St Bartholomew’s Hospital has been using an artificial intelligence tool to analyse thousands of outpatient letters to ensure that both the patient and the clinical team have taken the correct action.
The AI tool is trained to read the clinical letters and identify key phrases, which it then checks against patients’ electronic health records. It looks for actions such as booking a follow-up appointment, a scan request, or a referral to another specialist.
It has already prevented patients from missing out on key appointments within the congenital heart disease service and the uro-oncology department. Over six months at St Bartholomew’s it had read 1,500 letters and was able to spot 16 ‘high risk’ cases, where action was needed to prevent care delays or serious harm.
This is the first time AI has been used in this way at Barts Health NHS Trust.
The idea was conceived by Shane Cashin, information business partner at St Bartholomew’s. He worked alongside representatives from the trust’s clinical services, Queen Mary University of London and Andrew Houston from the Barts Life Sciences programme.
Cashin said: “If used correctly, AI has the potential to ease workloads for our admin teams, reduce waits for patients and improve safety.
“Whilst only a small sample, early analysis also suggests some ethnic groups are disproportionately affected by missed follow-up appointments, which could indicate a language or some other barrier when it comes to the contents of their letters.”
The team involved with the AI implementation have applied to Barts Charity for additional funding so they can refine and expand the pilot project and explore other AI opportunities.
AI is also being used within the NHS to help reduce the number of did not attends (DNAs). Following a successful pilot at Mid and South Essex NHS Foundation Trust, AI-powered software from Deep Medical has been rolled out to 10 more NHS trusts. The tool predicts the likely missed appointments using algorithms and anonymised data, and saw Mid and South Essex slash its DNA rate by almost a third in six months.