Measuring Upper Limb Function In Multiple Sclerosis: Enhancing The Abilhand’s Performance


      ASCEND is a phase 3, randomized, double-blind, placebo-controlled trial assessing whether natalizumab slows disability progression in secondary progressive multiple sclerosis (SPMS) patients. The aim of this current analysis was to use Rasch Measurement Theory (RMT) methods to evaluate the ABILHAND (56 item version), a PRO assessing manual ability, in SPMS patients, and to explore an optimized scoring structure based on empirical post-hoc analyses.


      Baseline blinded data from the 889 randomized patients in ASCEND were analyzed. In stage 1, RMT methods examined: scale-to-sample targeting, item fit, local dependency, and reliability. In stage 2, a post-hoc revision of the ABILHAND-56 scoring structure and conceptual grouping of items was conducted and re-evaluated using the same RMT methods.


      Stage 1 analyses showed adequate scale performance: minor item misfit (2/56); minimal dependency (4 pairs of items); good reliability (Person Separation Index = 0.93). However, there was ABILHAND-to-ASCEND sample mis-targeting (person location range [mean]: -7.31 to 1.83 [-3.62]; item location range [mean]: -5.75 to 4.27 [0.00]). In stage 2, all items were rescored on a dichotomous response scale (easy & difficult/impossible) in an attempt to improve targeting. Also, the 56 items were re-categorized into two conceptually clearer manual ability sub-scales: ‘fine motor skills’ and ‘power’. These ABILHAND-56 revisions improved targeting: ‘fine motor skills’ (person range [mean]: -5.39 to 5.34 [-1.75]; item range [mean]: -4.15 to 3.54 [0.00]), ‘power’ (person range [mean]: -4.47 to 4.38 [-1.72]; item range [mean]: -4.15 to 3.54 [0.00]). Sample measurement and item fit were consistent with the original ABILHAND-56.


      The ABILHAND-56 revised scoring demonstrated improved psychometric performance and provides an initial evidence-base for the enhancement of ABILHAND-56’s measurement performance in people with SPMS. Additional research will determine whether the two sub-scale structure of the revised ABILHAND-56 provides better interpretability of patient-reported manual ability.