Is the Brief-BESTest Brief Enough? Suggested Modifications Based on Structural Validity and Internal Consistency

Background. The Brief Balance Evaluation Systems Test (Brief-BESTest) could be a useful tool for balance assessment. Although some psychometric characteristics have been examined, others still need to be clarified. Objective. The objective was to assess the structural validity, convergent validity, discriminant validity, and internal consistency of the Brief-BESTest in neurological patients. Design. This was a cross-sectional study. Methods. Data were from 416 patients with neurological disease and related balance disorders. Patients were assessed with the 5-levels Activities-Specific Balance Confidence Scale (ABC 5-levels), Brief-BESTest, and some simple balance tests (ie, 1-leg stance, Timed Up & Go test, functional reach, and a fall history questionnaire). Three Brief-BESTest models were examined through confirmatory factor analysis, and the following indexes were calculated: Comparative Fit Index, Tucker-Lewis Index, and root-mean-square error of approximation. Convergent validity was assessed by calculating the correlation between Brief-BESTest and ABC 5-levels total scores. Receiver operating characteristics assessed the ability of each model to differentiate between people with falls and those without falls. Internal consistency was measured with Cronbach alpha and coefficient omega. Results. Confirmatory factor analysis showed model 3 (Comparative Fit Index = 0.97; Tucker-Lewis Index = 0.95; root-mean-square error of approximation = 0.05), with item 1 removed and error covariance between items 3 and 4 and between items 5 and 6, to have a significantly better structure than models 1 and 2. The correlation between Brief-BESTest and ABC 5-levels was 0.61 (Spearman rho) for all 3 models. The area under the curve of the receiver operating characteristics showed an acceptable accuracy (0.72) in distinguishing patients with a history of falls from those without a history of falls (95% confidence interval = 0.66-0.78) for all models and was superior to the areas under the curve of other simple balance tests (1-leg stance, Timed Up & Go test, functional reach). Cronbach alpha was good for Brief-BESTest models 1 (0.92) and 3 (0.92), but omega was greater than 0.80 only for model 3. Limitations. The sample was heterogeneous. Conclusions. The Brief-BESTest, after some changes, shows good validity and internal consistency in patients affected by different balance disorders.