Different Accelerometry Interpretation Methods Alter Physical Activity Classifications

Background: Multiple cutpoints exist for reducing raw accelerometry data into specific intensity categories (i.e. light, moderate, and vigorous). Historically, Freedson et al. (1998) uniaxial cutpoints have been most readily used. Multiple triaxial cutpoints now exist, but no standard interpretation methods have been accepted. Theoretically, using different interpretation methods may result in significantly different physical activity (PA) classifications from the same raw data.

Purpose: To compare the effect of using three different sets of cutpoints with regard to classifying subjects as meeting or not meeting PA guidelines.

Methods: Subjects (n=84; age 45.5±20.5 yr, BMI 24.9±4.5 kg/m2) wore an initialised (60Hz) ActiGraph GT3X+ accelerometer, positioned at their right hip, for a minimum of 4 days with 10 hours valid wear-time per day. Raw accelerometry data were processed in Actilife v6.8.0 using a 60s epoch. PA was quantified by scoring data using generalised cutpoints developed by Sasaki et al. (2011) and age-specific cutpoints developed by Miller et al. (2010) and Santos-Lozano et al. (2013). Subjects were classified as active or inactive based on the criteria of accumulating >150 minutes/week of MVPA, or >75 minutes/week of vigorous PA, in >10 minute bouts. PA classifications were compared across the three cutpoint outputs using a Chi-Square test, with significance set at p<0.05.

Results: The number of subjects classified as active were significantly different between each of the three sets of cutpoints (p<0.001). The number of subjects classified as active across the three methods were 28, 31, and 35 for Miller et al., Santos-Lozano et al., and Sasaki et al., respectively.

Conclusion: The use of different accelerometry cutpoints alters the classification of an individual’s PA status. A consensus should be established for cutpoints, to enable researchers and clinicians to accurately evaluate activity status. Future research should examine the use of relative intensity cutpoints to negate the misclassification of those meeting and not meeting PA guidelines.