Related Experiment Video
Updated: Jun 12, 2026
Visualization of Intensity Levels to Reduce the Gap Between Self-Reported and Directly Measured Physical Activity
Published on: March 7, 2019
Measuring physical activity change in broad-reach intervention trials
Marina M Reeves1, Alison L Marshall, Neville Owen
1Cancer Prevention Research Centre, University of Queensland, Brisbane, Australia.
Background:
We compared the responsiveness to change (prepost intervention) of 3 commonly-used self-report measures of physical activity.
Methods:
In a cluster-randomized trial of a telephone-delivered intervention with primary care patients, physical activity was assessed at baseline and 4 months (n = 381) using the 31-item CHAMPS questionnaire; the 6-item Active Australia Questionnaire (AAQ); and, 2 walking for exercise items from the US National Health Interview Survey (USNHIS). Responsiveness to change was calculated for frequency (sessions/week) and duration (MET x minutes/week) of walking and moderate-to-vigorous intensity physical activity.
Results:
The greatest responsiveness for walking frequency was found with the USNHIS (0.45, 95% CI: 0.19, 0.72) and AAQ (0.43, 95% CI: 0.19, 0.67), and for walking duration with the USNHIS (0.27, 95% CI 0.13, 0.41) and CHAMPS (0.24, 95% CI: 0.12, 0.36). For moderate-to-vigorous activity, responsiveness for frequency was slightly higher for the AAQ (0.50, 95% CI: 0.30, 0.69); for duration it was slightly higher for CHAMPS (0.32, 95% CI: 0.17, 0.47).
Conclusions:
In broad-reach trials, brief self-report measures (USNHIS and AAQ) are useful for their comparability to population physical activity estimates and low respondent burden. These measures can be used without a loss in responsiveness to change relative to a more detailed self-report measure (CHAMPS).
