The Development of a Cataract-specific Item Bank to Measure Patient-reported Outcomes

Abstract

Objective: To develop a Rasch-based item bank for the assessment of Cataract-specific patient-reported outcomes (PROs) that can be evaluated precisely, comprehensively and dynamically. Methods: In this cross-sectional study, patients diagnosed with cataract in the Eye Hospital of Wenzhou Medical University were enrolled. Cataract patients answered an item pool of cataract PROs that was established by following the guidelines of the international PRO development process. Rasch analysis was used to test the category probability curve (CPC), measure precision, infit/outfit MNSQ, unidimensionality, targeting and differential item functioning (DIF). The items were refined until reliable characteristics were developed for the item bank. A Pearson correlation analysis was used to test each dimension for its correlation with visual acuity. Results: A total of 296 patients (130 males) were recruited for this study. Of those, 60.1% were illiterate with a median age of 70 years, an interquartile range of 63 to 77 years, and a preoperative binocular visual acuity (LogMAR) range from 0.0 to 3.0 (median 0.6). The conceptual framework of the item bank consisted of three domains. The Rasch analysis led to the following item banks: Vision-related activity (23 items), Visual (14 items) and Emotional (11 items). All items had four response options. The vision-related activity limitation domain had a category threshold order of -1.84, 0.04, 1.79, a person separation reliability/person separation index (PSR/PSI) of 0.93/3.59, an eigenvalue of principal component analysis of 2.61, and targeting of -0.81 none of the items had a DIF. This domain had a moderate correlation with better visual acuity (r=0.619, P<0.01) and binocular visual acuity (r=0.622, P<0.01). The vision symptoms domain had a category threshold order of -1.83, -0.18, 2.00, a PSR/PSI of 0.88/2.68, an eigenvalue of principal components analysis of 2.53, and targeting of -1.35. Two items showing a DIF of ocular comorbidities. This domain had a moderate correlation with better visual acuity (r=0.482, P<0.01) and binocular visual acuity (r=0.492, P<0.01). The emotional well-being domain had a category threshold order of -2.83, 0.03, 2.80, a PSR/PSI of 0.88/2.68, an eigenvalue of principal components analysis of 2.79, targeting of -0.34. None of the items had a DIF. This domain showed moderate correlation with better visual acuity (r=0.425, P<0.01) and binocular visual acuity (r=0.426, P<0.01). All items were trimmed to meet the goodness of fit statistic criteria of 0.50-1.50. Conclusions: This study, which is based on the Rasch analysis, developed a cataract item bank that covers all three common concern domains of cataract patients. It is a comprehensive assessment denoting the PROs of cataract patients and is easy to understand and administer. It not only shows perfect psychometric properties but is also appropriate for measuring PROs in Chinese cataract patients. Key words: cataract; patient-reported outcomes; item bank; Rasch model