Validity of the Crouter 2-Regression Model for Predicting Energy Expenditure: Transitions from Rest to Activity

The Crouter 2-regression model (15) is a new approach for predicting MET values. However, the Crouter 2-regression model may misclassify walking/jogging during transition periods (i.e. during the first and last minute of the activity) due to the high coefficient of variation resulting in an overestimation of MET values. Purpose: The purpose was to examine the Crouter 2-regression model for predicting MET values during transitions from rest to activity and back to rest. It was hypothesized that starting a walking bout partway through a minute will lead to a misclassification, due to the high coefficient of variation, and therefore over-predict MET values. Methods: Thirty participants (age, 28.2 ± 7.7 yrs; BMI, 24.6 ± 3.6 kg^.m^-2) volunteered to perform 15 minutes of supine rest followed immediately by 8 minutes of each of the following activities: one-on-one basketball, seated rest, over-ground walking, and seated rest. An ActiGraph GT1M was positioned on the right hip and a portable metabolic unit (Cosmed K4b²) was worn during all activities. The participants were randomly assigned to start the activity bouts at 0-s, 20-s or 40-s into the minute (according to the ActiGraph clock). Results: The Crouter 2-regression model under-predicted the total energy cost of basketball compared to the Cosmed K4b² by an average of 12.8 MET·minutes (P < 0.001) (15.9%). During walking, the Crouter 2-regression model over-predicted MET values compared to the Cosmed K4b² during the transitional minutes in the 20-s and 40-s conditions by an average of 2.2 METs (P < 0.001). However, it accurately predicted MET values during minutes 2-7 part of the bout. Conclusion: The energy cost of basketball was under-predicted by the Crouter 2-regression model. During the walk-to-rest and rest-to-walk transitions, the Crouter 2-regression model significantly over-predicted the MET values in the 20-s and 40-s conditions. This over-prediction was due to the high coefficient of variation and the oxygen deficit and debt. Despite these limitations, the Crouter 2-regression model still provided a closer estimate of MET values compared to the Cosmed K4b² for basketball and walking compared to other commonly used single-regression equations for the ActiGraph. Supported by NIH Grant OIR21 CA122430-01