Purpose: To evaluate the psychometric properties of the Danish Keratoconus Outcome Research Questionnaire (KORQ).
Methods: Phase 1: Rasch analysis assessed the original Danish KORQ. Phase 2: Reengineering of the Danish KORQ to optimize the psychometric properties and functioning of the KORQ.
Results: The KORQ was self-administered by 195 patients. Phase 1 displayed trouble with local dependency and poor item fit. In phase 2, local dependency was addressed, and acceptable item fit was obtained for both subscales by collapsing item pairs into superitems (A6+A7, A9+A17, and S6+S9) and deleting 2 items (S4 and S2). After these corrections, the inter-item range was 0.28 to 0.72 for activity limitation scale (AL-S) and 0.14 to 0.54 for symptoms scale (S-S). Ordered thresholds were present, except in items A11, S5, and S10. No differential item functioning was present. Person separation reliability was 0.93 for AL-S and 0.81 for S-S. The person-item maps showed that the target was slightly above the study population. The maximal corneal curvature was the clinical measure of keratoconus with the highest correlation of 0.33 for AL-S and 0.24 for S-S. Evaluating individual questions showed that patients had the most trouble seeing at distance, driving at night, and wearing rigid gas-permeable lenses, whereas the least troublesome areas were walking up/downstairs, doing household tasks, avoiding objects on their path, and doing their job.
Conclusions: The revised Danish KORQ largely fulfilled the assumptions of the Rasch model and displayed satisfactory psychometric properties. The need to revise the Danish KORQ highlights that cross-cultural validation is of key importance when working with patient-reported outcomes.