AI technology at Barts Health spares at-risk diabetic patients from limb amputation

Clinicians at Barts Health NHS Trust, one of the largest trusts in the UK, say it would have taken a clinician over 100 years to review the volume of data analysed in a project which, thanks to AI technology, took just weeks to complete from start to finish.

The trust deployed new AI technology, supplied by Clinithink, to scan 14.2 million documents to find patients with diabetic foot disease, a potentially serious complication of diabetes.

The software trawled through medical records and notes to find 30% more patients with diabetes and 375% more patients with diabetic foot problems, making it easier for clinicians to schedule earlier treatments to save feet and limbs from amputation.

“Attempting this scale of analysis manually would have been frankly impossible,” said Dr Charles Gutteridge, Chief Clinical Information Officer at Barts Health. “Theoretically it would have taken one clinician over a hundred years to review that volume of documents. So not only does AI technology help us find patients who we couldn’t otherwise find, it also saves precious clinical time. This is a first and most important step in being able to treat many patients earlier than would have been possible using a manual process to find them and preventing the serious complications that may result in amputation.”

In the next phase of work planned by the Barts Health team, the characteristics extracted by the software in the cohort it identified, along with input from other sources, will be used to determine whether this approach can predict which patients are most likely to develop the severe complications associated with Diabetic Foot Disease (DFD).

Mr Sandip Sarkar, Consultant Vascular Surgeon at Barts Health and lead clinician for the project, added: “Using this advanced AI technology, we are very excited about the possibility of being able to predict which patients are most likely to experience the worst consequences of DFD.

“This will enable us to focus our precious clinical resource on those patients likely to benefit most from early intervention, which will also reduce the burden on hard-pressed acute services. This is how we need to manage chronic disease in the post-Covid era.”

Diabetes is the leading cause of non-traumatic limb amputation in the UK with analysis by the charity Diabetes reporting more than 176 leg, toe or foot amputations every week in England. Patients with diabetes are also at increased risk of death and that risk increases with the severity of foot ulceration. The estimated cost to the NHS of caring for patients with DFD was £1 billion in 2014/15.

Barts Health, which delivers healthcare to a diverse population of 2.5 million people in East London, was able to use this technology - which harnesses natural language processing - to analyse 14.2 million documents, related to 775,217 patients written between 2018 and 2020. This identified 61,756 patients with diabetes and, of these, 3,119 patients with DFD.

“It is very exciting to see how organisations like Barts Health caring for large patient populations with complex and diverse health needs can leverage our technology to help patients and deliver value,” said Chris Tackaberry, CEO and Co-founder at Clinithink.

“We have always seen a role for AI technology from a population health perspective in identifying and predicting at-risk groups. It is also very encouraging to see the saving in clinical time and effort that results from using our technology to do the heavy lifting that would otherwise have to be done manually, or not at all. We look forward to continuing our collaboration with Barts Health in the next phase of the project.”