28 October 2025
A new study published in British Journal of Ophthalmology has examined the effectiveness of an artificial intelligence (AI)-based diabetic retinopathy screening service in Australia.
The service was used in two general practice clinics and two endocrinology clinics in Victoria and an Aboriginal medical centre in Western Australia between August 2021 and June 2023.
The AI system was trained on more than 200,000 images from diverse clinical settings. It incorporates an automated fundus camera that provides audio prompts to guide patients for correct positioning and focus.
The camera automatically captures a macula-centred image of each eye with a 50° field of view.
All patients with diabetes aged 18 and older were offered screening. All patients aged 50 and over were also offered screening –regardless of whether they had a diabetes diagnosis.
Among the 863 participants who took part in the trial, the AI system achieved an accuracy of 93% for referable diabetic retinopathy detection.
A patient and care provider feedback survey found that 86% of patients thought the screening system was easy to use, while 85% of care providers thought that the technology improved their ability to conduct eye screening.
The researchers highlighted that the technology has the potential to address gaps in accessibility and early detection.
“With high diagnostic accuracy, the AI system shows promise for early detection of referable diabetic retinopathy in primary care and endocrinology clinics. Positive feedback from both patients and care providers supports its integration into routine clinical practice,” the authors concluded.
Comments (0)
You must be logged in to join the discussion. Log in