Here we explore some recent updates from across the NHS which share insight into how artificial intelligence tools are being utilised and tested.
Identifying abnormalities
At South Tyneside and Sunderland NHS Trust, a study sought to examine AI’s ability to identify abnormalities with the potential to lead to bowel cancer. A computer module powered by AI was utilised during colonoscopies alongside existing tech, with the AI highlighting a “green box around possible polyps” on screen, to help clinicians view potentially cancerous or precancerous polyps (adenomas) and decide whether to remove them.
The trial involved over 2,000 patients from 10 centres across the country and found that the AI was capable of locating an extra 0.36 adenomas during each colonoscopy, described as a “big difference” in comparison to existing methods of detection. It also reportedly identified “at least one adenoma in an extra eight out of every 100 people”, and found increased detection of a particularly concerning type of polyps.
The AI reportedly helped the team identify a higher number of smaller polyps and more “flat” polyps, with trial leader Professor Colin Rees commenting: “Crucially, we know that some of the polyps that lead to cancer are small polyps or flat polyps. The AI helped us find more of these lesions, the ones which are more likely to be missed with the human eye.”
Predicting Alzheimer’s
Cambridgeshire and Peterborough NHS Foundation Trust shared how researchers from the trust and the University of Cambridge have developed an AI tool designed to help predict the progress of Alzheimer’s disease. The machine learning model seeks to predict if an individual experiencing mild memory and thinking problems is likely to develop the disease, through analysis of data from cognitive tests and MRI scans.
They found that the algorithm could distinguish between people with stable mild cognitive impairment and people who developed Alzheimer’s within three years. The trust reports that the tool correctly identified individuals who went on to develop the disease in 82 percent of cases, those who didn’t in 81 percent of cases, and could also identify people at risk of rapid deterioration.
Senior author Professor Zoe Kourtzi from the university said that the tool is “much more sensitive than current approaches” and shared hopes for its potential to “significantly improve patient wellbeing, showing us which people need closest care, while removing the anxiety for those patients we predict will remain stable”.
Tackling waiting lists
George Eliot Hospital NHS Trust and South Warwickshire University Hospital NHS Foundation Trust have launched a trial aiming to help reduce waiting lists by combining AI with volunteer support.
The AI tool seeks to identify patients at high risk of missing appointments or in need of additional support such as transport assistance, with a volunteer force then calling up identified patients to offer help.
So far, the pilot reportedly indicates a 33 percent overall increase in the number of attendances and a 28.8 percent reduction in Did Not Attend rates, compared to the period before volunteer calls were made. According to George Eliot Hospital, this translates to an estimated £36,157 a week of total added value across both trusts.
The trial has also involved collaboration with Helpforce, a charity specialising in the optimisation of volunteer deployment in the health service, providing support alongside the AI supplier to train volunteers in use of the software.
Developing personalised care plans
A study held at Doncaster and Bassetlaw Teaching Hospitals seeks to explore how AI can help with improving prevention and management of pressure ulcers, with the aim of reducing incidences and severity of ulcers as well as tackling economic and carbon impacts of their treatment.
The trust describes how the AI system will take patient data such as age, weight and pressure ulcer risk factors, in order to develop a personalised care plan to suggest “the most efficient and effective treatment” whilst also providing the lowest carbon and economic cost.
At present, the study is in data collection phase at the trust; this is set to last for around a year, before the AI tool will begin to suggest personalised clinical pathways for patients.
Lead nurse for the skin integrity team, Kelly Phillips, shared hopes that the AI tool will help to save time and reduce variation in care “based on data and evidence, thereby providing that assurance to the patient and the healthcare professional that the right care is being provided at the right time”.
More on artificial intelligence
Earlier in the week, we shared insight into our recent panel discussion focusing on AI’s use in the healthcare space, with panellists sharing views on whether AI will live up to the current hype and their thoughts on managing bias in healthcare data. Click here to read more.
Other recent news includes The European Commission announcing that AI Act, a legal framework seeking to address the risks of AI in Europe by setting out clear requirements and obligations in support of “trustworthy AI”, has officially entered into force.
And in other news, HTN highlighted a partnership between Leeds Teaching Hospitals NHS Trust and health tech startup Newton’s Tree which will see the startup’s enterprise AI platform deployed across the trust to “rapidly scale” its ability to evaluate and implement AI applications; and we noted how the University of Huddersfield is working on the development of a secure threat intelligence sharing platform with the aim of helping to protect AI-enabled diagnostic tools from cyber attacks.
