The Mini-BESTest is a clinical balance test that has shown a high sensitivity in detecting balance impairments in elderly with Parkinson's disease (PD). However, its reproducibility between different raters and between test occasions has yet to be investigated in a clinical context. Moreover, no one has investigated the reproducibility of the Mini-BESTest's subcomponents (i.e. anticipatory postural adjustments; postural responses; sensory orientation and dynamic gait).
We aimed to investigate the inter-rater and test-retest reproducibility (reliability as well as agreement) of the Mini-BESTest, as well as its subcomponents, in elderly with mild to moderate PD, performed under conditions assimilating clinical practice.
This was an observational measurement study with a test-retest design.
Twenty-seven individuals with idiopathic PD (66 - 80 years, mean age: 73; Hoehn & Yahr: 2-3; 1-15 years since diagnosis) were included. Two test administrators, having different experiences with the Mini-BESTest, administered the test individually, in separate rooms in a hospital setting. For the test-retest assessment, all participants returned 7 days after the first test session to perform the Mini-BESTest under similar conditions. Intra-class correlation coefficients (ICC2.1), standard error of measurement (SEMagreement), and smallest real difference (SRD) were analyzed.
The Mini-BESTest showed good reliability for both inter-rater and test-retest reproducibility (ICC = 0.72 and 0.80). Regarding agreement, the measurement error (SRD) was found to be 4.1 points (accounting for 15% of the maximal total score) for inter-rater reproducibility and 3.4 points (12% of the maximal total score) for test-retest reproducibility. The investigation of the Mini-BESTest's subcomponents showed a similar pattern for both inter-rater and test-retest reproducibility, where postural responses had the largest proportional measurement error, and sensory orientation showed the highest agreement.
Our findings indicate that the Mini-BESTest is able to distinguish between individuals with mild to moderate PD; however, when used in clinical balance assessments, the large measurement error needs to be accounted for.