Estimating leisure-time physical activity energy expenditure in the Canadian population: a comparison of 2 methods

The purpose of this study was to determine the impact of 2 different calculation methods for total leisure-time physical activity energy expenditure (LTPAEE) and LTPAEE from different intensity activities on the classification of level of physical activity in the population. Nationally representative cross-sectional data from the Canadian Community Health Survey Cycle 4.1 (2007) were used for this study (N = 64, 397). LTPAEE was calculated using a metabolic equivalent (MET) value of 4.0 for all activities in the "other activity" category for method 1 (currently employed by Statistics Canada) and using activity-specific MET values for method 2. The weighted prevalence and 95% confidence intervals of active (> or =3 kcal.kg-1.day-1 (kkd)), moderately active (1.5-2.9 kkd), and inactive (<1.5 kkd) were determined for each method by demographic characteristics. The agreement between the 2 methods was assessed overall, and for light, moderate, and vigorous activities. There was no difference between methods in the proportion classified as active, moderately active, or inactive for any of the subgroups studied and there was no difference in the distribution or mean LTPAEE between methods. However, assessment of the agreement showed a large number of outliers and a tendency to underestimate LTPAEE from light and vigorous activities while overestimating LTPAEE from moderate activities at the individual level. The results of this study should be considered when performing inferential statistics on the relationship between physical activity and health outcomes.