Will AI save the NHS?

by Joe Ventre, digital campaign manager


My phone has no signal, and I’m wading through a pile of outdated newspapers and magazines to pass the time - I can only be in one place. Having taken a few non-emergency trips to various London hospitals recently, long waits in corridors leave little to do besides taking in one's surroundings. This time I’m in Charing Cross hospital. Chipped doors, scuffed walls, patchwork dropped ceilings, with cabling exposed through the missing tiles. We are told relentlessly that the National Health Service, which commands a colossal budget of over £160 billion, is in crisis. When you’re sat waiting in a dilapidated, off-white corridor (or perhaps more pertinently, when it’s your workplace), it’s difficult to disagree.


Charing Cross Hospital, April 2023


And the problems aren’t confined to the NHS estate. Waiting lists for treatments have reached astronomical lengths, ambulances take days (literally) to arrive, 111 calls go unreturned, GPs seemingly do everything to avoid seeing you face-to-face - and this is on the days when the demoralised workforce isn’t striking. Some would have you believe that is due to a purposeful program of underfunding, but that simply doesn’t mesh with the reality. In 2019, before Covid ramped up budgets, annual public healthcare spending per capita in the UK had increased by 30.8 per cent over the previous decade. We’re at the point where even Keir Starmer says that more funding won’t solve the issues.


So, how does one tackle the problems in the health service? That’s just the type of question you ask yourself during a long wait. We all have our own ideas, but being a bit of a techie, and given the recent surge of interest, my mind instantly turns towards artificial intelligence (AI). In my line of work, I often devise software solutions that boost my own efficiency - they call this “working smarter, not harder”. If the will is there, there’s no reason we can’t make the NHS work smarter too.


Before coronavirus reared its ugly head, bright-eyed health secretary Matt Hancock touted how artificial intelligence “will play a crucial role in the future of the NHS”. This was in late 2018, long before we could use ChatGPT to proofread our work, or Bing Image Creator to create complex images and logos; and only computer scientists and tech-boffins could have possibly envisaged the implications of that statement. By December 2019, a study found that 52 per cent of NHS trusts were already deploying AI solutions in their services, with a further 16 per cent purported to be planning to follow suit within the next two years. These movements were met with enthusiasm from the TPA, with our research (launched by Hancock himself) demonstrating how AI and other automation technologies could bring about £18.5 billion of yearly savings in health and social care.


What do these solutions look like in practice? A good example can be found in Addenbrooke hospital in Cambridge, where Microsoft’s InnerEye system is being deployed to automatically process scans for patients with prostate cancer. “The results are incredibly exciting” according to Addenrooke’s oncologist Dr Raj Jena. “The research has shown that a computer can perform in just a few minutes what would normally take me several hours to do.” And that really is the crux of the matter. Ramping up efficiency in this way has the potential to eliminate backlogs and remove human error, ensuring that patients get the care they need earlier, thereby saving lives, money and resources. As of March 2023, UK health bodies had invested £123 million in 86 such AI technologies to tackle conditions such as cancer, heart disease, diabetes, neurological disorders, and mental health conditions. The savings against expensive operations and the like could be game-changing.


In addition to the gargantuan price tag on treatments, the NHS also incurs enormous administrative costs - some more avoidable than others. Every year, 8 million hospital appointments are missed, costing the NHS an eye-watering £1.2 billion annually. You would hope that a good dose of human decency would be the answer to tackling this waste, but yet again, the machines may beat us to the chase. An AI solution being piloted at Mid and South Essex NHS Foundation Trust considers the reasons why someone may miss their appointment (taking into account things like weather, traffic and working hours), and arranges appointments at what it believes to be the most convenient time for patients. In addition to taking a sledgehammer to the enormous cost of missed appointments, it is predicted this tool will allow as many as 100,000 extra patients to be seen at the trust every year. An additional five trusts are planning to trial the solution this year.


I’m aware that this may all sound a bit utopian. Many policy-makers retain a healthy suspicion of handing over such functions to machines. Amidst the excitement of the AI boom, a number of industry experts have voiced grave concerns and our own PM has been in America making this point. “Mitigating the risk of extinction from AI should be a global priority alongside other societal-scale risks such as pandemics and nuclear war” a cohort of scientists and tech leaders recently warned. Less drastic scenarios could see software bugs leading to false test results, or a health service too dependent on gadgets being vulnerable to cyber attacks. In a field as important as medicine, where outcomes can quite literally mean the difference between life and death, it will likely take many years before patients feel comfortable with an increasingly AI-driven NHS. After all, how does one trust software to abide by the Hippocratic oath?


There is always the sorry possibility that these undertakings collapse under the weight of NHS IT incompetency, too. Taxpayers are all too familiar with the health service’s woeful track record of upgrading computer systems, with deadlines pushed back and costs spiralling out of control. Who can forget the dismal failure of The National Program for IT (NPfIT)? This ambitious, early-noughties project sought to bring the NHS up-to-speed with the technology of the time, promising electronic patient records systems, online appointment booking services, computerised referral and prescription systems, and upgrades to network infrastructure. After nine years of cost overruns and failure to deliver, the project was put to bed, having racked up a total bill of around £10 billion and very little to show for it. A report from the Public Accounts Committee would later call the program one of the "worst and most expensive contracting fiascos" in public sector history.


Amidst the enthusiasm for these cutting-edge solutions, there is always the risk that the health service will be forking out huge sums to run before it can walk. For context, this is an institution which has had to set targets to eliminate the use of fax machines, and sees one in seven trusts still relying on paper-based record keeping. What good are AI-powered patient care plans tools if that patient’s records are stored on paper? Are “virtual wards” feasible when smart monitoring systems are rendered useless due to network failures? In the foreword to our aforementioned research paper on automation in health and social care, Matt Hancock stressed the importance of “sorting out the NHS’s data layer” - and it’s an important point. Advanced technological structures like the one envisaged for the health service must be built from the ground up, with clean, thorough, standardised data forming the foundation. Without this, otherwise worthwhile investments will be rendered useless.


It’s clear that NHSX, the unit created in 2019 to accelerate the digital transformation of health service and social care, has a monumental task ahead of them. The emergence of AI means that the entire fabric of society is on the cusp of seismic disruption, and no one is quite sure whether this will lead to a digitised paradise, a Huxley-esque dystopia, or an outright apocalypse. What we do know is that the genie is out of the bottle, and the NHS, like most large institutions, is already devoting hundreds of millions of pounds to put AI tools to work. If lessons can be learned from previous debacles, if the groundwork is laid correctly, and if investments are made with due care and consideration, there is every reason to believe that the AI boom will equip the health service to do more with less; simultaneously cutting costs, easing pressures, and saving lives. Given the enormous challenges that an ageing population is already placing on NHS capacity, this venture can hardly afford to fail.

It’s unlikely that artificial intelligence is going to fix the ceiling tiles, or make hospital visits a more pleasant experience anytime soon. But with the right implementation, a digitally-transformed NHS could pave the way for the world-leading service that hard working taxpayers deserve; one in which the dreaded hospital corridor can at least be properly maintained. We’re all familiar with politicians promising to deliver the salvation that the healthcare system so badly needs, but in the words of Professor Sir John Bell, regius professor of medicine at Oxford University, “AI might be the thing that saves the NHS.”

This website uses cookies to ensure you get the best experience.  More info. Okay