Development of a new measure to assess trunk impairment after stroke (trunk impairment scale): its psychometric properties.

Abstract

The purpose of this study was to investigate reliability, validity, internal structure, and responsiveness of our newly developed Trunk Impairment Scale (TIS) for patients with stroke. A total of 73 patients with stroke participated in this prospective study. Interrater reliability (weighted kappa statistics), content validity (principal component analysis), concurrent validity (Spearman's rank correlation with the Trunk Control Test), predictive validity (prediction of discharge FIM scores), and responsiveness (standardized response mean values) were examined. Internal consistency and item difficulties were analyzed with Rasch analysis. The weighted kappa of each TIS item ranged from 0.66 to 1.0. Principal component analysis revealed that the TIS measured a domain similar to the Stroke Impairment Assessment Set trunk items but different from the Stroke Impairment Assessment Set motor and visuospatial items. The TIS correlated with the Trunk Control Test (r = 0.91). To predict discharge FIM motor scores, addition of the TIS as one of the predictors to age, time from onset, and admission FIM score increased the adjusted R2 from 0.66 to 0.75. With Rasch analysis, the misfit was acceptable, except for the abdominal muscle strength item. The difficulty patterns were similar at admission and discharge, except for the abdominal muscle strength item. The responsiveness of the TIS was satisfactory and comparable with that of the Trunk Control Test (standardized response mean values, 0.94 and 1.06). Our newly developed TIS is reliable, valid, and responsive for use in stroke outcome research.