Interpretability of Change Scores in Measures of Balance in People With COPD.

Abstract

Balance deficits and an increased fall risk are well documented in individuals with COPD. Despite evidence that balance training programs can improve performance on clinical balance tests, their minimal clinically important difference (MCID) is unknown. The aim of this study was to determine the MCID of the Berg Balance Scale (BBS), Balance Evaluation Systems Test (BESTest), and Activities-Specific Balance Confidence (ABC) scale in patients with COPD undergoing pulmonary rehabilitation. We performed a secondary analysis of data from two studies of balance training in COPD (n = 55). The MCID for each balance measure was estimated using the following anchor and distribution-based approaches: (1) mean change scores on a patient-reported global change in balance scale, (2) optimal cut-point from receiver operating characteristic curves (ROCs), and (3) the minimal detectable change with 95% confidence (MDC95). Data from 55 patients with COPD (mean age, 71.2 ± 7.1 y; mean FEV1, 39.2 ± 15.8% predicted) were used in the analysis. The smallest estimate of MCID was from the ROC method. Anchor-based estimates of the MCID ranged from 3.5 to 7.1 for the BBS, 10.2 to 17.4 for the BESTest, and 14.2 to 18.5 for the ABC scale; their MDC95 values were 5.0, 13.1, and 18.9, respectively. Among patients with COPD undergoing pulmonary rehabilitation, a change of 5 to 7 points for the BBS, 13 to 17 points for the BESTest, and 19 points for the ABC scale is required to be both perceptible to patients and beyond measurement error.