16 March 2022

We’re told apps, AI and video calls will transform the NHS. What about GPs?

The 111 computer algorithm cannot make diagnoses or manage risk and uncertainty.

With Ukraine dominating the news, a proposed overhaul of English general practice, published by the right-wing think tank Policy Exchange on 4 March, received little media attention. This is a shame because, with its commendatory foreword by the Health Secretary Sajid Javid, we can assume the proposal, titled At Your Service, is pretty close to the government’s vision for primary care.

One prominent idea is that we will in future make contact with our health service through something called NHS Gateway. Sharing a name with a long-defunct supermarket chain, NHS Gateway is conceived as a central hub that all GP surgery phone lines and NHS 111 will feed into. There, the unwell or anxious patient would be met by the latest “symptom checker” apps and AI technology, which would then – so the vision goes – divert a big chunk of people down a “self-care” route or to other sorts of help. This will reduce doctors’ workloads, meaning the government can continue trying to get away with having fewer GPs per head than virtually any other developed nation.

Only it won’t. We know this because it’s already been tried with NHS 111, the non-emergency helpline. Whatever the symptom, there is usually a wide range of possible causes, from trivial to life-threatening. NHS Pathways, the computer algorithm that powers 111, cannot make diagnoses or manage risk and uncertainty. So it has to err on the safe side, sending 20 per cent of its callers straight to A&E or the ambulance service (in my experience, a GP handling the same phone calls will send around 2 per cent). That’s not a problem unique to Pathways: no algorithm or machine-learning program can make diagnoses or manage risk, though plenty of profit-hungry tech entrepreneurs tell think-tankers otherwise. Policy Exchange should have consulted Jason Maude, co-founder of Isabel, which is by far the best AI diagnostic support tool. He would, I am sure, have given them a realistic picture.

If a patient does eventually clear NHS Gateway and obtain a consultation with a GP, the doctor is likely, in areas with the worst shortages, to be based elsewhere – even overseas – working via video link. I’m not sure what happens when patients need to be examined: one can only assume they will be booked into a hub run by an entity called something like NHS Palpate. That Javid put his name to this document after siding with the Daily Mail vitriol last year about the lack of face-to-face GP appointments is breathtakingly hypocritical.

An issue that no one seems to grasp is training. The clinicians currently diagnosing and risk-managing patients remotely can do so because they learned through traditional face-to-face practice. They have extensive experience of correlating patients’ stories with their true clinical state. There is a fundamental question about how this new world of remote medicine will affect our ability to train future practitioners.

There are some constructive ideas in the report, not least on premises costs. And the policy nerds have at last acknowledged the overwhelming evidence that what really matters is continuity of care: where doctor and patient actually know each other. If we can restore this extraordinarily simple concept – widespread in the NHS 20 years ago – it will enhance the patient and doctor experience, and research shows it would substantially reduce hospital admissions and referrals. Submissions to the Health and Social Care Select Committee inquiry into the future of general practice show how some practices have managed to greatly enhance continuity again, yet At Your Service contains no policies for restoring this vital aspect of primary care.

Above all else, we need many more GPs. Training one takes ten years, but there is a huge pool who could return to practice or increase their time commitment right now. All the government needs to do is make the job sustainable and attractive again, including in deprived communities. Few wanted to be GP partners in the early 2000s, but within a year of a new, fit-for-purpose contract doctors were flooding back in.