The repeatability of self reported physical activity patterns in rural South India

The objective of this study was to develop a physical activity assessment questionnaire for use in poorly educated and low socio-economic rural communities in developing countries and assess the repeatability of the measurements over a one-year duration. Physical Activity Level, as a composite measure of daily physical activity (PAL = estimated 24 hour energy expenditure / estimated basal metabolic rate) was derived from the semi-pictorial, interviewer based questionnaire assessing physical activity over the previous 1 week. PAL of both men and women was unchanged across the different seasons of the year. Correlations between the first physical activity questionnaire and the mean of at least 4 questionnaires administered during the course of one year were generally high (males: PAL 'r' = 0.69, individual domains 'r' between 0.57 and 0.82; females: PAL 'r' = 0.7, individual domains 'r' between 0.41 and 0.76, all p < 0.05). Within- individual % coefficient variations (CV) across 4 physical activity questionnaires were low for sleep duration and PAL (< 14%) but high for the other domains of physical activity (77-100%). The data, at least in this community, suggest that while within-individual variations in specific domains of physical activity may be considerable, the questionnaire that we designed can rank individuals in terms of their PAL or individual domains of physical activity fairly consistently. A single physical activity assessment using this instrument may therefore be adequate to rank individuals.