This week I attended London Tech Week. The entire Tuesday schedule for the ‘Impact Stage’ (one of seven stages at the three-day event) was dedicated to sessions on technological innovation in healthcare, and featured some interesting discussions.
AI was, unsurprisingly, the overarching theme which speakers kept circling back to. In light of this, and the fact that HBI will be releasing a Special Report in the coming months looking at how digital and AI are transforming the sector, I thought it would be worth taking stock of where AI seems to have the biggest potential for impact in healthcare.
Radiology
Diagnostic imaging is the area of healthcare provision where integration of AI is most advanced. An oft-cited statistic is that radiology AI tools account for around 80% of all the AI healthcare tools approved by the US’ Federal Drug Administration. According to the UK’s Royal College of Radiologists, over 60% of cancer centres and 70% of radiology departments in the UK already use AI. The digital-ready nature of the diagnostic image data makes it particularly suitable for applying classical AI to allow much faster, and even in many cases more accurate, analysis of images than is possible with humans alone. The proliferation of start-ups developing tools in this space, and the development of platforms such as DeepHealth to integrate these tools into diagnostic imaging workflows, suggests roll out will only accelerate in the coming years.
Remote monitoring
Another area where AI is already having a huge impact is remote monitoring. AI can be used to analyse data taken from wearable or ambient home devices and alert care providers when, or even before, an intervention is needed for patients, particularly in mental health, post acute and elderly care settings. UK home care provider Cera Care is one of the biggest pioneers in this space, and has raised hundreds of millions of venture capital dollars to roll out its AI-based care.
Drug discovery
The area that most people at London Tech Week seemed to agree holds the biggest promise in the coming decades is applying AI to drug discovery. This will not only reduce the cost in time and money to develop new drugs but could also radically increase the scope of what types of drugs are possible.
A panel discussion on the topic featured Oliver Vince, Co-Founder of Basecamp Research, a London based tech company that is attempting to use generative AI to move drug discovery well beyond what is possible today based on current human understanding and knowledge of biology.
“It’s not a trivial thing to teach computers biology, but the fact that generative AI models can learn unsupervised means they can learn patterns we haven’t taught them,” Vince said. “Humans have such a simplistic understanding of biology relative to what’s out there, which is why they have such a high failure rate with new trials. At least we have a roadmap to overcome this, and eventually develop types of drugs that are far more sophisticated than what we’ve got today. We’re working with Nvidia to build the largest foundation models ever, for generative designs for new molecules for a wide range of applications.”
Reducing admin workload
A less transformational application of generative AI, but one which is already available right now to be deployed and reap the rewards of, is ambient scribe tools, which can significantly reduce the amount of time clinicians have to spend on admin and boost their productivity and patient throughput by 20-30% by taking their notes for them.
Roll out of these tools is quite advanced in US hospital systems, but in Europe and the UK has been slow. However, Caroline Clarke, NHS England’s Regional Director for London, said the roll out of this “here-and-now obvious business case” is well underway in a number of settings in London NHS facilities, including A&E, and is working “pretty well”.
The move towards predictive and personalised medicine
AI has the potential to do far more than increase clinician productivity and patient throughput for healthcare providers. There are already examples of healthcare providers that are using AI to radically transform their model of care delivery, in particular shifting towards more preventive care. Israeli integrated payor-provider Clalit is probably the most advanced at doing this. Ran Balicer, the group’s Chief Innovation Officer, explained at HBI’s 2023 conference how it works.
Ultimately providers should be able to move towards more personalised as well as preventive care delivery, assuming the regulatory and logistical barriers to properly harnessing health data are overcome.
Finn Stevenson, Founder & CEO of Flok Health, a UK digital physiotherapy provider doing a lot of work for the NHS, spoke about the crucial role data plays in this at a panel on how tech is driving personalisation of care:
“Personalisation is almost a code word for good care: everyone has different health needs. The reason it exists as a concept is that historically any healthcare model you can think of has had to deal with a trade off between scalability and the level of quality (i.e. personalisation) of the care. But AI kind of breaks that trade off: for digital/AI-first delivery models, the more you put through the system, the better it becomes.”