The sensitivity of three commonly used outcome measures to detect change amongst patients receiving inpatient rehabilitation following stroke

Objective: To investigate the sensitivity of three commonly used functional outcome measures to detect change over time in subjects receiving inpatient rehabilitation post stroke. Design: Subjects were assessed within one week of admission and one week of discharge from an inpatient rehabilitation facility. Several parameters of sensitivity were calculated, including floor and ceiling effects, the percentage of subjects showing no change and the effect size of the change between admission and discharge. Setting: The medical rehabilitation ward of an inpatient rehabilitation facility. Subjects: Seventy-eight subjects receiving inpatient rehabilitation following a first or recurrent stroke. Measures: Five-metre walk, comfortable pace (gait speed), the Berg Balance Scale and the Motor Assessment Scale. Results: Sixty-one subjects had complete admission and discharge data. Gait speed and the Berg Balance Scale were both sensitive to change and demonstrated large effect sizes. The Motor Assessment Scale item five also showed a large effect size and was able to detect change amongst lower functioning subjects. The other items of the Motor Assessment Scale were less useful, in particular, the effect sizes for upper extremity change scores were small (d = 0.36-0.5) and the majority of subjects (44.3-63.9%) showed no change over time on these measures. Conclusion: Gait speed, the Berg Balance Scale and the Motor Assessment Scale item five were sensitive to change over time in this sample.