• L E Barrett, S J Cano, J P Zajicek, and J C Hobart.
• Clinical Neurology Research Group, Plymouth University Peninsula Schools of Medicine and Dentistry, Devon, UK.
• Mult. Scler. 2013 May 1;19(6):806-15.

BackgroundHand dysfunction is common in multiple sclerosis (MS). Recent interest has focused on incorporating patient-reported outcome (PRO) instruments into clinical trials. Nevertheless, examinations are rare in MS of existing manual ability measures.ObjectivesThe objective of this paper is to evaluate the 23-item ABILHAND, developed for use after stroke, in people with MS, comparing the findings from two psychometric approaches.MethodsWe analysed ABILHAND data from 300 people with MS using: 1) traditional psychometric methods (data completeness, scaling assumptions, reliability, internal and external construct validity); and 2) Rasch measurement methods (including targeting, item response category ordering, data fit to the Rasch model, spread of item locations, item scoring bias, item stability, reliability, person response validity).ResultsTraditional psychometric methods implied ABILHAND was reliable and valid in this sample. Rasch measurement methods supported this finding. The three-category scoring function worked as intended and item fit to Rasch model expectations was acceptable. The 23 items (location range -3.16 to +2.73 logits) mapped a continuum of manual ability. Reliability was high (Person Separation Index (PSI) = 0.95).ConclusionBoth psychometric evaluations supported ABILHAND as a robust manual ability PRO measure for MS. Rasch measurement methods were more informative and, consistent with its role of detecting anomalies, identified ways of advancing further ABILHAND's measurement performance to reduce any potential for type II errors in clinical trials.

23 keywords...

hide…