Preventing Diabetic Peripheral Neuropathy Using Technology To Transform Clinical Outcomes

Background: Personalised medicine in diabetes could be greatly enhanced through the use of cutting-edge technology, enabling improved treatment recommendations, particularly for the prevention of disability from long-term complications. Early microvascular changes in the retina strongly indicate deterioration in other microvascular beds.

Early detection of change in retinal blood vessels could assist in reducing the burden of diabetic peripheral neuropathy (DPN) by providing an “early warning system” and an indicator of disease status, enabling improved treatment recommendations.

Aims: (1) To develop a novel artificial intelligence (AI) driven image-based complication risk prediction model (requiring retinal images only) and assess its predictive value in quantifying individual 5-year risk of developing DPN. (2) To map the development of DPN and associated longitudinal risk factors using large-scale primary care electronic health records (EHRs).

Methods: We will use a comprehensive longitudinal cohort from the national community-based Diabetic Eye Screening Service Wales, where retinal images have been annually taken for up to 15-years on 160,000+people. Routine-collected EHR clinical data is available via the Secure-Anonymised-Information-Linkage-(SAIL)-Databank.

Summary: The project has the potential to improve healthcare outcomes for people with diabetes by providing an ‘early warning system’, meeting the sustainability development goal of good health and wellbeing.