ActiGraph GT3X+ cut-points for identifying sedentary behaviour in older adults in free-living environments

Abstract

Objectives: To determine the ActiGraph GT3X+ cut-points with the highest accuracy for estimating time spent in sedentary behaviour in older adults in free-living environments. ActivPAL(3)™ was used as the reference standard.

Design: Cross-sectional study.

Methods: 37 participants (13 males and 24 females, 73.5 ± 7.3 years old) wore an ActiGraph GT3X+ and an ActivPAL(3)™ for 7 consecutive days. For ActivPAL(3)™, variables were created based on posture. For ActiGraph GT3X+, sedentary behaviour was defined as (1) vector magnitude and (2) vertical axis counts for 1-s, 15-s and 1-min epochs, with cut-points for 1-s epochs of <1 to <10 counts, for 15-s epochs of <1 to <100 counts and for 1-min epochs of <1 to <400 counts. For each of the ActiGraph GT3X+ cut-points, area under the receiver operating characteristic curve (area under the curve), sensitivity, specificity, and percentage correctly classified were calculated. Bias and 95% limits of agreement were calculated using the Bland-Altman method.

Results: The highest areas under the curve were obtained for the vector magnitude cut-points: <1 count/s, <70 counts/15-s, and <200 counts/min; and for the vertical axis cut-points: <1 count/s, <10 counts/15-s and <25 counts/min. Mean biases ranged from -4.29 to 124.28 min/day. The 95% limits of agreement for these cut-points were ± 2 h suggesting great inter-individual variation.

Conclusions: The results suggest that cut-points are dependent on unit of analyses (i.e. epoch length and axes); cut-points for a given epoch length and axis cannot simply be extrapolated to other epoch lengths. Limitations regarding inter-individual variability and misclassification of standing activity as sitting/lying must be considered.

Keywords: Actigraphy; ActivPAL; ROC analysis; Sensitivity and specificity; Triaxial accelerometer.