back to article £250m fund for NHS artificial intelligence laboratory slammed as tech for tech's sake

UK health secretary Matt Hancock has been accused of being "obsessed by technology" for its own sake following the UK government's vague announcement about injecting £250m into a AI laboratory for the NHS. The fund is aimed at improving cancer screening by speeding up the results of tests – including mammograms, brain scans, …

  1. Anonymous Coward
    Anonymous Coward

    It Needs Something

    My wife has now undergone 18 months of different treatments for an aggressive cancer. It was missed by the current methods and it was only after months of increasingly stressful insisting that there was something wrong, did the investigations find the problem. Current methods need augmentation, I worry that 2 years could be a long time to wait for progress, but sometimes one has to live on hope. At least those two years should allow the flesh to be hung on the frame and allow further evidence to be collated. Improvements cannot come too soon for those who will otherwise suffer and die.

    1. Joe W Silver badge

      Re: It Needs Something

      That is of course true, but the question is (in this case) if there is actually a real plan to do something, instead of this being an "upper manglement wants shiny-shiny"-situation, or just a "we'll throw some money at AI, because AI".

      I stumbled across an article in a newspaper last week that was actually discussing ML for cancer detection, and at the moment it does not work well enough to wrap it into a product. Yes, further research is needed, but that should be a strategic decision. I am also always for funding research (and development, though once it is product development it would preferably be market driven, but let's not discuss that point...), but I'm against just throwing money at stuff because somebody thinks it is cool and then loudly announces this as a major step (while other problems persist that would need much more money to fix). AI and ML is over hyped snakeoil (in many cases, not all of them, I know)...

      1. Anonymous Coward
        Anonymous Coward

        Re: It Needs Something

        Particularly unacceptable when the NHS is deprived of cash to achieve the basics, get wait times consistently down to under half an hour, sort the disaster that is mental health care provision (seriously people get "parked" for years waiting for treatment instead of the system being proactive and getting them into treatment ASAP to stop them getting any worse) and then we'll talk about potential "desirables" like AI/ML etc....

    2. Alan Johnson

      Re: It Needs Something

      I am sorry about your wife clearly there is a need for better cancer diagnostics, screening and therapy but is AI/ML the answer? I am sure there are areas where it can help but suspect it will play a relatively small part compared to biomedical/biochemical research. AI/ML has the potential to marginally improve what we get out of the exisiting diagnostics and optimise therapies but that is all. Gains worth having but not game changing.

      If we are going to invest this amount of money AI is probably not the best place to put the bulk of the money.

      I think this has all the hallmarks of buzzword and pre-election gesture politics.

      In this area there are continuing improvements but they are incremental and cumulative rather than dramatic.

  2. Wellyboot Silver badge
    Go

    it's worth a go.

    Sometimes you just have to take a gamble to see if something works and I don't think automating image analysis is a shot in the dark.

    £250M is loose change compared to the overall NHS budget and an absolute bargain if it improves cancer diagnosis and treatment.

    1. codejunky Silver badge

      Re: it's worth a go.

      @Wellyboot

      And if it works we will reduce the work on doctors and nurses. I dont know if I am more cynical of the politicians or the NHS in this case

      1. CountCadaver

        Re: it's worth a go.

        And if it works or not we will reduce the number of doctors and nurses

    2. Anonymous Coward
      Anonymous Coward

      Re: it's worth a go.

      I really wish I could wholeheartedly agree with you and my heart says yes but my mind says it's just a data slurp to get our data into a decent format to sell it to the highest bidder. If they could sell it now they would and 250m sounds about right for cleaning that vast amount of data and getting systems in line because they aren't doing any A.I. (statistical analysis) till that's done anyway.

    3. Duffy Moon

      Re: it's worth a go.

      "£250M is loose change compared to the overall NHS budget"

      True, nut how many doctors could they train for that?

      I had all the necessary qualifications and experience, but tried three or four times to get a place at medical school, to no avail (due to the insane amount of competition for so few places). Then they put the fees up to £9000 per year, killing all my hopes.

      1. GreenReaper

        Re: it's worth a go.

        Couldn't you just get a loan for that? I know it's not chump change, but it's also relatively low amount of money in the long term, considering the wages of a doctor.

        1. Duffy Moon

          Re: it's worth a go.

          "Couldn't you just get a loan for that?"

          I don't know. That's a lot of money to owe. I'm 49 now and I wouldn't qualify until I was about 56. You don't get well-paid until you've been in the job for a while. I don't think I have enough years left, unfortunately.

  3. AMBxx Silver badge

    Wish they'd stop arguing

    If we get bogged down in the argument of 'need more nurses and doctors' we'll never improve how the NHS operates.

    The problems are demographic, not political.

    1. Nick Kew Silver badge

      Re: Wish they'd stop arguing

      The problems are multifaceted: demographic and political being just two faces. I'd say both of those are probably eclipsed by lawyers (and fear-of-lawyers), by the media (traditional and social) raising expectations, and above all by NHS culture and management.

      (p.s. I'm not one of your downvoters. I expect I'll get a few of my own).

  4. Anonymous Coward
    Anonymous Coward

    Just don’t spend the money on Watson

    1. Korev Silver badge
      Holmes

      You forgot the icon... -->

  5. hammarbtyp Silver badge

    While investment in new technology for the NHS is always welcome, you do wonder whether a piecemeal local only approach with unclear objectives will result in the 250 million being swallowed up with very few results.

    It would be far better to for links with other health services and have a much larger and better funded research where the larger scale will make getting the necessary industry and university involvement far easier and be more cost effective and have less risk.

