Citation

Phillips LJ, Petroski GF, Markis NE. Res. Gerontol. Nurs. 2015; 8(5): 213-219.

Copyright

(Copyright © 2015, Healio)

DOI

10.3928/19404921-20150429-03

PMID

25942386

Abstract

Older adults' gait disorders present challenges for accurate activity monitoring. The current study compared the accuracy of accelerometer-detected to hand-tallied steps in 50 residential care/assisted living residents. Participants completed two walking trials wearing a Fitbit(®) Tracker and waist-, wrist-, and ankle-mounted Actigraph GT1M. Agreement between accelerometer and observed counts was calculated using concordance correlation coefficients (CCC), accelerometer to observed count ratios, accelerometer and observed count differences, and Bland-Altman plots. Classification and Regression Tree analysis identified minimum gait speed thresholds to achieve accelerometer accuracy ≥0.80. Participants' mean age was 84.2 and gait speed was 0.64 m/s. All accelerometers underestimated true steps. Only the ankle-mounted GT1M demonstrated positive agreement with observed counts (CCC = 0.205). Thresholds for 0.80 accuracy were gait speeds ≥0.56 m/s for the Fitbit and gait speeds ≥0.71 m/s for the ankle-mounted GT1M. Gait speed and accelerometer placement affected activity monitor accuracy in older adults. [Res Gerontol Nurs. 20XX; X(XX):XX-XX.].

Language: en