Abstract
Abstract: :
Purpose: To assess the reliability of a lissamine green grading scale for conjunctival images. Methods: A 20 second video clip of the right eye of 288 contact lens wearers was recorded using a digital slip lamp camera system after instilling 10 µL of lissamine green. Optimal nasal and temporal still images from each patient’s video clip were selected for grading. A masked grader used the Oxford grading scale to grade the temporal and nasal images on two occasions separated by 1 week to determine within–grader reliability. A second masked grader graded each image on one occasion to assess between–grader reliability. Simple kappa statistics and 95% confidence intervals (95% CI) were used to determine the test–retest reliability both within– and between–graders overall and stratified based on the staining severity score from grader 1 at reading 1 (non–diseased = grades 0–2 vs. diseased = grades 3–5). Results:Of the 288 video clips, 284 nasal and 285 temporal images were obtained that were gradable and used in the analyses. The average age of the sample was 31.0 ± 11.3 years and 65% were female. There was substantial within–grader reliability for both the nasal (Kappa = 0.69, 95% CI = 0.63 to 0.75) and the temporal (Kappa = 0.73, 95% CI=0.67 to 0.79) images. There was moderate between–grader reliability for both the nasal (Kappa=0.51, 95% CI=0.44 to 0.58) and the temporal (Kappa=0.51, 95% CI=0.44 to 0.58) images. When staining severity was considered, within–grader reliability was moderate–to–substantial for diseased eyes (Kappanasal = 0.55, 95% CI = 0.42 to 0.68; Kappatemporal = 0.48, 95% CI = 0.30 to 0.66) and non–diseased eyes (Kappanasal = 0.66, 95% CI = 0.56 to 0.75; Kappatemporal = 0.74, 95%CI = 0.67 to 0.81). The kappa values from the temporal region were significantly different (p = 0.01) while the nasal region comparison did not differ (p = 0.19). Between–grader reliability was fair–to–moderate for diseased eyes (Kappanasal = 0.33, 95% CI = 0.21 to 0.44; Kappatemporal = 0.34, 95% CI = 0.18 to 0.50) and non–diseased eyes (Kappanasal = 0.50, 95% CI = 0.40 to 0.60; Kappatemporal = 0.47, 95% CI = 0.38 to 0.56). The kappa values from the nasal region were significantly different (p = 0.03) while the temporal region comparison did not differ (p = 0.17). Conclusions: Within– and between–grader reliability of lissamine green staining appears to be at least substantial–to–moderate, although it appears that the extent of conjunctival staining impacts this reliability.
Keywords: conjunctiva • clinical (human) or epidemiologic studies: biostatistics/epidemiology methodology