Simplified Post-stroke Functioning Assessment Based on ICF via Dichotomous Mokken Scale Analysis and Rasch Modeling

• Published in: Frontiers in neurology 2022-04, Vol.13, p.827247-827247
• Main Authors: Feng, Chun, Jiang, Zhong-Li, Sun, Ming-Xue, Lin, Feng
• Format: Article
• Language: English
• Subjects: Disability and Health (ICF); ICF core set; International Classification of Functioning; item response theory; Mokken scale analysis; Neurology; Rasch modeling
• ISSN: 1664-2295; 1664-2295
• DOI: 10.3389/fneur.2022.827247

This study aims to accomplish two tasks for (ICF) application among persons with stroke: (1) to make an ICF tool for measuring personal abilities with simplified assessment operations; (2) to quantitatively evaluate ICF categories for being functioning rather than being disabled. A total of 130 inpatients with stroke convenience sampling were evaluated by the extended comprehensive ICF core set for stroke, modified Rankin scale, and modified Barthel index (MBI). This study investigated the responses to 118 stroke-related ICF items (59 items in b and d domains individually) using Mokken scale analysis followed with Rasch modeling. A Mokken scale with 47 items was extracted from the binary data (1 as no-impairment or mild-impairment and 0 as moderate to complete impairment). A Rasch model with 45 items was derived from the Mokken scale. The conversion chart was available involving the original ordinal scores to Rasch-transformed scores from 0 to 100 (interval scale). Total scores exhibited a high correlation with the personal abilities estimated by the Rasch model. The personal ability also demonstrated a significantly strong correlation with the score of the MBI. Thus, the 45 ICF items were suggested to rate potential functional ability as a single measurement. Based on simple "functioning or disabled" judgment tasks, ICF assessment can be simplified to a questionnaire with answering "yes-or-no" questions for each category. Functioning level for each person and difficulty of being functioning for each category can be estimated by the Rasch model of this questionnaire.