Measurement of physical activity to assess health effects in free-living populations.

For physical activity surveys that would identify relations to chronic diseases, quality of life, and longevity, the method of choice remains the questionnaire, especially if it can be standardized and administered in uniform fashion to large populations. A sample questionnaire derived largely from epidemiological experience with the Harvard Alumni Health Study is presented that requests anthropometric estimates; physician-diagnosed diseases by year of onset; contemporary physical activities including walking, stair-climbing, and recreational pursuits; food-frequency data that estimate nutrient values and caloric intakes; and social habits affecting health. The questionnaire presents opportunities for cross-sectional, retrospective, and prospective studies. Personal characteristics, physical activities, and other elements of lifestyle may be used as predictor or outcome variables in testing specific hypotheses. Representative surveys are described that have validated and used questionnaires of various complexities, some complemented by measures of physiological fitness. The epidemiological survey questionnaire, when properly designed and administered, can measure effectively energy intake, energy retention, energy expenditure, physiological fitness, quality of life, and health maintenance.