Cross-validation of Actigraph derived accelerometer cut-points for assessment of sedentary behaviour and physical activity in children aged 8-11 years

Abstract

Aim: To cross-validate previously calibrated Actigraph cut-points in children.

Methods: Twenty-eight children (50% boys) aged between 8 and 11 years of age (9.4 ± 1.4 years) performed a series of 5 minute bouts of activity reflective of different levels of PA from sedentary behaviour (SB) to moderate to vigorous physical activity (MVPA); $\dot{V}$ O₂ was assessed using breath-by-breath indirect calorimetry, and activity was assessed using Actigraph accelerometers worn on the hip and non-dominant wrist. The $\dot{V}$ O₂ values were then converted into age-specific METs (measured METs) and coded as SB, light PA and MVPA. Accelerometer data were analysed using previously calibrated cut-points at different epochs, that is 5, 15, 30 and 60 seconds.

Results: Receiver operating characteristic (ROC) curve analysis indicated that there was excellent discrimination of SB using the Evenson et al (15 seconds), Romanzini (15 seconds), Treuth et al (30 seconds), Freedson et al (60 seconds), Treuth et al (60 seconds) and Troiano et al (60 seconds) cut-points. ROC analysis indicated poor discrimination for LPA irrespective of the cut-points used. Good discrimination of MVPA was evident for all existing cut-points using a 60-second epoch.

Conclusion: There is considerable variation in the performance of existing cut-points for assessment of SB, LPA and MPA in children.

Keywords: cut-points; energy expenditure; indirect calorimetry; motor skills.