Validation and Establishment of an Interval-Level Measure of the Balance Assessment in Sitting and Standing Positions in Patients With Stroke.

Abstract

To validate the psychometric properties of the Balance Assessment in Sitting and Standing Positions, including validity (unidimensionality and concurrent validity), reliability (Rasch reliability), and responsiveness (compared with the Postural Assessment Scale for Stroke Patients [PASS]) and to transform the Balance Assessment in Sitting and Standing Positions from an ordinal-level measure into an interval-level measure. Retrospective cross-sectional study. Medical records from a medical center. Patients with stroke (N=1193). Not applicable. The 4-item Balance Assessment in Sitting and Standing Positions was used, assessing static sitting balance, dynamic sitting balance, static standing balance, and dynamic standing balance. Data of 1193 patients with stroke were included for Rasch analysis. The 4 items of the Balance Assessment in Sitting and Standing Positions constituted a unidimensional construct (infit/outfit mean square, .75-1.05), had good concurrent validity (r=.70-.90), and had sufficient Rasch reliability (.93). The Balance Assessment in Sitting and Standing Positions had large responsiveness (effect size, 1.20; standardized response mean, 1.51) and was comparable with the PASS (effect size, .90; standardized response mean, 1.32). The Balance Assessment in Sitting and Standing Positions has sound psychometric properties. The transformed-Rasch scores of the Balance Assessment in Sitting and Standing Positions can be used to identify patients' balance function and detect patients' changes.