September 2023
Whitaker, Kara M.; Jones, Melissa A.; Boonstra, Daniel E.; Dziewior, Jaclyn; Unke, Mackenzie; Gibbs, Bethany Barone
Medicine & Science in Sports & Exercise
The 24-hour activity cycle includes three interrelated behaviors: sedentary behavior (SB), physical activity (PA), and sleep. To assess these behaviors with high rigor, multiple wearable devices are worn concurrently. Questions remain on how to best integrate data across devices. Purpose: To describe a novel method to integrate data from a research grade sleep monitor to quantify SB and PA more accurately. Also, to compare SB and PA estimates using this novel method with a standard approach using a diary. Methods: In the ongoing Pregnancy 24/7 Cohort Study, participants wear two devices concurrently for seven days of 24-hour wear to assess SB and PA (thigh-worn activPAL3 micro) and sleep (wrist-worn Actiwatch Spectrum Plus). Participants also complete a diary to identify non-wear and sleep times. To quantify SB and PA, epoch data are exported, cleaned, and reduced using an R program that removes diary-reported non-wear time and integrates Actiwatch-determined sleep onset/offset. PA is calculated using MET- and cadence-based data. A sub-sample of participants (n = 20) in the second trimester were used to compare this novel method to a standard approach where non-wear and sleep are removed manually using only diary data to estimate SB. PA is estimated using cadence-based information from activPAL software. Descriptive statistics, t-tests, and intraclass correlation coefficients (ICCs) compare SB and PA data between methods. Results: There were minimal differences between processing methods, with high ICCs observed (all >0.96, see Table). The relative difference between processing methods for total SB time was 0.3% ± 1.3% (p = 0.23, ICC = 0.98) and for total stepping time was 0.0% ± 0.2 (p = 0.82, ICC = 0.99). Conclusions: While some statistically significant differences were observed between processing methods, these were small and likely not clinically relevant. In the absence of sleep monitor data, diary-reported information provides reliable estimates of SB and PA.
10.1249/01.mss.0000982004.56705.9b
Whitaker, Kara M.; Jones, Melissa A.; Boonstra, Daniel E.; Dziewior, Jaclyn; Unke, Mackenzie; Gibbs, Bethany Barone Integrating Actiwatch And Activpal Data To Estimate Physical Activity And Sedentary Behavior In Pregnant Women: 717 Medicine & Science in Sports & Exercise 2023. https://journals.lww.com/10.1249/01.mss.0000982004.56705.9b.