Background: In multiple sclerosis (MS), the impact of spasticity on the patient's life is a key issue, and it is fundamental that existing tools measuring the patient's perspective undergo psychometric analysis and refinement to optimize confidence in their use in clinical practice and research.
Objective: We examined-by Rasch analysis (RA)-the main metric characteristics of the 88-item Multiple Sclerosis Spasticity Scale (MSSS-88) to: (i) further validate its Italian version (MSSS-88-IT), previously validated through classical test theory methods only and (ii) independently verify the measurement properties of the original scale.
Methods: MSSS-88 data from a convenience sample of 232 subjects with MS underwent RA, mainly examining item fit, reliability indices, test information function, dimensionality, local item independence, and differential item functioning (DIF).
Results: Most items fitted the Rasch model, but 13/88 items showed a misfit in infit and/or outfit values. Rasch reliability indices were high (> 0.80). Test information functions in most subscales showed a sharp decrease in measurement precision as the ability level departs from the quite limited central range of maximal information. The unidimensionality of each subscale was confirmed. Thirteen item pairs showed local dependency (residual correlations > 0.30) and three items presented DIF.
Conclusion: Reliability, dimensionality and some internal construct validity characteristics of the MSSS-88-IT were confirmed. But, drawbacks of the original MSSS-88 emerged related to some item misfit, redundancy, or malfunctioning. Thus, further large independent studies are recommended, to verify the robustness of previous findings and examine the appropriateness of a few targeted item replacements.
Keywords: Multiple sclerosis; Patient-reported outcome measure; Psychometrics; Rasch model; Spasticity.