Abstract des Autors

Declining physical function is common among aging adults; however, physical activity may decrease the rate of functional loss among adults with or at risk of developing osteoarthritis. Previous studies on this topic used subjective measures of physical activity and were subject to higher selection bias. This study employed propensity score matching techniques to reduce selection bias when measuring the association between objectively measured physical activity and future physical function. Participants (n = 687) of the Osteoarthritis Initiative were included in either a physically active group (n = 280) or a physically inactive group (n = 407) according to baseline accelerometer data. Participants’ performance on functional tests 4 years later was compared using both ordinary least squares linear regression and average treatment effects from propensity score matching. Analysis revealed physical activity to be associated with significantly better performance on a 400-m walk test, but not 20-m gait speed or chair stand tests.