Responsiveness and predictability of gait speed and other disability measures in acute stroke

Objectives: To identify the most responsive method of measuring gait speed, to estimate the responsiveness of other outcome measures, and to determine whether gait speed predicts discharge destination in acute stroke. Design: A prospective cohort study. Setting: Five acute-care hospitals. Patients: Fifty subjects with residual gait deficits after a first-time stroke. Interventions: Five- (5mWT) and 10-meter walk tests (10mWT) at comfortable and maximum speeds, with 2 evaluations conducted an average +/- standard deviation (SD) of 8 +/- 3 and 38 +/- 5 days poststroke. Main Outcome Measure: Standardized response mean (SRM = mean change/SD of change) was used to estimate responsiveness for each walk test, the Berg Balance Scale, the Barthel Index, the Stroke Rehabilitation Assessment of Movement (STREAM), and the Timed Up and Go (TUG). Results: The SRMs were 1.22 and 1.00 for the 5mWT, and .92 and .83 for the 10mWT performed at a comfortable and maximum pace, respectively. The SRMs for the Berg Balance Scale, the Barthel Index, the STREAM, and the TUG were 1.04, .99, .89, and .73, respectively. The probability of discharge to a rehabilitation center for persons walking at less than or equal to 0.3m/s or > 0.6m/s at the first evaluation was .95 and .22, respectively. Conclusions: The 5mWT at a comfortable pace is recommended as the measure of choice for clinicians and researchers who need to detect longitudinal change in walking disability in the first 5 weeks poststroke.