Is the Berg Balance Scale an internally valid and reliable measure of balance across different etiologies in neurorehabilitation? A revisited Rasch analysis study.

Abstract

To assess, within the context of Rasch analysis, (1) the internal validity and reliability of the Berg Balance Scale (BBS) in a sample of rehabilitation patients with varied balance abilities; and (2) the comparability of the BBS measures across different neurologic diseases. Observational prospective study. Rehabilitation ward of an Italian district hospital. Consecutively admitted inpatients and outpatients (N=217); for 85 participants, data were collected both on admission and discharge, giving a total sample of 302 observations. Not applicable. BBS. Most of the BBS items had to be rescored, and 2 items (static sitting and standing balance) had to be deleted, to attain adequate internal construct validity (χ(2)(24)=35.68; P=.059). The reliability of the Rasch-modified BBS (BBS-12) (total score, 0-35) was high (.957), indicating precision of measurement at the individual level. The analysis of differential item functioning (DIF) showed invariance of the item calibrations across patients' sex, age, and etiology. After adjusting for the possible effect of repeated measurements on person estimates, the analysis of DIF by timing of assessment confirmed the stability of the item hierarchy across time. A practical ruler was provided to convert item raw scores into Rasch estimates of balance ability. This study supports the internal validity and reliability of the BBS-12 as a measurement tool independent of the etiology of the neurologic disease causing the balance impairment. In view of some sample-related issues and that not all possible etiologies encountered in the neurorehabilitation settings were tested, a larger multicenter study is warranted to confirm these findings.