Rasch Validation of the Mini-BESTest in People With Parkinson Disease.

Abstract

Background and purpose: The Mini-BESTest is a balance measure with robust psychometric properties widely used in people with Parkinson disease. The aim of this study was to examine-with advanced psychometric techniques-some key properties of the Mini-BESTest (including unidimensionality, functioning of rating categories, internal construct validity, reliability indexes) in a consecutive sample of individuals with Parkinson disease admitted for balance rehabilitation.

Methods: Confirmatory factor analysis and Rasch analysis (partial credit model) were performed on 193 individual raw scores of the Mini-BESTest items.

Results: Confirmatory factor analysis fit indices and principal component analysis of the residuals confirmed the scale's unidimensionality. At Rasch analysis, the 3-level rating scale demonstrated appropriate functioning. All items fitted the Rasch model. Item response dependence was negligible. No differential item functioning was found across gender and age groups.

Discussion and conclusions: We confirmed and extended the evidence (demonstrated in different populations) on the general psychometric soundness of the Mini-BESTest, even when tested with rigorous statistical methods. In addition, 2 forms were created: (i) to transform raw scores into linear estimates of dynamic balance; and (ii) to compare the individual's item responses with those expected by the Rasch model (thus providing an aid for tailored interventions) and manage missing responses. Further independent studies using advanced psychometric techniques are warranted, also in people with balance disturbances of different etiology. Our study further increases the confidence in using the Mini-BESTest to assess dynamic balance in people with Parkinson disease and provides some useful additional clinical aids for interpreting the results and calculating more precise change scores.See the Supplementary Video, available at: http://links.lww.com/JNPT/A384.