UK’s NHS trials AI app as alternative to medical helpline

UK’s NHS trials AI app as alternative to medical helpline

UK's NHS trials AI app as alternative to medical helpline
UK's NHS trials AI app as alternative to medical helpline

The UK National Health Service is to trial an artificially intelligent (AI) medical application in place of its NHS 111 non-emergency helpline.

The app comes from Babylon Health, a UK start-up specializing in remote healthcare applications. It enables users to type their symptoms into a chat box – much like texting – and the app will perform triage for urgent but non-life-threatening conditions.

The trial is set to start at the end of January, meaning that for the next six months more than 1.2m people in north central London will be able to access the app, according to the FT.

Can AI do better?

The NHS 111 helpline has been accused of making things worth within the UK health service, actually adding strain on GPs rather than alleviating it by sending too many too many people to accident and emergency unnecessarily.

Perhaps, then, an AI assistant can do better? Londoners in the boroughs of Camden, Islington, Enfield and Barnet – among others – will have the change to text the Babylon app about their medical needs.

After each text, the app carries on the conversation by asking further questions to acquire more details. For every answer it receives it consults a database of every symptom and illness in the world, and uses an algorithm to assess the urgency of the situation before giving advice.

At what cost?

Clearly, the AI chatbot can do things a normal human beings – or non clinical professionals as many 111 handlers are – cannot.

What’s more, it’s quicker. The process requires roughly 12 texts from both parties and takes a minute and a half on average, compared with 10-12 minutes for an average 111 call.

Babylon’s chief executive, Ali Parsa, told the FT that cost savings would be “substantial compared to the current model”. Currently, every NHS 111 call costs between $14.75 (£12) and $19.67 (£16) depending on the service provider.

“They have armies of people sitting there on the phone, costing them a fortune. If they get sick, they need a replacement in place. The cost difference is not little, it’s huge,” Parsa said.

Whereas Babylon’s own medical app offers a free triage service and paid for video consultation via smartphone, which only costs $6.15 (£5) a month or $30.77 (£25) for a one-off consultation.

Related: Connected medical devices improve patient engagement, but not for long

Defining success in the NHS

Babylon obviously hopes for a successful trial and widespread adoption of its product following the six-month trial. For the NHS, the decision is much more difficult.

Keith McNeil, chief clinical information officer for the NHS illustrated that challenge. “What we really need is real-world evidence to show that apps like Babylon are scalable across a whole population in a real-world scenario,” he said.

He went on to confirm that the app’s success would be assessed by “a variety of means, including trial outcomes for how many people are using it and how that experience has been for them, follow-on effect on GP surgeries, A&E etc.”

An encouraging step for digital health

Internet of Business spoke to Alex Butler, partner Open Health and managing director of digital health company, The EarthWorks, about the potential for this solution.

“The NHS should be commended for the trial of new technologies that could improve the provision of care,” he said. However, he noted that “Artificial intelligence and machine learning technologies at this stage are often over hyped and to some extent misappropriated. The public’s understanding of AI is formed through science fiction, with robotic ‘sentient’ beings.

“In truth even some of the most powerful systems at this stage are not more than vast algorithmic ‘decision trees’. Automated systems in healthcare have often frustrated users because even though they can effectively categorize the risk factors that could be red flags in need of urgent attention, more nuanced differential diagnosis has not broadly been possible and they end up giving bland ‘you should consult your doctor’ or unhelpful conclusions that give possible causes of symptoms a broad range from relatively small concerns through to cancer.

“If technology is actually able to learn, to build on the knowledge it already has to better diagnose and support users it could be exciting. Technology should provide us more humanity in our interaction. Let us hope that this pilot can demonstrate this.”

Digital health futurist, Maneesh Juneja, also told IoB that he was encouraged by the news.

“This is not a time to be fearful of change, but a time to explore new tools like Babylon,” he said. “We don’t yet know how effective this will be, hence the quest for real world evidence is welcomed.

“An app may not suit everyone, but it’s critical we offer patients a full spectrum of choices. I know many who have found phoning 111 and speaking to a human to be ineffective, so what have we got to lose with this trial?”

Related: Healthcare professionals hit by Internet of Things reality check