Psychometric validation of the Chronic Ocular Pain Questionnaire (COP-Q).

Abstract

Background: The Chronic Ocular Pain Questionnaire (COP-Q) is a newly developed patient-reported outcome (PRO) measure intended to assess symptoms and impacts associated with Chronic Ocular Surface Pain (COSP). This study assessed the psychometric properties of the COP-Q to determine the adequacy of the COP-Q as a 'fit-for-purpose' instrument to derive trial endpoints for future clinical studies in COSP.

Methods: Patients with COSP completed the COP-Q daily for four weeks on an electronic, touch-screen, tablet device as part of a longitudinal, observational study in the United States (N = 124). Analyses were conducted to assess item properties, dimensionality and scoring, reliability and validity, and interpretation of scores. In addition, 4-hour and 24-hour recall period versions of the COP-Q Symptom Module were compared.

Results: Item responses were distributed across the full response scale for most COP-Q items. Inter-item correlations did not identify any redundant items (r > 0.90) and all items correlated at > 0.40 in their respective module. Confirmatory factor analysis (CFA) provided acceptable support for the unidimensional structure of the multi-item scales in the COP-Q and calculation of a total score for each module. However, CFA and Rasch analysis outlined potential redundant items for the COP-Q Visual Tasking Module (VTM), which were removed, resulting in a six-item VTM. The multi-item COP-Q modules had excellent internal consistency (α range = 0.91-0.96) and suggested fair to excellent test-retest reliability (ICC/Kappa range = 0.651-0.940) for all COP-Q modules. Construct validity for all COP-Q modules was supported by a logical pattern of correlations with concurrent measures and evidence of ability to distinguish between known-groups, with statistically significant differences between COSP severity groups. Paired t-tests, coefficient of determination (CoD) and concordance correlation coefficients (CCC) showed statistically significant differences between the two recall period versions of the Symptom Module, although the magnitude of the difference was small, and each version shares a high level of reproducibility in scores.

Conclusions: Findings provide evidence that the COP-Q is a valid and reliable measure of patient-reported COSP symptoms and impacts for use in future clinical trials in COSP. Both 4-hour and 24-hour Symptom Module recall period versions are likely to yield consistent results and are equally robust.