Comparison of self-reported and accelerometer-measured physical activity among Canadian youth

Generally, correlation and agreement between self-reported and accelerometer-measured physical activity are low. The objective of this study is to compare estimates of physical activity from a newly developed Canadian questionnaire with measurements by accelerometer among 12- to 17-year-olds.

Physical activity was self-reported by domain (transportation, recreation, school, and occupational/household) as part of the new Physical Activity Youth Questionnaire (PAYQ) in the Canadian Health Measures Survey (CHMS; 2014-2017; n = 975) and the Canadian Community Health Survey (CCHS; 2015-2016; n=7,619). The CHMS also collected moderate-to-vigorous physical activity (MVPA) data using the Actical accelerometer. Descriptive statistics and correlation and agreement analyses were used to compare and contrast self-reported and accelerometer-measured physical activity variables. Linear regression was used to assess the association between physical activity and obesity.

The average daily MVPA measured by accelerometry was 49.7 minutes per day. According to the PAYQ, Canadian youth reported an average of 78.2 minutes of physical activity per day from all domains, including recreation (31.3 minutes per day), transportation (15.5 minutes per day), school (25.8 minutes per day), and occupational/household (5.6 minutes per day). According to accelerometer-measured MVPA, 23.1% of youth met the physical activity guideline. The inclusion of all domains of self-reported physical activity resulted in a higher percentage of youth meeting the physical activity guideline (58.6%) than was the case for the recreation domain only (18.5%) or the sum of the recreation and school domains (34.0%). Overall, self-reported and accelerometer-measured physical activity estimates were poorly correlated (R ⟨ 0.2).

Population-level estimates of physical activity and the percentage of youth meeting the physical activity guideline were well-aligned between the Actical and the PAYQ; however, large differences were evident at the individual level. Therefore, caution should be exercised in using data from these two methods since their values may not be interchangeable.