A Comparison Between Estimated Physical Activity from an Open-Ended Self-Report Questionnaire and the SenseWear Armband in Older Australians

Joseph M Northey, Ben Rattray, Kate Pumpa, Disa Smee, Nicolas Cherbuin, Kaarin Anstey

Research output: Contribution to journalMeeting Abstractpeer-review


Introduction: Differences in self-reported and objective estimates of physical activity (PA) have been reported in the literature (Scheers, Philippaerts, & Lefevre, 2012), however few studies have assessed these differences in adults aged 70 and over. This study compared intensity-specific estimates of self-reported PA and the SenseWear ArmbandTM (SWA) in a sub-group of The Personality and Total Health Through Life project (PATH). Methods: A sub-group (n = 105, 44 females; mean age: 75.3 ± 1.3 years) of the 60–64 year cohort of the PATH study (Anstey et al., 2012) were analysed in this study. Participants were asked in an open-ended format to report the average time per week undertaking PA of light (e.g., walking or housework), moderate (e.g., dancing), and vigorous (e.g., running) intensities (Bielak, Cherbuin, Bunce, & Anstey, 2014). The SWA (BodyMedia, PA, USA) was worn for seven consecutive days, and total time accumulated (SWATotal) and time accumulated in 10-min bouts (SWAG10) were calculated for light (1.5–3.0 METs), moderate (3.0–6.0 METs), and vigorous (> 6 METs) PA. Moderate to vigorous physical activity (MVPA) was calculated by combining the respective domains. Spearman correlations and Bland-Altman were used to test agreement between each estimate. Results: Small significant associations were found between self-reported PA and both the SWAG10 and SWATotal estimates of PA (r = .23–.33). Mean bias and 95% limits of agreement between self-report PA and SWATotal were large, with estimates of PA from self-report lower for light (–821.0 [–2320.2 to 678.2] min·week–1), moderate (–155.0 [–736.7 to 426.1] min·week–1), vigorous (–2.1 [–236.1 to 231.9] min·week–1), and MVPA (–157.5 [–825.6 to 510.7] min·week–1). However, mean bias between self-reported PA and SWAG10 was small for light (–73.7 [–1489.5 to 1341.0] min·week–1), moderate (+39.02 [–475.1 to 553.2] min·week–1), vigorous (+23.9 [–176.5 to 224.2] min·week–1), and MVPA (+16.9 [–576.4 to 610.2] min·week–1). Conclusion: Despite the significant associations between estimates of PA, large absolute differences were found between SWATotal and self-reported estimates of PA. In general, self-reported PA time was lower than SWATotal, but was more comparable to SWAG10. A combination of objective and self-reported measures of PA may provide interesting and novel insights into any link between PA and improved health outcomes. Acknowledgements: This project was funded by the Dementia Collaborative Research Centre – Early Diagnosis and Prevention, as part of an Australian Government Initiative.
Original languageEnglish
Pages (from-to)107-108
Number of pages2
JournalJournal of Aging and Physical Activity
Issue numberS1
Publication statusPublished - 2016
Event9th World Congress on Active Ageing - Melbourne, Australia
Duration: 28 Jun 20161 Jul 2016


Dive into the research topics of 'A Comparison Between Estimated Physical Activity from an Open-Ended Self-Report Questionnaire and the SenseWear Armband in Older Australians'. Together they form a unique fingerprint.

Cite this