A standardized grading system for scleritis

Abstract

Objective: This study evaluated the performance of a standardized grading system for scleritis using standard digital photographs.

Design: Cross-sectional interobserver agreement study.

Participants: Photo archives from the National Eye Institute.

Methods: Three uveitis specialists from 3 different centers graded 79 randomly arranged images of the sclera with various degrees of inflammation. Grading was done using standard screen resolution (1024×768 pixels) on a 0 to 4+ scale in 2 sessions: (1) without using reference photographs and (2) with reference to a set of standard photographs (proposed grading system). The graders were masked to the order of images, and the order of images was randomized. Interobserver agreement in grading the severity of inflammation with and without the use of grading system was evaluated.

Main outcome measures: Interobserver agreement.

Results: The proposed grading system for assessing activity in scleritis demonstrated a good interobserver agreement. Interobserver agreement (pooled κ) was poor (0.289) without photographic guidance and improved substantially when the "grading system" with standardized photographs was used (κ = 0.603).

Conclusions: This system of standardized images for scleritis grading provides significantly more consistent grading of scleral inflammation in this study and has clear applications in clinical settings and clinical research.

Conflict of interest statement

No authors have any financial/conflicting interests to disclose except Dr Eric Suhler who receives financial support from Celgene, Abbott, Genentech, Novartis, the Department of Veterans' Affairs, and institutional support from Research to Prevent Blindness.

Copyright © 2011 American Academy of Ophthalmology. Published by Elsevier Inc. All rights reserved.

Figures

Figure 1. Scleritis grading system

Standardized digital photos of scleritis of varying severity is illustrated. Graders used this illustration as reference photos in session 2.

Figure 2. Comparing level of agreement for sessions 1 and 2

Zero represents no agreement among 3 graders, 2 represents 2 of the 3 graders agreeing and 3 represents full agreement by all 3 graders.

Figure 2. Comparing level of agreement for sessions 1 and 2

Zero represents no agreement among 3 graders, 2 represents 2 of the 3 graders agreeing and 3 represents full agreement by all 3 graders.