Human healthcare and its algorithmic counterfeit

One small part of the rich tapestry of current misery, in Britain at least, is the ongoing difficulty of getting to see your doctor since COVID closed all the GP practices. I haven’t heard that this is an issue beyond the jurisdiction of our Established Religion, the NHS, but maybe it’s been tried in other parts of the world too. Let me know in the comments.

According to one recent news item, a large number of practices have continued to deal mainly in online- and telephone-consultations, with very few face-to-face encounters. An article in today’s Mail by Dr Martin Scurr describes the problem (I’m pretty sure I know him, either from training or working with him somewhere, or perhaps because we both wrote for the same medical periodicals back in the 1980s).

It’s interesting to see how the correspondence pages include a minority of enthusiasts for the idea, based on how much more efficient it is. Few patients, and some conscientious GPs, disagree. And so do I.

For the first couple of decades of my time in practice, we covered our patients’ needs out-of-hours. That meant dealing with urgent issues, issues that patients thought were urgent, and the odd time-waster with no idea of what “urgent” meant. But in all cases it meant telephone consultations, and I became pretty good at the art.

Bear in mind that this system meant that, having worked a full day until 7pm, and being due to start another at 8am – or in the case of weekends simply covering the practice in lieu of a weekend off – I was incentivised to limit face-to-face consultations or collapse from lack of sleep. This is because it meant my visiting the patient, that being more time-efficient than dragging us both to a meeting point like our surgery.

Even then, the NHS’s “free at point of contact” ethos, and the increasing sense of entitlement of the public towards such a free service, meant that many issues could easily have waited until the next open clinic in the morning, so there was an educational element involved too, encouraging people in risk assessment. One was surprised at how many parents, for example, had picked up the idea from magazines that a mild fever in a child is a dangerous sign that requires hospitalisation to prevent fatal convulsions.

Be that as it may, the skill of the phone consultation involved not only asking the right questions, but listening to the answers in the right way – including whatever non-verbal cues could be sensed down the telephone line. Obviously it helped to know many of the patients personally, and hence how to gauge the significance of unspoken anxiety, and so on. Maybe 70% of diagnosis comes, in traditionally trained medicine, from the history, which put me on reasonable – though not infallible – ground.

The first point to make is that my telephone skills arose directly from the fact that every working day I was doing a score or more of face to face consultations, where examination and investigations could correct my errors and hone my history-taking. As soon as you begin to major on virtual consultations, though, that training modality will rapidly diminish.

The second point is that my telephone conversations were always backed up by:

  • If in doubt, visit.
  • Telling the caller that if my predictions turn out to be wrong (for example, the vomiting does not settle in such-and-such a time, or the fever does not ease on paracetamol), they must call again (when the adage was, “There is doubt – so visit”).
  • Even when reassurance is given, inviting the patient to be seen (in person!) early at the next surgery if necessary.

This gave patients full and well-informed confidence, corrected up most errors, and avoided having to see most things that actually were self limiting (which also taught patients confidence in self-management for next time). Very rarely, then, did I have the embarrassment of learning that my patient had had to call an ambulance in the small hours because I loused up: there is a world of difference between that embarrassment and the patient calling me in the small hours, and my organising not only the ambulance, but a bed and waiting information on the relevant ward.


The rot began to set in with the Labour Government’s idea of NHS Direct, essentially a bolt-on, algorithm-based and nurse-manned help-line advertised as a cushion for the worried well. “Worried about that painful knee you’ve had for a month? Call any time and get professional advice.” Since that advice, whether appropriate or not, was usually to call the GP out of hours, it served to undo all the “patient health training” we’d been doing for decades, thus abolishing the urgent-routine distinction almost entirely. One wonders how much of the chronic staffing problems in the profession have arisen from the resulting increase of trivia referred as “urgent by the call-centres.

But NHS Direct also served to introduce the “algorithms know best” mentality to medicine. Before too long (and just before I retired, seeing the writing on the wall), even the GP out of hours co-operatives that were the new kid on the block for a few years began to be replaced with a universal out-of-hours “service” not even manned by nurses. Any virtual consultation “nous” gained from real consultations disappeared at a stroke, to be replaced by an endless set of questions designed mostly to exclude a few urgent conditions rather than to diagnose actual ones.

More often than not, the risk-averse nature of the algorithms caused more unnecessary admissions, A&E attendances, and GP follow up appointments than real harm, and hence it did not become a scandal in the press. It did, however, lead to much frustration from patients, as I found on the few occasions since retirement when I have wanted to contact a doctor for a problem in the family. Time wasted answering questions designed to exclude heart attacks and strokes (all to be repeated when put through to some other assessor) is potentially dangerous, but definitely annoying to even a non-medical patient with a degree of intelligence… well, we’re all familiar with the experience of call-centres for banks, insurance companies and so on, aren’t we? That’s NHS 111 with a different product..

But for 15 years or so that is all any member of the public has known out of hours, apart from those isolated areas where GP co-operatives or heroic 24-hour providers remain. So it is not surprising that some people think it is a positive move to make it the norm for GP contacts, nor surprising that younger, or less experienced, GPs, feel the same way. But it’s not positive at all.

At best, it is blind to the centrality of human interaction in health-care. Though I never agreed with all his methods, Dr Michael Balint was right to say that the best medicine is the doctor himself, by which he meant that “the doctor’s person, feelings and reactions constitute a key diagnostic and therapeutic instrument.” Balint, of course, assumed that these would be responding to the person, feelings and reactions of a living patient, and not to an e-mail or a diagnostic algorithm.

At worst, as Martin Scurr rightly reminds us, it can kill people.

About Jon Garvey

Training in medicine (which was my career), social psychology and theology. Interests in most things, but especially the science-faith interface. The rest of my time, though, is spent writing, playing and recording music.
This entry was posted in Medicine, Politics and sociology. Bookmark the permalink.

2 Responses to Human healthcare and its algorithmic counterfeit

  1. Ben says:

    (Nothing to do with this article particularly, but thought you might like this)
    An outbreak of common sense in the UK? https://www.youtube.com/watch?v=JhRb5hnTseU

    Stop testing people who aren’t ill.
    ‘Herd immunity’ is impossible.
    COVID is with us for good.
    Don’t quarantine people who aren’t ill.
    etc.

    • Jon Garvey says:

      Hi Ben

      I’m not the most avid John Campbell follower, but I did see this one and he’s making good sense of the science he cites. His proposals are eminently sensible and practical.

      His problem (in general, it seems to me) is that he assumes that the best science will be accepted by the scientific community, who will then inform the politicians on rational policy, which they will then implement. The trouble is, that the scientific community has largely been bought, and even today SAGE and the government are bucking the advice of the vaccine advisory group to impose childhood vaccinations. Sadly, Joe Public overall listens to the SAGEs of this world rather than the Campbells.

      But at best, the structures of government, unlike the scientific ideal, don’t change their mind when new data becomes available, being unable to lose face. At worst, it was never about science in the first place, but power.

Leave a Reply