Comparison of the structural validity of three balance evaluation systems test in older adults with femoral or vertebral fracture

Abstract

Objective: To clarify and compare the structural validity of 3 Balance Evaluation Systems Tests (BESTest, Mini-BESTest, and Brief-BESTest) in older adults with femoral or vertebral fractures. Design: Cross-sectional study. Subjects: Ninety-four older adults (age ≥65 years) with femoral or vertebral fractures, who could walk without physical assistance. Methods: Four BESTest models (BESTest, one-factor Mini-BESTest, four-factor Mini-BESTest, and Brief- BESTest) were examined using confirmatory factor analysis, and the models' goodness-of-fit was assessed. Unidimensionality of the best-fitting model was confirmed by Rasch principal component analysis on the residuals. Results: Confirmatory factor analysis showed that the four-factor Mini-BESTest model (comparative fit index = 0.952; Tucker-Lewis index = 0.937; root-mean square error of approximation = 0.060; standardized root-mean-square residual = 0.062) has a better structure than other models. The principal component analysis of standardized residuals showed that the variance attributable to Rasch factor was good, with eigenvalues < 2, confirming the factor's unidimensionality. Conclusion: The four-factor Mini-BESTest model shows good structural validity in older adults with femoral or vertebral fracture. Evaluating dynamic balance by focusing on 4 components (anticipatory postural adjustments, postural responses, sensory orientation, and stability in gait) may help therapists in making clinical decisions.