It’s as if the government doesn’t know what it’s doing, isn’t it? And that cannot possibly be true about the NHS, an absurd thought. For we all know that it’s so exquisitely managed that no system in the world comes near it.

18 thoughts on “An extract”

  1. Quote from Tim W article

    Yea, even if the septics are involved.

    Tim did you really mean to say this, is it a typo or a pun?!
    Please explain

  2. Rhyming slang.

    Yank rhymes with tank. What’s a phrase that has tank as the second word? Septic tank. Therefore septic means yank.

  3. AI in the form of Watson etc could very soon handle much of the GP’s workload, giving better diagnostics for pennies.

    The only problem is that we are not allowed to self-medicate, so GPs will continue to enjoy handsome incomes purely on the rent-seeking basis of being holders of the prescription pad.

  4. Wat dabney:
    Agreed, plus the impressionability of NICE to the pharmaceutical industry which has removed cheap effective remedies over the counter, to be replaced by more expensive ones prescription only,( nappy rash, nocturnal asthma offhand).

  5. “Using NHS data to train an AI to make the NHS work better is terrible.”

    Not NHS data. Our data. And they are supposed to ask us first before giving it to others. That’s why the Deepmind deal was poorly conceived.

    It also seems like a sledgehammer to crack a nut. Descriptions of acute kidney injury seem to mainly revolve around dehydration. Get patients to drink water.

  6. GPs aren’t merely holders of the prescription pad: they’re gatekeepers to most other NHS services. Got a worrysome mole? You can’t go straight to a dermatologist, you have to go to your GP and beg for a referral.

  7. Our data. And they are supposed to ask us first before giving it to others.

    No, they are supposed to tell us. DPA s33 applies but there is no expemtion from the notification requirement.

  8. The apparently contrary positions come from 2 different organisations, the Department of Health and the NHS. I have never understood the logic of having the two organisations, apart from the limitless possibilities for the two organisations to waste time and money having meetings with each other.

  9. These yahoos need to get East Anglia involved. They have determined what the temperature of the earth will be in 2050, with NO DATA AT ALL!

  10. I am 6’3″ tall and weigh 185 pounds.

    Use the data any way you wish.

    There is nothing insidious about sharing data, as long as no names are attached to it.

    Tim, your use of ‘AI’ is suspect. Computers running analysis of data teasing out information is old. My dad was doing it 65 years ago. I was doing it 40 years ago.

  11. “AI in the form of Watson etc could very soon handle much of the GP’s workload, giving better diagnostics for pennies.”

    Incrementally for free. But IBM isn’t in the business of doing things for free.

    Having written many analysis programs over the years, I can say that the problem with computers, vs doctors, is that computers are literal. AI might be capable of introducing nuance, but you should realize there is no market for it.

    Doctors see the same sh!t day after day. Just different names on the charts. A system “giving better diagnostics” is hardly needed. Doctors are usually right with their diagnosis. An AI system would only produce noise.

    A diagnostic system could be useful in extreme cases, but I see no application for AI. A system containing all known symptoms. Practitioner enters all observed symptoms. System produces a list of all known afflictions exhibiting said symptoms. Then the doctor ignores it, sticking with his gut feeling. A computer making a diagnosis is not going to work. Making suggestions would be innocuous, perhaps helpful.

    One other thing. My physician of many years retired recently. One of the reasons, he told me, was that he was tired of fvcking with computers. They greatly increased his workload. The computer companies have executed masterful marketing, getting the politicians to believe that computerization is important to medicine.

  12. I’d like to see AI tackle the ‘label’ conditions which are undefined medically like autism.
    We might get some consistent analysis, rather than letting behaviourists say if someone has it or not. And an acknowledgement of the link to parenting.

  13. Agree with Gamecock (sort of).

    AFAICS, you turn up at the GP with a sore throat and a rash. The very first thing the GP will check for, are any other symptoms, bit of a sore neck?, that might mean you’ve got something that could kill you within days.
    Once that’s out of the way, they’ll basically work down the list of things they see most often, “take these three times a day, come back if you don’t feel any better by Friday”, so, no, actual diagnosis isn’t really the problem.
    And I don’t see how AI/ML helps with the first set of checks here. Regardless of the reported symptoms, someone still has to stick the probe in your ear and take a butchers. With the second check, it might, but you still need someone to carry out the interview/verification process, but it’s still going to be a decision based on probabilities, so a sorted lookup table will do. You could probably even get away with printing that out once a month and sticking it on the wall in the waiting room.

    No, getting in front of the GP in the first place is the problem. That then gets transformed into a getting in front of the specialist problem. And the NHS still writes bloody letters, and sticks them in the sodding post.

    There’s a wodge of behaviours that have become embedded in the system. AI/ML might provide the excuse to start having nurses or even non-medically qualified staff carrying out a form of triage in the process, so that could improve the response times experienced by the patient.

  14. Oh, and the other thing it does, once you have actually seen a specialist, is test for *one* thing at a time. The total round trip time for, say a blood test, with the necessary letters, is about two weeks, plus a bit for the next appointment with the specialist. So two blood tests, with three appointments, is getting on for two months.

    I *once* managed to convince a consultant that doing two separate tests on my precious bodily fluids concurrently might be a good idea. He looked at me as if I’d just sprouted a third leg.

  15. BWTM . . . Doc knows about the crud that’s going around. Seen 30 cases in a week. Someone comes in with symptoms, he’s done with the patient in a minute.

    AI will require all sorts of other – non productive – activity. And a high probability of being wrong.

  16. The Meissen Bison

    Ducky McD: getting in front of the GP in the first place is the problem

    Quite. Which is why

    “take these three times a day, come back if you don’t feel any better by Friday”

    needs to be amended to

    “take these three times a day, if you don’t feel any better come back whenever you can get an appointment or never, whichever is sooner.

Leave a Reply

Your email address will not be published. Required fields are marked *