Physical activity assessment for epidemiologic research: the utility of two simplified approaches

We assessed the utility of two simplified approaches for the assessment of physical activity in a random sample of 348 college students (218 white, 130 black), mean age 19.3 years. Self-assessment of physical activity level was obtained from the response (4-point scale) to the question "Compared with others your age and sex would you consider yourself to be" (I) much more active to (IV) somewhat less active. In addition, the frequency of exercise-induced sweating (days/week) was assessed. Responses were compared with measurements of resting heart rate, triceps skinfolds, and physical activity as assessed by the Harvard Alumni Survey. Results for the self-assessment question indicated significantly lower resting heart rates (beats/min) (73.0 vs 64.6, P less than 0.01), triceps skinfolds (mm) (10.9 vs 20.4, P less than 0.001), and higher Harvard Survey scores (kcal/week) (5,654 vs 1,310, P less than 0.001) for those responding (i) compared with (iv). Similar results were noted for the sweat-episode question. Those reporting 5-7 sweat episodes per week had significantly lower resting heart rates (67.0 vs 74.8, P less than 0.01), triceps skinfolds (14.7 vs 17.3, P less than 0.01), and higher Harvard Survey scores (5,717 vs 1,453, P less than 0.001) than those reporting 0-1 sweat episodes per week. The results suggest that these simplified approaches may provide useful indices of physical activity for epidemiologic research and warrant further investigation in other populations.