Minimal clinically important difference of the Berg Balance Scale and comfortable walking speed in patients with acute stroke: A multicenter, prospective, longitudinal study.

Objective: To determine the minimal clinically important difference between the Berg Balance Scale and comfortable walking speed in acute-phase stroke patients.

Design: Multicenter, prospective, longitudinal study.

Setting: Inpatient acute stroke rehabilitation.

Subjects: Seventy-five patients with acute stroke, mean (SD) age 71.7 (12.2) years.

Intervention: Inpatients with acute stroke were assessed with the Berg Balance Scale and comfortable walking speed before and after rehabilitation. Physiotherapy was conducted to improve balance and gait over a 2-week period: an average of 40 min/day on weekdays and 20 min/day on weekends and holidays.

Main Measures: The patients' Berg Balance Scale, comfortable walking speed, Global Rating of Change scale (patient-rated and physiotherapist-rated), and motor score of the Functional Independence Measure were obtained. Minimal clinically important differences were estimated using both anchor- (receiver operating characteristic curves and change difference) and distribution-based approaches (minimal detectable change and 0.5× the change score [SD]).

Results: The baseline scores were 31.2 (18.9) for the Berg Balance Scale and 0.79 (0.35) m/s for comfortable walking speed. The minimal clinically important difference in the Berg Balance Scale was 6.5-12.5 points by the anchor-based approach and 2.3-4.9 points by the distribution-based approach. The minimal clinically important difference in comfortable walking speed was 0.18-0.25 m/s by the anchor-based and 0.13-0.15 m/s by the distribution-based approach.

Conclusions: A change of 6.5-12.5 points in the Berg Balance Scale and 0.18-0.25 m/s in the comfortable walking speed is required in these measurements' anchor-based minimal clinically important differences to be beyond measurement error, and to be perceptible by both patients and clinicians.