Abstract
OBJECTIVES: From among physiological attributes commonly targeted in rehabilitation, to identify those in which changes led to clinically meaningful differences (CMDs) in mobility outcomes. DESIGN: Secondary analysis of data collected for a randomized controlled trial of exercise using binary outcomes defined by recording a large CMD (Short Physical Performance Battery (SPPB)=1 unit; gait speed (GS)=0.1 m/s). Iterative models were performed to evaluate possible confounding between physiological variables and relevant covariates. SETTING: Outpatient rehabilitation centers. PARTICIPANTS: Community-dwelling mobility-limited older adults (n=116) participating in a 16-week randomized controlled trial of two modes of exercise. MEASUREMENTS: Physiological measures included leg power, leg strength, balance as measured according to the Performance-Oriented Mobility Assessment (POMA), and rate pressure product at the maximal stage of an exercise tolerance test. Outcomes included GS and SPPB. Leg power and leg strength were measured using computerized pneumatic strength training equipment and recorded in Watts and Newtons, respectively. RESULTS: Participants were 68% female, had a mean age of 75.2, a mean of 5.5 chronic conditions, and a baseline mean SPPB score of 8.7. After controlling for age, site, group assignment, and baseline outcome values, leg power was the only attribute in which changes were significantly associated with a large CMD in SPPB (odds ratio (OR)=1.48, 95% confidence interval (CI)=1.09-2.02) and GS (OR=1.31, 95% CI=1.01-1.70). CONCLUSION: Improvements in leg power, independent of strength, appear to make an important contribution to clinically meaningful improvements in SPPB and GS.