiple regression techniques were performed to develop prediction equations, and these equations were then applied to data from a separate sample for cross-validation purposes. Reliability was also examined. Results: The correlations between the raw counts from each monitor and the measured metabolic variables ranged from r = 0.74-0.88 for the Biotrainer and from r = 0.81-0.91 for the Actitrac. The equations predicting EE (kcal·min-1) from counts yielded strong validation results for both the Biotrainer (R2 = 0.88, SEE = 1.47) and the Actitrac (R2 = 0.91, SEE = 1.24). When used on the cross-validation sample, the correlations between measured and predicted EE were r = 0.93 (Biotrainer) and r = 0.94 (Actitrac). Intraclass reliability coefficients computed between the left and right monitors ranged from 0.60 to 0.71 (Biotrainer) and 0.80 to 0.87 (Actitrac). When the equation developed from one side was applied to data from the monitor on the other side, there were no significant differences in predicted and measured EE for most comparisons. Conclusion: The results support the validity of Biotrainer and Actitrac monitors for estimating energy expenditure under controlled conditions.