Simulating a colour-blind ophthalmologist for diagnosing and staging diabetic retinopathy

In the absence of pre-admission testing for colour blindness, many of the currently practicing ophthalmologists are colour blind, accordingly their accuracy of distinguishing fine diabetic retinopathy (DR) changes is still unknown. This study aims to assess the accuracy of diagnosing and staging diabetic retinopathy and macular oedema among protonopic, deutronopic and tritanopic ophthalmologists.

Cross-sectional assessment of fundus images that were prepared to simulate the appearance in cases of colour blindness. We assessed the accuracy of staging diabetic retinopathy and macular oedema by a retina specialist on colour-blind simulated images. We used randomiser.org to randomly select images to be simulated by the previously validated Vischeck colour blindness simulator.

A total of 150 simulated images were reviewed, 50 images for each of simulated protanopia, deuteranopia and tritanopia. We found that the accuracy for staging DR and macular oedema for protanope grader were 50% and 60%, respectively. Accuracy within one stage difference for DR and macular oedema were 88% and 90%, respectively. For deuteranopes, 56% and 64% accuracy for DR and macular oedema, respectively. Accuracy within one stage difference for DR and macular oedema were 86% and 90%, respectively. For Tritanope, 62% and 84% accuracy for DR and macular oedema, respectively.

Colour vision is important for distinguishing fine details during retina assessment in diabetic retinopathy patients. Colour blindness is associated with low accuracy in staging diabetic retinopathy and macular oedema, particularly among protonopic graders, and to a lesser extent in tritanopic graders.