Classification Accuracy of Actigraph MVPA Cutpoints in Pre-school Children: A Whole Body Calorimetry Study

Objective To evaluate the classification accuracy of five ActiGraph moderate-to-vigorous physical activity (MVPA) cut-point definitions in pre-school children using energy expenditure measured by whole-room calorimetry as the criterion measure.

Methods 20 children aged 4-6 years (5.1 ± 0.8years) completed a ~50 min whole-room calorimeter protocol involving age-appropriate moderate and vigorous physical activities (eg. rumling on the spot, playing mini basketball, hopscotch). Each child wore an ActiGraph GT3X for the duration of the protocol. Physical activity intensity was classified every 15 seconds for five ActiGraph cut-points: ?420 counts/15sec (Pate et al., 2006); >573 counts/15sec (Evenson et al., 2008); >584 counts/15sec (Van Cauwenberghe et al., 2010); ?800 counts/15sec (Puyau et al., 2002), and; >890 counts/15sec (Sirard et al., 2005). Classification accuracy was evaluated using Kappa statistics, sensitivity, specificity and area under the receiver operating characteristics curve.

Results The percent agreement in classifying moderate-to-vigorous intensity physical activity will be presented for each Actigraph equation. Specificity and sensitivity levels in detecting moderate-to-vigorous intensity physical activity will be reported along with the classification accuracy as measured by the area under the ROC curve.

Discussion The results will be discussed against the backdrop of current literature in this area and explanations provided as to the usefulness of each equation in classifying MVPA in pre-school children.