EXCLUSIVE: Artificial Intelligence – An NHS lifesaver or tech bogeyman?
AI tech promises much for future health services delivery, but do healthcare workers have the confidence to use it? StackCare UK CEO Noel Verbruggen discusses the key role AI can play in increasing the capacities of the NHS and how to get healthcare professionals on board with its implementation with an inclusive and educational approach.
Last month Health Education England (HEE) published its first of two reports looking at the use of, and understanding of, AI in the NHS. HEE worked with NHS AI Lab to explore the factors influencing healthcare workers’ confidence in AI technologies. The report followed the 2019 Topol Review, which recommended developing skills within the healthcare service to use AI and robotics.
The report makes for interesting reading, not least of all in that it addresses the fast-moving area of AI development and how it can be used to help deliver first class healthcare. AI has the potential to revolutionise many areas of healthcare provision, not least in the delivery of clinical capabilities in diagnosis and screening, drug discovery, digital epidemiology and personalised medicine.
It could help with organisational resources, system efficiencies and clinical workflows. The very nature of AI in that it works in the same way as human thought processes means it represents a potentially huge extension to the capacities of the health service.
However, there is a large fly in the ointment.
The report rightly points out that most healthcare workers lack direct experience of working directly with AI technologies. The report cites a 2020 survey of over 1,000 NHS staff by the Health Foundation, which found that 75 percent had seen or read “not very much” or “nothing at all” about automation and AI.
This represents a real problem. We can laud the possibilities of AI technology and its world-changing possibilities but what about the people we expect to consider it, implement it and work with it? If health sector staff don’t understand AI, don’t trust it, and therein don’t have confidence in it, how can we expect them to roll it out successfully across our healthcare network?
The last mile challenge
HEE also surveyed 240 AI technologies and found that over 20 percent of them were estimated to be ready for large scale deployment within 2022 and a further 40 percent in the next three years.
This is great, but again, we don’t seem to be addressing the real need for the people on the ground in our healthcare sector to understand these technologies and feel confident in working with them.
I have to put my hands up at this point and cite personal experience. My firm StackCare is based on AI technology. Our software uses AI and machine learning to understand the behaviours of our users, and the understanding of AI amongst both healthcare professionals and the public is often limited at best.
We spend a lot of time explaining how it works and its potential. The response my colleagues and I get is often one of amazement. To be fair, people are really impressed by the capabilities of AI technology and are quick to see how it can provide crucial capacity in helping monitor and report on the wellbeing of our elderly and vulnerable. There is a huge ‘wow factor’ to AI – once people understand it, they’re quick to fast forward to the benefits and potential it represents.
But how do we get our health sector professionals on board? Educating clients or members of the public is one thing, but educating a major industry sector and its manifold parts is quite something else.
Additionally, given the pressures on the NHS in terms of manpower, targets and increasing public expectations of service delivery (not least of all due to recently increased taxes!), the need to get AI integrated through the NHS, and beyond, to help solve these problems is time critical.
Education, understanding and acceptance
HEE’s report rightly recommends an approach focused on strong implementation of AI – through confident leadership and management that advocates AI as a strategic asset – essentially supporting the implementation of AI both from a top-down approach and also by engaging end-users through early-on discussions about their needs and how AI can help. AI adoption needs to be inclusive, not forced without discussion.
Of course, AI use will be adding to a real spaghetti bowl of existing NHS IT infrastructure. The UK Government has recently put forward its thoughts and aspirations on digitisation of more NHS services, and the addition of more tech-led solutions can look like adding even more to the plethora of ‘solutions’ optimistically designed to help. AI could get lost in this and that needs to be avoided.
The people on the ground
A failed approach to AI implementation across the NHS will lead to wasted resources, not least in terms of workflows and patient care. Because of how our healthcare is delivered there is also the possibility that some parts of the country will benefit from successful AI roll out whereas other won’t. Are we facing another potential health postcode lottery? I fear we could be.
At this point we can avoid this; there’s no reason why AI adoption needs to be another health service IT failure or missed opportunity. There is also no reason why people should miss out on the many, huge benefits that AI can bring.
I welcome HEE’s report whole-heartedly – it highlights both the opportunities that AI presents and real concerns about how it’s adopted and it will hopefully get people thinking. The report’s main recommendation, that we develop a programme of education for health professionals at all levels to help them confidently evaluate, adopt and use AI, is a sensible one and would go a long way with helping ensure uniform adoption of AI and avoiding potential up-take issues.
AI represents such a massive opportunity for our healthcare provision we can’t afford to ignore it, or adopt it badly. It would be our healthcare end users who suffer. We need to seize AI with both hands and ensure we maximise its potential to help deliver a truly future-proof health service.