Comparison of two methods to assess PAEE during six activities in children.

Purpose: The purpose of this study was to compare the accuracy of physical activity energy expenditure (PAEE)-prediction models using accelerometry alone (ACC) and accelerometry combined with heart rate monitoring (HR+ACC) to estimate PAEE during six common activities in children (lying, sitting, slow and brisk walking, hop-scotch, running). Three PAEE-prediction models derived using the current data, and five previously published prediction models were cross-validated to estimate PAEE in this sample. Methods: PAEE was assessed using ACC, HR+ACC, and indirect calorimetry during six activities in 145 children (12.4 +/- 0.2 yr). One ACC and two HR+ACC PAEE-prediction models were derived using linear regression on data from the current study. These three new models were cross-validated using a jackknife approach, and a modified Bland-Altman method was used to assess the validity of all eight models. Results: PAEE predictions using the one ACC and two HR+ACC models derived in the current study correlated strongly with measured values (RMSE = 97.3-118.0 J[middle dot]min-1[middle dot]kg-1). All five previously published models agreed well overall (RMSE = 115.6-245.3 J[middle dot]min-1[middle dot]kg-1), but systematic error was present for most of these, to a greater extent for ACC. Conclusions: ACC and HR+ACC can both be used to predict overall PAEE during these six activities in children; however, systematic error was present in all predictions. Although both ACC and HR+ACC provide accurate predictions of overall PAEE, according to the activities in this study, PAEE-prediction models using HR+ACC may be more accurate and widely applicable than those based on accelerometry alone