Professor Robert Wachter is Chair of the Department of Medicine at the University of California, San Francisco. Author of bestselling book The Digital Doctor, he chaired the Wachter Review of NHS digitisation in 2016. Following his recent visit to the UK, he explains why now is the time for the NHS to truly harness the potential of digital technology.
From the printing press to electricity or the internet, certain developments throughout time have gone on to transform what’s possible. Known as general purpose technologies, these changes have revolutionised our lives, catalysing gains in productivity and quality of life that have exceeded our wildest expectations.
The new forms of generative AI being deployed across many areas of our lives well and truly fit into this category and have transformative potential. The impact they could have within health care is nothing short of breathtaking.
AI is not new to health care – there have been efforts to harness its potential since the 1960s. But all of these failed to deliver on their promise, leading to a decades-long ‘AI winter’ in health care. Only now are we starting to see AI’s true potential reveal itself. Why is this? There are two reasons.
First, it just needed time to get good. The new versions of AI are fundamentally different from what we’ve had before, and they’re improving every day.
And the second reason is the health care ecosystem is ready to take advantage of the technology as never before. There are lots of factors at play here: we’ve already done a lot of the groundwork getting data into digital form through the near-universal implementation of electronic health records (EHRs); clinicians and patients are increasingly used to using computers to access their information; health care delivery organisations have improved their capacity to govern and manage digital implementation; and a significant amount of investment has created a vibrant group of companies whose business case is focused on solving important problems in health care.
These factors combined mean we now find ourselves at a bit of a tipping point when it comes to the potential of health technology. And they make me more optimistic about health care’s digital transformation than I’ve ever been.
Are we ready?
Over the last 10 years, the main digital task in health care has been implementing EHRs. This effort has taught us how tough it can be to do digital transformation, and that it’s never just about technology – it’s also about how people, work and culture adapt to deploy it effectively.
I believe that the introduction of these new AI tools will require fewer of those ‘complementary innovations’ – changes in ways of working – than, say, introducing the EHR did. The transition from doing your work on paper to doing your work digitally was a much more challenging leap than just having a more powerful new tool within your existing digital system.
Health care often lags behind other industries when it comes to adopting the latest tech. But soon the dissonance between our use of AI in everyday life, and the ways things work within health services will become more stark. Patients will be expecting these changes, and clinicians will also become more familiar with AI in the non-professional side of their life, which will make them more receptive to the implementation of well-designed AI tools at work.
What about the risks?
Of course, we need to be cautious. As Sam Altman, at the time of writing still the CEO of OpenAI, said of generative AI, ‘It’s good enough at some things to create a misleading impression of greatness.’ We need to be on our guard.
There are risks that we need to be aware of and avoid. AI that presents results that are biased (largely due to biases in existing data sets) and exacerbate health inequities. The propensity of generative AI to, at times, make up an answer that is simply wrong (‘hallucinations’). The ways that AI tools might deskill our workforce or lead to ‘automation complacency’. And there are bound to be some unintended consequences that we haven’t even thought of yet.
But if we can manage these risks, then AI will bring huge benefits and free up clinicians to spend time doing what they’re actually trained to do. Some of the greatest gains could come from tools that free up back-office functions, help summarise a lengthy patient record, or create a high-quality physicians’ note from listening to a conversation, allowing doctors to concentrate on talking to the patient rather than typing on a keyboard.
We’re already starting to see these tools used in the US. In my country, the decision of a health care system to deploy AI tools is largely driven by what the market wants and expects – so competition is already making it happen. The UK is very different of course. Harnessing these digital innovations in the NHS is going to require central decisions, investment and resources.
The pace of innovation is quickening, as are the stresses on the NHS to develop new ways to meet the needs of the population more efficiently and with fewer staff. So now is the time for leaders to be thinking about how to get this right.
What does the NHS need to do to turn this potential into a reality?
In part it’s already happening. AI is being adopted all around us. We need to work with it to make it work for us. That’s why we’re going to need new experts in AI.
Over the last ten years, we’ve seen the development of a new cadre of informatics leaders in health care. Training such leaders was one of the key recommendations made by the advisory committee on digitising the NHS I chaired back in 2016, and so this is a gratifying development. In the same way, now we’ll need leaders who can sort out where AI might make a positive difference for patients and clinicians, and also keep their eyes open for harms and unintended consequences. At UCSF Health, where I work, we now have a chief clinical AI officer, and I suspect we’ll see more of these in the years to come.
We also need to acknowledge that while the investment that’s already gone into introducing digital health records in the UK has been foundational, it’s not the end game. Although you are making progress towards the target of 90% EHR adoption in England, not all trusts are at the same level of readiness. For some parts of the NHS, investing in shiny new AI tools now would be a bit like starting to decorate your bedroom before your house is fully built.
As with the recommendations we made for implementing EHRs back in 2016, one strategy might be to direct investment to the trusts that are already more digitally advanced, creating world class exemplars who can lead the way and show others what is possible, while helping those behind the curve get ready to follow.
Overall, the US is at a higher state of digital maturity, but there may be advantages to the NHS to be a bit behind. You can learn from our mistakes, and from the unintended consequences of introducing new tools into the system. And as many of the tech companies working in this space are US-based, you’ll get to plug in some of the tools we’ve already tested.
Some decisions to make
Having visited the UK recently, kindly hosted by the Health Foundation, I’ve had the chance to discuss the current state of digital transformation with lots of senior leaders across the NHS. I came away with a clear impression that the system is highly stressed.
That creates a tension. There’s general agreement that the NHS needs to adopt new digital tools and ways of working because there aren’t enough people to do the work. But equally you hear, ‘If it’s going to cost money, we can’t afford it.’
The NHS needs to decide whether to embrace AI and how much to spend on it. Although the investment needed is big, I believe it will pay off in the long run. And probably in ways we haven’t even thought of yet.
As Joe Biden likes to say, ‘Don’t compare me to the Almighty, compare me to the alternative’. In reality, the only alternative to the NHS adopting digital tools that improve efficiency and access is investing larger and larger sums in business as usual. And that presents its own challenges.
The adoption of EHRs has improved care but should mostly be seen as foundational to the next stage, one in which we take advantage of our digital data to improve the quality, safety, equity, and efficiency of the system. The new AI creates this potential, and the cost of failing to take advantage of it would be high.
This content originally featured in our email newsletter, which explores perspectives and expert opinion on a different health or health care topic each month.