Abstract
OBJECTIVES: To assess domain-specific effects of physical activity (PA) in the relationship with health care utilization and to investigate whether a measure that aggregates PA across domains (leisure, transport, work) is appropriate. METHODS: Data were retrieved from a longitudinal cohort study conducted in Southern Germany (women n = 1330, men n = 766). The number of physician visits was regressed on total PA and on PA differentiated by the domains leisure time, travel time and working time in a negative binomial model. RESULTS: For women, no association with physician visits is found for total PA, while high leisure time physical activity (LTPA) is associated with 22% more visits. The effect of high LTPA is statistically different from the effect of high total PA. For men, no significant associations are found for both measures. CONCLUSIONS: The specific, positive effect of high LTPA on physician visits among women shows that using an aggregate measure of PA is inappropriate for analyzing the relation between PA and health care utilization. Further, the positive relationship should be considered in attempts to promote physical activity.