Glaucoma Risk Underestimated by Ophthalmologists, Trainees

Larry Hand

May 07, 2014

Ophthalmology trainees and comprehensive ophthalmologists who participated in an Internet-based multinational study to assess optic discs were about twice as likely to underestimate as to overestimate the likelihood of glaucoma, according to an article published online April 3 in JAMA Ophthalmology.

Evelyn C. O'Neill, MBBCh, MD, from the Centre for Eye Research Australia, University of Melbourne, East Melbourne, Victoria, and colleagues conducted the study involving training and practicing ophthalmologists from 22 countries who participated in the Glaucomatous Optic Neuropathy Evaluation (GONE) Project between December 1, 2008, and June 30, 2010.

GONE is a self-assessment program that presents 42 monoscopic optic disc photographs chosen from 2500 healthy and glaucomatous optic discs to represent a range of physiological and glaucomatous characteristics. Participants grade 9 topographic features and clinical features associated with the retinal nerve fiber layer. They then assess glaucoma likelihood, using a 4-point scale ranging from unlikely to possible, to probable, to certain. The program has a time limit of 90 seconds for each image.

Trainees and practicing ophthalmologists have learned about GONE through conference presentations, publications, and word of mouth. The current analysis includes participants who self-registered and completed assessment of the 42 discs between December 1, 2008, and June 30, 2010. Of 197 participants, 109 were trainees, 51 were comprehensive ophthalmologists, and 37 were glaucoma subspecialists.

For optic discs that glaucoma subspecialists graded as probable or certain glaucoma, ophthalmology trainees underestimated glaucoma likelihood in a mean of 22.1% of discs compared with comprehensive ophthalmologists, who underestimated the likelihood in a mean of 23.8% of discs.

For optic discs that glaucoma subspecialists graded as possible or unlikely glaucoma, trainees overestimated glaucoma likelihood in a mean of 13.0% of discs compared with a mean of 8.9% discs for comprehensive ophthalmologists.

4 Main Disc Characteristics

The researchers used multivariable linear regression analysis to determine which optic disc characteristics had the greatest effect on assessments. They found 4 characteristics that had the most influence: retinal nerve fiber layer loss, rim loss, hemorrhage, and vertical cup:disc ratio (P < .01).

When participants correctly assessed those 4 characteristics, underestimation of glaucoma likelihood fell to 10.9% for trainees and 3.1% for comprehensive ophthalmologists, and overestimation dropped to 2.9% and 1.6%, respectively.

However, when participants incorrectly assessed all 4 of these characteristics, the mean percentage of discs underestimated for glaucoma likelihood rose to 43.0% for trainees and 42.9% for comprehensive ophthalmologists, with no significant change in overestimation.

"The most important take-home message from this paper is the importance of accurate assessment of the optic nerve head and [retinal nerve fiber layer] in the diagnosis and monitoring of glaucoma. When these are inappropriately assessed, diagnoses may be missed," Dr. O'Neill told Medscape Medical News. "We all know population-based screening for glaucoma is not cost effective or efficient; thus, I would recommend improved awareness and case detection through screening of high-risk individuals and better assessment of the optic nerve head in all patients attending ophthalmic services."

She added, "One surprise result in this study was that small disc size did not increase the underestimation of glaucoma likelihood. This may have been due to sample size. However, it was unexpected."

"This paper outlines specific areas in optic nerve assessment that are associated with the grader making a false diagnosis," co-investigator Jonathan G. Crowston, PhD, MBBS, head of glaucoma research and managing director of the Centre for Eye Research Australia, and senior author, added. "In learning or teaching optic nerve assessment, we can focus on specific aspects of the examination that deserve special attention to avoid making a false diagnosis."

Underdiagnosis Is "Biggest Surprise"

"The results of the study weren't a huge surprise, in the sense that we already knew there was a lot of variability among different people who evaluate the optic nerve," Brian Song, MD, a glaucoma specialist at the Massachusetts Eye and Ear Infirmary and instructor in ophthalmology at Harvard Medical School, Boston, told Medscape Medical News.

"The biggest surprise is [that] when an ophthalmology trainee or comprehensive ophthalmologist misinterprets these 4 key features, the rate of underestimating goes up to as high as about 43%. If anything, I would've expected that trainees and comprehensive ophthalmologists might err on the side of overdiagnosing glaucoma (because they don't want to miss someone who has early disease) and referring that patient to a glaucoma specialist...just to play it safe."

He continued, "The take-home message is we still have a long way to go as far as diagnosing glaucoma accurately. The study points out that even among glaucoma specialists, there can sometimes be a significant variability in terms of their interpretation of what someone's optic nerve looks like and what their glaucoma risk is as a result of that evaluation.

"There's no good biomarker for diagnosing glaucoma. There's no single finding that we can use to say that this is definitely glaucoma or this is not glaucoma, which is different from something like high cholesterol. There are a bunch of different findings and characteristics that you try and piece together. Some of those pieces are bigger than others.

"If you interpret one of those features differently than one of your colleagues, then you may come to 2 different conclusions."

Going Forward

Projects like GONE and similar training modules can be helpful going forward in terms of training and continuing medical education, Dr. Song said, as well as integrating information technology and using modern technology to access large databases of normal vs glaucoma optic nerves.

Among the potential weaknesses of the study is the use of monoscopic disc images, which precludes any depth perception, the researchers write.

They conclude, however: "The results of this study will facilitate creation of targeted teaching tools for ophthalmology trainees and continual medical education modules for comprehensive ophthalmologists, focusing on specific aspects of disc examination. We recommend that the assessment parameters be highlighted in optic nerve head examination teaching and continuing medical education modules."

"Since this study we have implemented a training session for ophthalmic trainees in our unit, highlighting a systematic approach to optic disc examination," Dr. O'Neill said. "We are also in the process of creating an online interactive teaching system, TutorGONE, which will highlight difficulties individuals have in disc assessment and provide targeted feedback and teaching to them. Hopefully, both of these will improve disc assessment both locally and on a wider scale."

This research has been supported by the Victorian Government, Allergan, the Helen McPherson Trust Grant, the National Health and Medical Research Council Practitioner Fellowship, and the Dorothy Adele Edols Charitable Trust. The authors and Dr. Song have disclosed no relevant financial relationships.

JAMA Ophthalmol. Published online April 3, 2014. Abstract


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.