Data Use Application 010

This project aims to improve our capacity to identify patients at risk of end-stage renal disease within the next five years. Currently chronic kidney disease (CKD) has a high prevalence in the UK and it is estimated to affect around 10% of the population, rising to 20% in those above the age of 80. Prior studies from the NHS indicate that between 900,000 and 1.8 million people have undiagnosed CKD in the moderate to severe stages 3-5. CKD can be difficult to diagnose as it requires blood testing to confirm diagnosis. However, patients in the early stages are more likely to be asymptomatic until later stages of the disease.

This research uses machine learning techniques to identify links between routinely captured retinal imaging and a patient’s current kidney health. Computer vision techniques are used to unlock features in these retinal scans that can be used to identify those at risk of end-stage renal disease. The outputs would be robustly compared to the current standard – kidney risk failure equation – to verify that the research methods deployed do not introduce significant bias. The outcome of this project will be a tool that can identify those at risk of renal decline within routine care, improving standards of treatment and quality of life for patients with CKD.