Using AI to screen for diabetic retinopathy in Rwanda

New research published in the British Journal of Ophthalmology has found that AI led to accurate referrals for eye disease from diabetes clinics

blood sugar test
Pixabay/Tesa Robbins

A new study published in British Journal of Ophthalmology has highlighted the potential of artificial intelligence (AI) in interpreting retinal images for signs of diabetic retinopathy.

Between March and June in 2021, 827 participants at four diabetes clinics in Rwanda were screened for diabetic retinopathy using retinal images with AI interpretation.

Among the study participants, there was a high level of satisfaction with AI screening. More than half of participants (64%) preferred AI to human grading of retinal images.

Compared to human grading, the sensitivity for referable diabetic retinopathy was 92% with a specificity of 85%.

More than half of participants (64%) preferred AI to human grading of retinal images.

“Notably, almost two-thirds of respondents reported a preference for AI over human grading, and the very large majority were satisfied with the AI screening process,” the authors highlighted.

The scientists noted that a barrier to the implementation of screening programmes in low and middle income nations is a low level of ophthalmologists.

“Using non-ophthalmologists trained in capturing retinal images, in tandem with AI-supported interpretation for referrals as this study did, will be crucial to improve access to sight-saving screening services in Africa, the global region poised for the greatest rise in the diabetes burden,” the authors observed.