Comparative assessment of ActiGraph data processing techniques for measuring sedentary behavior in adults with COPD

Area of Science:

• Physical activity measurement
• Wearable sensor technology
• Chronic disease management

Background:

• Sedentary behavior (SB) measurement in adults with chronic obstructive pulmonary disease (COPD) lacks standardized ActiGraph processing techniques.
• Accurate SB assessment is crucial for understanding health outcomes in individuals with chronic illnesses.

Purpose of the Study:

• To compare various ActiGraph data processing techniques against activPAL for sedentary time measurement in COPD patients.
• To identify the optimal ActiGraph processing method for sedentary adults with COPD.

Main Methods:

• Secondary analysis of 59 adults (≥50 years) with COPD, wearing ActiGraph GT9X and activPAL3 for 7 days.
• ActiGraph data (vertical axis, vector magnitude) processed using combinations of filters, non-wear durations, and cut-points.
• Bland-Altman analysis assessed concordance between 30 ActiGraph techniques and activPAL-measured sedentary time.

Main Results:

• Moderate to strong agreement (concordance correlation 0.614–0.838) observed between ActiGraph and activPAL across all techniques.
• The optimal technique involved vector magnitude data, low-frequency extension filter, 120-minute non-wear algorithm, and a <40 counts/15s cut-point (concordance correlation 0.838).
• Wide limits of agreement were noted, indicating variability in individual measurements.

Conclusions:

• ActiGraph vector magnitude data with LFE filter and specific processing parameters are recommended for sedentary adults with COPD.
• While optimal techniques exist, other ActiGraph processing methods may be acceptable with appropriate cut-point selection.
• Findings provide guidance for standardized ActiGraph data processing in clinical and research settings for COPD populations.