Abstract
Abstract: :
Purpose: To determine the inter–observer and intra–observer repeatability of ocular surface stain grading using the Oxford Grading Scheme and the NEI/Industry Workshop Scale. Methods: Using a standardized set of eight photographs of eyes stained with lissamine green or fluorescein, four observers independently graded the surface staining using the Oxford Scale and the NEI/Industry Workshop Scale in separate sittings. Several weeks later the observers repeated their grading of the same photos. The observers included a broad range of experience levels (cornea attending, cornea fellow, first–year ophthalmology resident, and fourth–year medical student). Statistical analysis of inter–observer repeatability was done using the Fleiss extension of Kappa statistic (Statistica Neerlandica 1975, 29:151–161) as a measure of agreement between multiple independent raters. Statistical analysis of intra–observer repeatability was done with a non–parametric association analysis (Spearman Rank Correlation Coefficient). Results: Our data show good inter–observer repeatability of ocular surface stain grading using the NEI/Industry Workshop Scale (k=0.54, 0.60 (two independent trials) for cornea grading and k=0.60, 0.56 for conjunctiva grading) and the Oxford Grading Scheme (k=0.45, 0.29). Association analysis for intra–observer repeatability over time revealed excellent correlation (correlation measures ranged from 0.804 to 0.994) and these values were comparable between NEI grading and Oxford grading. Conclusions: Our data suggest that ocular surface stain grading using both the NEI/Industry Workshop Scale and the Oxford Grading Scheme is reproducible between observers and consistent for a single observer when evaluating a standardized set of photos, despite a range of experience between observers.
Keywords: cornea: tears/tear film/dry eye