advertisement

Editors Selection IGR 23-2

Screening and Detection: Recommendations for Further Evaluation of Screening Hits

Comment by Angelo Tanna on:

104345 Comparison between the Recommendations of Glaucoma Specialists and OCT Report Specialists for Further Ophthalmic Evaluation in a Community-Based Screening Study, Ramachandran R; Joiner DB; Joiner DB; Patel V et al., Ophthalmology. Glaucoma, 2022; 5: 602-613

Find related abstracts

Ramachandran and colleagues studied the agreement between glaucoma specialists and 'OCT report specialists' in the determination of whether to refer participants in a community outreach glaucoma screening program for further evaluation by a glaucoma specialist. The investigators subsequently evaluated the influence of access to a customized OCT report¹ as well as the senior report specialist's (DCH) comments on the glaucoma specialists' determinations regarding the decision to recommend referral.

The investigators analyzed 344 eyes of 196 subjects who represented a subset of participants in a larger screening study in communities in New York City with populations at high risk for undiagnosed glaucoma.² A total of 957 participants were screened in the larger study that was conducted in 2017. It is not entirely clear how the investigators arrived at the final smaller subgroup of subjects.

Three glaucoma specialists had access to the commercial OCT report and clinical data gathered during the screening process, including FDT perimetry. Based on their judgement without pre-set criteria, the physician graders were asked to recommend referral to a glaucoma specialist, an ophthalmologist (for evaluation of non-glaucoma pathology) or an optometrist in two to three years.

The two report specialist graders were masked to clinical and demographic characteristics, and only had access to the commercially generated and customized widefield OCT reports.¹ These graders rated the suspicion of glaucoma from 0‐100% and used a threshold of >50% for glaucoma referral. One of the report specialist graders (DCH) annotated the reports with comments regarding suspected abnormalities. Kappa statistics were used to assess the level of agreement between and among graders on the binary determination of whether a glaucoma referral was recommended or not.

Unanimous agreement was achieved for about half of the study eyes. Agreement among three glaucoma specialists was moderate (Kappa 0.43), with unanimous agreement for 60% of eyes. Agreement between the two report specialists was substantial (Kappa 0.77, representing agreement in 95% of eyes).

Agreement between the senior report specialist and the majority determination of the three glaucoma specialists was fair (Kappa 0.32). Disagreement between the two types of specialists occurred in 91 eyes, with the glaucoma specialists recommending referral in 86. These eyes tended to have elevated IOP and/or cup-to-disc ratios ≥0.5. In phase 2 of the study, glaucoma specialists gained access to the customized OCT report including Dr. Hood's comments. The Kappa level of agreement improved to 0.53 after the glaucoma specialists regraded all eyes for which there was initial disagreement. This increase in Kappa was largely driven by many recommendation changes from referral to a glaucoma specialist to referral to an ophthalmologist.

The authors conclude that "it is possible that OCT reading centers can be leveraged in the design of screening protocols to decrease the numbers of unnecessary specialist referrals while still maintaining a high level of quality comparable to those of more comprehensive screening strategies." Since there was no follow-up of the participants and there are no reference standard diagnoses, the graders' diagnostic accuracy cannot be assessed. Moreover, the improvement in agreement in phase 2 of the study was in large part due to glaucoma specialists changing their referral recommendation from glaucoma specialist to ophthalmologist ‒ this does not decrease the number of specialist referrals. Evidencebased methods for identifying participants at greatest risk for vision loss, such as the OHTS risk calculator, were not employed.

Much work is needed to develop and optimize the methodology and to demonstrate the utility and cost effectiveness of glaucoma screening. This report represents another important piece of that foundation.

• ← Previous
• Next →