• Louise Barrett, Stefan Cano, John Zajicek, and Jeremy Hobart.
• Clinical Neurology Research Group, Plymouth University Peninsula Schools of Medicine and Dentistry, UK.
• Mult. Scler. 2015 Apr 1;21(5):612-21.

BackgroundOur examination in multiple sclerosis (MS) of the ABILHAND, a patient-reported outcome (PRO) instrument measuring manual ability, identified limited measurement range and precision. These deficiencies could lead to type II errors in clinical trials.ObjectivesThis paper aims to determine if ABILHAND's measurement performance in MS can be improved by adding relevant items from the Disabilities of the Arm, Shoulder and Hand scale (DASH).MethodsThe 23-item ABILHAND and 30-item DASH were administered to 461 people with MS. Data from the ABILHAND were combined with 16 DASH items to create a 39-item scale (AD-39). Using Rasch Measurement Theory methods, we compared the psychometric properties of AD-39 with ABILHAND.ResultsData were analysed from 300 people. AD-39 performed robustly as a measure and had greater measurement range, lower floor and ceiling effects, and higher reliability (person separation index 0.97) than ABILHAND. Surprisingly, AD-39 appeared no better than ABILHAND at detecting group differences in self-reported hand function.ConclusionDespite improving some psychometric properties, adding 16 DASH items to the ABILHAND did not improve its measurement performance to the degree expected. Our explanations for this anomaly emphasise the importance of evidence-based, conceptually driven scale modifications guided by hypothesis testing psychometric methods.© The Author(s), 2015.

33 keywords...

hide…