Following the highly successful Digital Health AI and Data in October, all eyes will be on the AI, Data and Analytics Stage at Rewired 2024. Three members of the programme committee for the stage, Haris Shuaib, Ruth Holland and Hassan Chaudhury, answer questions put to them by Digital Health News on AI in the NHS and healthcare
The stage will focus on the rapid evolution of AI, data and analytics technologies and services, such as ChatGPT, and explore how these technologies are being used currently by healthcare organisations and how they will next provide a pathway to powerful and disruptive new innovation capabilities.
There will be a ‘live AI bake-off’ between ChatGPT and human doctors at the end of the first day on the stage at the NEC in Birmingham, where a panel of expert doctors will critique diagnostic responses to a series of model patient diagnostic queries fed in advance into ChatGPT and dedicated health large language models (LLMs), with the results shown on stage.
Other highlights on the AI, Data and Analytics Stage include a keynote session featuring NHS England chief data and analytics officer Ming Tang, and day two keynote session where Dr Nadine Hachach-Haram, founder and CEO of Proximie, and Professor Erika Denton, interim national medical director for transformation at NHSE will speak.
Ahead of Rewired24, programme committee members Haris Shuaib, head of clinical scientific computing at Guy’s and St Thomas’ NHS Foundation Trust, Ruth Holland, deputy CIO at Imperial College Healthcare NHS Trust, and Hassan Chaudhury, commercial director at DATA-CAN, answered questions on AI in the NHS and healthcare put to them by Digital Health News.
Digital Health: What have been the most promising AI developments in recent months in the NHS and wider healthcare sector?
Shuaib: “I think the UK is at the leading edge of applications of LLMs in healthcare; we have a lot of UK startups like Tortus and Phare health as well as a lot of European startups with traction here like Corti.”
Holland: “Tools that help with some of the tough and high impact issues the NHS is facing – clinician fatigue and burnout (e.g. auto-generated discharge summaries, ambient AI supporting the creation of clinical notes and patient and GP letters in outpatients) and patient flow in the emergency care pathway (predicting patient discharge).”
Are there any notable success stories or case studies showcasing the positive impact of AI in the health sector?
Shuaib: “I think the notable success is the fact that there are so many deployments now across the country.”
Holland: “Mainly in imaging so far.”
Are there any limitations or drawbacks to the current use of AI in healthcare that need to be addressed?
Shuaib: “I think the lack of understanding and regulation around some LLM uses cases and also continuously learning AI is a limitation.”
Holland: “Immature data infrastructure in most acute trusts and lack of investment in this area. We need platforms that can process big data. Shortage of data scientists, data literacy and training of staff to effectively interact with AI tools.”
Chaudhury: “We can’t take it for granted that we will scale the adoption of AI in healthcare. The barriers are steep. Innovators typically voice concerns around accessing data, navigating regulations and standards, generating robust evidence, interoperability on deployment, and progressing (beyond promising pilots) towards repeatable sales.
“Each step has costs, and engineering teams are already expensive. Without making this journey smoother many exciting businesses will fail despite their innovation. We specifically need more platform support for medical AI innovation.
“That is why there is such excitement around Haris Shuaib’s new startup, Newton’s Tree, building on his initial invention and development of the AI Deployment Engine (AIDE) at GSTT.”
Does the increasing presence of AI pose any sort of threat to human doctors and if so, why? What is your opinion on the current regulatory environment for AI and the potential impact of the AI safety summit?
Shuaib: “Yes it does pose a risk in the medium term. Doctors roles and responsibilities will be radically redefined. I think the current regulatory environment is quickly changing and adapting thanks to political momentum. The impact of the AI safety summit overall was limited though.”
Holland: “Appointment of AI Officers will help (either as an extension of chief data/analytics officer role or an independent role similar to a Caldicott Guardian). This is a requirement of the UK Artificial Intelligence (Regulation) Bill that had first reading in House of Lords on Nov 23.”
Chaudhury: “The discussion of AI replacing doctors appears to me to be a red herring. The context is that there is a huge shortfall in the global medical workforce. This is an urgent matter, with international demand for staff outstripping qualified supply.
“We’re competing to attract the best talent to move to our country already. The question is therefore how best to augment the staff we have. It’s also another compelling reason, if one is required, for why we should be paying junior doctors more than high street baristas.”
What further impact do you expect AI to have in the sector over the coming year and will it live up to the hype?
Shuaib: “I think we haven’t even realised how radical the transformation LLMs will create will be.”
Holland: “It feels like AI is finally coming of age in healthcare in the UK but there are a lot of foundations that need to be put in place before the benefits can be delivered at scale across the system.”
Chaudhury: “Workforce burnout is very real, and AI presents the most promising and transformative opportunity. We simply don’t have a better way of reimagining our tasks for the future than AI.
“We want AI that is powerful enough to match or exceed human reasoning. We want startups to get to repeatable sales and then scale so they can continue to innovate.
“We then want to use the AI to augment our people, save us from busy work, and save us from a future where our workforce burns out, underpaid and unappreciated, waiting for AI innovation that never got past its (many) pilots.”
Rewired is at the Birmingham NEC, March 12-13. Find out more about the AI, Data and Analytics Stage, explore the programme, and register here.
1 February 2024 @ 11:47
I don’t doubt the ability of AI to speed up some processes or support (or replace) clinicians in some roles – for example imaging interpretation as in the article. However difficult to escape the notion that the general complexity of health systems is deliberately downplayed, as with so many things where there is commercial imperative; sure we can relieve those one-dimensional bottlenecks where AI may excel, but we’ll soon bump into a more complex one where (current) AI can’t!
18 January 2024 @ 19:13
This is fantastic