Predicting Chinese Children and Youth's Energy Expenditure Using ActiGraph Accelerometers: A Calibration and Cross-Validation Study

Abstract

Purpose: The purpose of this study was to develop and cross-validate an equation based on ActiGraph accelerometer GT3X output to predict children and youth's energy expenditure (EE) of physical activity (PA). Method: Participants were 367 Chinese children and youth (179 boys and 188 girls, aged 9 to 17 years old) who wore 1 ActiGraph GT3X accelerometer on their right hip during the following tests/activities: resting metabolic rate (RMR), six 5-min treadmill walk/runs (tested at different speeds: 3 km·h^− 1, 4 km·h^− 1, 5 km·h^− 1, 6 km·h^− 1, 7 km·h^− 1, and 8 km·h^− 1), 1 broadcast gymnastics, and 2 table-tennis exercises. Participants' oxygen consumption was measured using Cosmed K4b². The participants were randomly divided into a calibration group (n = 331, 90%) and a cross-validation group (n = 36, 10%). The calibration group's data were used to determine the relationship between EE and triaxial vector magnitude counts (VM) using the Pearson correlation and to derive the equation using a stepwise multiple regression. In the cross-validation group, differences between measured and predicted EE were evaluated using pairwise t tests. Results: VM activity counts had a moderately high correlation with EE (r = .758, p <  .01). An EE prediction equation was developed: EE (kcal·min^− 1) = 0.00083 × VM+0.073 × weight − 2.01 (R ²  = .72, SEE = 1.45 kcal·min^− 1). According to the cross-validation study results, this equation could predict the EE within the range of known accuracy (i.e., about 20% error). Conclusions: An equation based on ActiGraph accelerometer VM activity counts was derived to predict EE of PA in Chinese children and youth within the range of known accuracy.

Keywords:

• cross-cultural
• physical activity
• regression
• validation