Wrist accelerometer cut-points for classifying sedentary behavior in children

Introduction To examine the validity and accuracy of wrist accelerometers for classifying sedentary behavior (SB) in children. Methods Fifty-seven children (5-8y and 9-12y) completed a ~170min protocol including 15 semi-structured activities and transitions. Nine ActiGraph (GT3X+) and two GENEActiv wrist cut-points were evaluated. Direct observation was the criterion measure. The accuracy of wrist cut-points was compared to that achieved by the ActiGraph hip cut-point (≤25 counts/15s) and the thigh-mounted activPAL3^TM. Analyses included equivalence testing, Bland-Altman procedures and area under the receiver operating curve (ROC-AUC). Results The most accurate ActiGraph wrist cut-points (Kim, vector magnitude: ≤3958 counts/60s and vertical axis: ≤1756 counts/60s) demonstrated good classification accuracy (ROC-AUC = 0.85-0.86) and accurately estimated SB time in 5-8y (equivalence p=0.02; mean bias: 4.1%, limits of agreement [LoA]: -20.1-28.4%) and 9-12y (equivalence p<0.01; -2.5%, -27.9-22.9%). Mean bias of SB time estimates from Kim were smaller than ActiGraph hip (5-8y: 15.8%, -5.7-37.2%; 9-12y: 17.8%, -3.9-39.5%) and similar to or smaller than activPAL3^TM (5-8y: 12.6%, -39.8-14.7%; 9-12y: -1.4%, -13.9-11.0%), although classification accuracy was similar to ActiGraph hip (ROC-AUC = 0.85) but lower than activPAL3^TM (ROC-AUC = 0.92-0.97). Mean bias (5-8y: 6.5%, -16.1-29.1%; 9-12y: 10.5%, -13.6-34.6%) for the most accurate GENEActiv wrist cut-point (Schaefer: ≤0.19g) was smaller than ActiGraph hip, and activPAL3^TM in 5-8y, but larger than activPAL3^TM in 9-12y. However, SB time estimates from Schaefer were not equivalent to direct observation (equivalence p>0.05) and classification accuracy (ROC-AUC = 0.79-0.80) was lower than for ActiGraph hip and activPAL3^TM. Conclusion The most accurate SB ActiGraph (Kim) and GENEActiv (Schaefer) wrist cut-points can be applied in children with similar confidence as the ActiGraph hip cut-point (≤25 counts/15s), although activPAL3^TM was generally more accurate.