Objectives: The purpose of this study was to determine the number of monitoring days needed to reliably measure weekly activity behaviours with the ActivPAL (AP) monitor in adults.
Design: Cross-sectional study.
Methods: Participants (90 adults (51.1% men); age=39.1±12.43 years) wore an AP for 7 consecutive days. Mean time spent sitting/lying, standing and stepping per day, and mean number of transitions from sitting to standing per day were calculated for each participant using 7 days of monitoring (reference). Estimates for these activities were also derived from a combination of randomly selected days (from 1 to 6 days), and randomly selected weekdays and weekend days, and compared with the reference using ANOVA, correlation coefficients and Bland-Altman methods Spearman-Brown Prophecy Formula, based on Intraclass correlation of 0.8, was used to predict the minimum number of days needed to represent activity behaviours as measured with the AP.
Results: At least five days of monitoring were necessary to achieve a reliability of 0.8 for all postures and transitions. Correlation coefficients between estimates derived from any combination of 5 days and the reference were high (rho=0.96-0.98). When using a combination of weekdays and weekend days, mean biases were comparable with those shown by any combination of days, but 95% limits of agreement were narrower.
Conclusions: When using the AP for a week, data from a combination of any 5 days provided reliable estimates of all activities and transitions per day, but more precise estimates were achieved if at least 1 weekend day was included.
Keywords: Measurement; Physical activity; Reliability; Sedentary behaviour; Sitting.
Copyright © 2018 Sports Medicine Australia. Published by Elsevier Ltd. All rights reserved.