ActiGraph-Measured Breaks in Sedentary Behavior; Are They Real Transitions From Sitting to Standing?

Barreira T, Schuna, J, Zderic T, Hamilton M, Tudor-Locke, C

International Journal of Exercise Science: Conference Proceedings: 9 (3), Article 9

Previous publications have shown a link between free-living ActiGraph accelerometer-measured breaks in sedentary behavior (BREAKS) and various markers of cardio-metabolic health. To our knowledge, there is limited evidence supporting the validity of ActiGraph-measured BREAKS. A true BREAK has been defined as a transition from sitting to standing. However, the measurement of sedentary behavior by the ActiGraph relies on a lack of movement (usually <100 counts/min) without any indication of body posture and a BREAK has been defined as a transition from <100 to ?100 activity counts/min. PURPOSE: To evaluate the accuracy of the number and time of BREAKS measured by the ActiGraph. METHODS: The activPAL served as the criterion measure since it has been shown to accurately measure BREAKS using the manufacturer s proprietary algorithm that determines a transition from sitting (or lying) to upright behaviors. A total of 15 participants wore both an ActiGraph GT3X+ at the waist and an activPAL on the right thigh for 7 consecutive days (24 h/day - removing them only when in contact with water). This analysis focused on data collected between 7am and 10pm only. Two participants data were excluded due to a lack of compliance with the study protocol. ActivPAL and ActiGraph BREAKS were determined. Data from both devices were matched on minute-by-minute timestamps while also applying a 3-min allowance window to account for known clock drift. Dependent t-tests were used to infer statistical significance between ActiGraph and activPAL BREAK count estimates. RESULTS: The activPAL detected 39 11 BREAKS/day (mean SD) while the ActiGraph detected 74 15 BREAKS/day (p<0.001). On average, the ActiGraph detected 67% of the activPAL BREAKS while 65% of the ActiGraph-measured BREAKS did not correspond with activPAL BREAKS. Fifty-two percent of those non-corresponding ActiGraph-determined BREAKS, occurred when the activPAL indicated that the participant was sitting, 42% when standing, and 6% when transitioning from standing to sitting. CONCLUSION: The ActiGraph detected a significantly higher number of BREAKS when compared to the activPAL and 65% of its detected BREAKS did not correspond to posture-based definitions of BREAKS in free-living. For the most part, ActiGraph BREAKS do not correspond to sitstand transitions.

10.1080/07448481.2021.1923505

Barreira T, Schuna, J, Zderic T, Hamilton M, Tudor-Locke, C ActiGraph-Measured Breaks in Sedentary Behavior; Are They Real Transitions From Sitting to Standing? International Journal of Exercise Science: Conference Proceedings: 9 (3), Article 9 2015. http://digitalcommons.wku.edu/ijesab/vol9/iss3/9/.