A new artificial intelligence (AI) model predicts, up to three years in advance, whether people with diabetes are at risk of developing severe diabetic retinopathy. The research, led by a team of data scientists and clinicians working across the HDR UK INSIGHT health data hub, demonstrates how securely linked, large-scale imaging and health data can be used to train AI tools to improve patient care.

By analysing thousands of retinal images alongside clinical data, the model can identify both those at high risk who may need more frequent monitoring, and those at low risk who can be safely screened less often. This has the potential to make NHS screening programmes more efficient, reduce unnecessary appointments, and help prevent avoidable blindness through earlier intervention.

The work has been supported by the West Midlands Secure Data Environment, which enables safe, privacy-protecting access to high-quality health data for research and innovation. Read the full story on the HDR UK website – New AI model predicts future risk of diabetes-related blindness