The trend of patient-side AI agent disruption overseas strongly suggests that 1800Medicare, as launched only two months ago, has to be entirely rethought now.
If our government is still wondering just how seriously we should be taking AI healthcare disruption then they need look no further than their recently launched grand national triage service 1800Medicare, which, based on overseas trends, looks set to be redundant in its current format, possibly as early as the end of this year.
That’s AI in healthcare for you. It’s terrifyingly fast.
Patient-side AI agents aren’t really in full swing in Australia yet. ChatGPT Health has launched as a limited pilot so far and Claude Health is due here soon.
This article originally ran on TMR’s sister site, Health Services Daily. TMR readers can sign up for a discounted subscription.
But in the UK, despite ChatGPT Health launching in January and then shutting down due to UK data regulations, other agents such Claude Health and Wellness AI are gaining patient traction so fast that the NHS is in a dilemma about how it will manage its comprehensive information triage service – an equivalent idea to 1800Medicare which has been around more than a decade.
In essence, it is already looking like the patient-side AI agents will become the patients’ default digital front door in the UK and that’s causing a lot of heartache in the NHS which doesn’t know whether it should build out its own AI agent using its massive store of longitudinal population health data, or somehow partner with the new agents to retain some level of control and governance in the triage process.
The battle, which is directly relevant to how we move on 1800Medicare, has been playing out over the last week week on LinkedIn.
Dr Charlotte Refsum, director of Health Policy and the Tony Blair Institute for Global Change, set the cat among the NHS pigeons with this post and story in which she says at the speed at which the new AI agents are taking the ground of the main digital front door for patients, the NHS has very little choice but to consider how to partner with, and integrate somehow, with one or more of these new agents.
“The NHS faces a strategic threat that has nothing to do with waiting lists, workforce shortages, or budget constraints. It’s about who controls the front door,” said Dr Refsum.
“Right now, millions of people are turning to ChatGPT and Claude for health advice. They’re uploading GP letters and test results. They’re asking: ‘Should I see a doctor?’ ‘Is this urgent?’ And they’re getting answers that are fast, personalised and – crucially – good enough.
“Patients know these platforms are not always 100% right. They know they’re not regulated medical devices. But they solve a problem the NHS has struggled with for years: navigating an impossibly complex system.
“The result is that these platforms are fast becoming the default first point of contact for health concerns. Not 111. Not NHS.uk. Not your GP reception. An AI assistant that knows your entire health history and responds in seconds.”
Dr Refsum sees the disruption as so serious that the very future of the NHS may be decided on which way the organisation chooses to go: try to hold on to control and build out its own AI agent themselves at enormous expense, or partner with one or more of the agents.
“The NHS does not need to outbuild big tech, it needs to work with it to shape these products – embedding guardrails, integrating systems and informing pathways,” she said.
This whole unfolding UK drama is directly relevant to Australia and its aspirations and plans for 1800Medicare and Healthdirect.
So it was interesting to see one of the first comments on Dr Refsum’s post was from the CEO of Healthdirect, Bettina McMahon.
Related
And she seemed to make it clear that her view is 1800Medicare should be looking to work with the AI agents – not outside them or against them in any way – although it’s not clear yet what Ms McMahon’s masters – every state and territory health department – think of this stance.
Notwithstanding, the question for Healthdirect and Australian governments would be, will big tech come to the table and work with them on some sort of integrated data model and governance?
NHS systems architect Andy Meiner outlines some of the problems in some of his posts.
He points out that OpenAI (ChatGPT), Anthropic (Claude) and Google are all saying the right things about safe deployment and clinical governance in health and they are actively seeking private and public partnerships with a view to improving clinical accuracy.
He also says no one wants the reputational damage that might come from a high-profile patient safety failure.
On the other hand, Mr Meiner says the AI giants are all about accumulating data, engagement and money, not constraint, which is counterintuitive often to good clinical governance and the concept of “guardrails”.
“The commercial logic cuts against exactly the safeguards we’d want,” says Mr Meiner.
“And procurement timelines on the NHS side mean ‘coming to the table’ in any meaningful sense could take years, by which point the default behaviours are already set.”
Another issue that the NHS is struggling with, and we will too, is data sovereignty. Mr Meiner talks to this problem in article posted here.
Another unique Australian problem: apparently OpenAI and Anthropic in Australia are all over the place because they are trying to set up the basics for a proper presence here. In that environment, it’s been hard to find anyone to talk to in these organisations, let alone someone with a good health grounding who would be engaged in the idea of early government involvement.
In Australia, so far, the only people accessing the new local management are said to be the giant private companies with lots of customers, transactions and data – read into this groups like the big four banks and Coles and Woolworths.
It would be hard to think of an institution with more customers, data swirling around, complexity and daily transactions than Medicare so you’d have to think that the Department of Health, Disability and Ageing would have some leverage to get a seat the table sooner rather than later.
All these are issues which 1800Medicare will have to contemplate if the government sets it free to go and investigate deals with the patient-side agents.
Things are a bit more embryonic in Australia but the NHS national front door dilemma and contemplating AI health agent partnerships has emerged in just two months.
That doesn’t give us much time to plan for what is almost certainly coming.
Then our biggest issue will be keeping up. AI is moving at a speed which government, doctors and policy people simply can’t match.
It’s going to need a lot of creative thinking. Fast, creative thinking.
