Rasch Analysis of the Korean Version of the Tinnitus Handicap Inventory.

Abstract

Tinnitus is the perception of abnormal sounds in the ears or head without external auditory stimulation. While classical test theory is often used in tinnitus questionnaire development, it has limitations in assessing item characteristics. Item response theory (IRT) offers more precise individual ability estimations and identifies key and less important items, making it superior for reliable measurement tools. This study investigated the suitability of the Korean version of the Tinnitus Handicap Inventory (K-THI) as a patient-reported outcome measure (PROM) for clinical trials. Using Rasch analysis based on IRT, we evaluated K-THI's measurement of tinnitus-related disability in 545 patients (40.4% men, 59.6% women). Five items (2, 7, 8, 19, and 24) did not fit the Rasch model, yet a unidimensional scale and good fit for person and item data emerged (person: 0.89; item: 0.98). The three-point rating scale in K-THI proved suitable. IRT allowed precise evaluation of K-THI's properties, vital for reliable PROMs in patient-centered care. Our findings highlight IRT's role in questionnaire development, contributing to the advancement of PROMs.