Beyond the reach of traditional analyses: using Rasch to evaluate the DASH in people with multiple sclerosis.

ABSTRACT

BACKGROUND: Few upper limb functioning patient rating scales have been used in multiple sclerosis (MS) research and none developed specifically for people with MS. OBJECTIVES: In this study, we examined the Disabilities of the Arm, Shoulder and Hand (DASH) to determine its utility as a useful, scientifically robust and clinically meaningful tool in MS. METHODS: DASH data from 300 people with MS underwent two independent phases of psychometric analyses (1) a traditional psychometric analysis (including data quality, scaling assumptions, reliability and validity); and (2) a Rasch analysis (including response option thresholds ordering, tests of fit, spread of item locations, residual correlations, and person separation index). RESULTS: Overall, the traditional psychometric analysis supported the DASH as a reliable and valid measure of upper limb function in people with MS. However, several issues were raised by the Rasch analysis that questioned the validity of the DASH, including misfit in 13/30 items, disordered item response option thresholds for 9/30 items, and six pairs of items with high residual correlations (> 0.60). CONCLUSION: Rasch analysis highlights areas for potential improvement for the use of the DASH. Our findings further support our previous arguments that traditional psychometric methods provide weak scale evaluations and can mislead clinicians as to the reliability and validity of outcome measures.