A validation study of the use of smartphones and wrist-worn ActiGraphs to measure physical activity at different levels of intensity and step rates in older people

Rick Yiu Cho KWAN, Justina Yat Wa LIU, Deborah LEE, Choi Yeung Andy TSE

Research output: Contribution to journalArticlespeer-review

Abstract

Background: Physical activity promotes healthy ageing in older people. Accurate measurement of physical activity in free-living environment is important in evaluating physical activity interventions. Research question: What is the criterion validity of 1)an ActiGraph to identify physical activity at different intensity levels and 2)an ActiGraph and a smartphone to measure step rate?
Methods: Community-dwelling older people aged≥60 were recruited. The index tests were using ActiGraph worn in different positions (i.e., both wrists and hip) to measure physical activity intensity and step rate and using smartphone (i.e., Samsung J2 pro and Google Fit) worn in different positions (i.e., trousers pocket and waist pouch) to measure the step rate. The reference standards were using indirect calorimetry (i.e., CosMedK4b 2) to measure physical activity intensity and using direct observation for step rate. Subjects were exposed in different physical activity intensity levels (i.e., sedentary:MET < 1.5,light: MET = 1.5–2.99, moderate:MET = 3.0–6.0, vigorous:MET>6) and step rates through walking on a treadmill at different speeds (i.e.,2−8 km) for approximately 30 min. Spearman’s rho, ROC analysis, and percentage error were employed to report the criterion validity. Results:31 participants completed the tests. ActiGraphs worn in different body positions could significantly differentiate physical activity intensity at the levels of “light- or-above” (VM cut-off = 279.5–1959.1,AUC = 0.932−0.954), “moderate-or-above” (VM cut- off = 1051.0–4212.9,AUC = 0.918−0.932), and “vigorous” (VM cut-off = 3335.4–5093.0, AUC = 0.890−0.907) well with different cut-off points identified. The step rate measured by direct observation correlated significantly with ActiGraph and smartphone (rho = 0.415−0.791). Both ActiGraph and smartphone at different positions generally underestimated the step rate (%error= -20.5,-30.3).
Significance: A wrist-worn ActiGraph can accurately identify different physical activity intensity levels in older people, but lower cut-off points in older people should be adopted. To measure step rate, a hip-mounted ActiGraph is preferable than a wrist- worn one. A smartphone employing Google Fit generally underestimates step rate but it gives a relatively more accurate estimation of step rate when the older people walk at a speed of 4−8 km/h. Copyright © 2020 Elsevier B.V. All rights reserved.
Original languageEnglish
Pages (from-to)306-312
JournalGait & Posture
Volume82
Early online date28 Sep 2020
DOIs
Publication statusPublished - Oct 2020

Citation

Kwan, R. Y. C., Liu, J. Y. W., Lee, D., Tse, C. Y. A., & Lee, P. H. (2020). A validation study of the use of smartphones and wrist-worn ActiGraphs to measure physical activity at different levels of intensity and step rates in older people. Gait & Posture, 82, 306-312. doi: 10.1016/j.gaitpost.2020.09.022

Keywords

  • ActiGraph
  • Physical activity
  • Step rate
  • Older people
  • Smartphones

Fingerprint Dive into the research topics of 'A validation study of the use of smartphones and wrist-worn ActiGraphs to measure physical activity at different levels of intensity and step rates in older people'. Together they form a unique fingerprint.