Wrist Accelerometer Estimates of Physical Activity Intensity During Walking in Older Adults and People Living With Complex Health Conditions: Retrospective Observational Data Analysis Study

• DOI: 10.2196/41685
• Publication Date: 2023-03-15T13:16:06Z
• Authors: Kyle S Weber; F Elizabeth Godkin; Benjamin F Cornish; William E McIlroy; Karen Van Ooteghem
• Keywords: Cadence

Accurate measurement of daily physical activity (PA) is important as PA linked to health outcomes in older adults and people living with complex conditions. Wrist-worn accelerometers are widely used estimate intensity, including walking, which composes much PA. However, there concern that wrist-derived data these cohorts unreliable due slow gait speed, mobility aid use, disease-related symptoms impact arm movement, transient activities living. Despite the potential for error intensity estimates, their use has become ubiquitous research clinical application.The goals this work were (1) determine accuracy wrist-based estimates during known walking periods cerebrovascular disease (CVD) or neurodegenerative (NDD) (2) explore factors influence estimates.A total 35 (n=23 CVD NDD) wore an accelerometer on dominant wrist ankle 7 10 days continuous monitoring. Stepping was detected using accelerometer. Analyses restricted bouts ≥60 seconds long a cadence ≥80 steps per minute (LONG walks) identify purposeful, likely reflect moderate-intensity activity. Wrist analyzed within LONG walks 15-second epochs, published thresholds applied classify epochs sedentary, light, moderate-to-vigorous (MVPA). Participants stratified into quartiles based percent classified examined differences behavioral demographic traits between top bottom quartiles. A case series performed illustrate behaviors can affect walking.Participants averaged 107.7 (SD 55.8) median 107.3 10.8) minute. Across participants, classification 22.9% 15.8) 27.7% 14.6) 49.3% 25.5) MVPA walks. All participants measured statistically lower proportion than expected (all P<.001), 80% (n=28) had at least 20 minutes time misclassified sedentary estimates. highest quartile significantly (t16=4.24, P<.001) more variable movement (t16=2.13, P=.049) compared those lowest quartile.The current best practice method prone misclassifying multidevice approach may be warranted advance methods accurately assessing groups.