    I see the EU has a AI research roadmap, so maybe we should join in with that one ;)

    1. Pascal Monett Silver badge

      Just out of curiosity, why is it the NHS should be the one to invest in this tech ? Shouldn't some other university or research organization front the costs, trial the thing, prove it works and then the NHS buys the solution ?

      I mean sure, I'm all for improving cancer diagnostic and treatment, but the NHS doesn't really have the means to do the research part, does it ?

      1. Richard Jones 1

        Pascal Monett, the NHS is pursuing a range of research projects at the moment all in partnership with others such as universities and potential suppliers to the health business. Sometimes this can introduce limitations as to who can participate. While that may be necessary or even desirable, it may also rule out some whose co-morbidities may well be markers for their major condition. Thus robbing some evidence from the studies.

        Sadly, nothing is perfect and wishful thinking is endless.

      2. Chris the bean counter

        NHS should do the work as public don't trust data leaving NHS to be used by others.

      3. TKW

        The NHS constitution provides for it to carry out research, and this is such.

      4. Nick Kew Silver badge
        Facepalm

        Perhaps a more pertinent question is how these decisions are reached. Why is a politician telling the NHS where to direct its R&D efforts? Sounds like the central-planning mindset that caused socialism and communism to go bust last century.

  6. Chris G Silver badge

    I'm curious

    As to how this proven muppet arrived at the £250M figure 5 minutes after his appointment?

    What is the figure based on, a lobbyist's Christmas list, somebody's reseach or is it just a 'look at me' ploy?

    Announcements like this are almost better off being ignored in the hope that they will go away, at least Hancock is not likely to last very long before they shunt him into a cupboard instead of the cabinet.

    1. Nick Kew Silver badge

      Re: I'm curious

      5 minutes? He's been in the job 13 months, and already has a track record of dodgy[1] links to commercial companies in the healthcare biz.

      [1] Dodgy as in a conflict of interest for a minister (or civil servant) in the department of health.

  7. SVV Silver badge

    Dazzled by dreams of magical tech, the clueless waste vast amounts of money

    I read lots on this story last week, and whilst there certainly is potential for algorithims to help in pattern matching applications such as diagnostics, there was lots of ridiculous nonsense that suggested that much of this was dreamt up by fools who know nothing about which they speak.

    The main thing that caught my eye was Hancock's idea that data input and management systems could be automated with AI, so that they would work all by themselves and free up doctors to spend all their time with patients. As nobody has ever achieved this ever in any business anywhere in the world, with budgets of much more than £250 million to achieve such a potentially miraculous feat, I cannot wait to see the end result of some slightly pre-filled input forms which is the best they will possibly manage (and can be done pretty easily anyway with a small amount of non AI programming).

    I suspect that Dominic Cummings will be featuring much more often here in the future, as he sounds like even more of a moron than Hancock Half Wit. He has spent time touring Silicon Valley and gazing slack jawed at big tech companies and this has convinced him that he is now a tech guru and he must now tear the civil service apart and turn it into some sort of ghastly wannabe tech giant. Example numero uno : he wants to take out the current table and chairs setup of the cabinet room and remodel it on NASA mission control, presumably with ministers sitting in rows behind little screens gawping at banks of data. As they're obviously not going to listen to anyone who tells them their sci-fi tech fantasies are bullshit, we will have to observe all this and moan about it here as usual

    1. Adrian 4 Silver badge

      Re: Dazzled by dreams of magical tech, the clueless waste vast amounts of money

      Politician suggests throwing money at something he doesn't understand. Film at 11.

    2. Tim Soldiers

      Re: Dazzled by dreams of magical tech, the clueless waste vast amounts of money

      "Hancock's idea that data input and management systems could be automated with AI, so that they would work all by themselves and free up doctors to spend all their time with patients."

      Sorry but you are wrong, many parts of the NHS are already using process Bots

  8. Anonymous Coward
    Anonymous Coward

    Why?

    Why does adding a tickbox to stuff like breast cancer treatments ('you consent to allow your scans and diagnosis to be anonymised and used as training data to better detect cancer in others'), collating say, a few million images etc per year, churning it through a machine learning algo, then checking whether unseen stuff results in more or less accurate diagnoses, on perhaps a smallish GPU cluster, cost 250 million and take two years to get started?

    I could literally do this on my desktop right now, if I had the data.

    1. Adrian 4 Silver badge

      Re: Why?

      Might work better if we didn't hear that deanonymisation is pretty trivial, and mining of data by BigCo common.

  9. Anonymous Coward
    Anonymous Coward

    It's no substitute for more trained hospital and laboratory staff, it's like having lots shiny new scanners but can't afford the buildings to put them in or the staff to operate them.

    1. Nick Kew Silver badge

      In principle[1] it's a perfect substitute for a resource that is necessarily limited. Trouble with all those staff is that the more you rely on them, the sooner the NHS consumes 100% of GDP and still needs much more.

      [1] In practice, YMMV. I don't know.

    2. Anonymous Coward
      Anonymous Coward

      A previous article (https://www.theregister.co.uk/2019/05/21/ai_lung_cancer/) suggests that it MAYBE a replacement for more trained hospital and laboratory staff, as it frees up time spent checking diagnostic images to spend more time with patients while also resulting in less false positives/negatives

      While it won't address chronic understaffing or a lack of scanners, it at least improves efficiency of the current system, benefiting both staff and patients.

  10. Wiretrip

    Missing the point

    What is really needed is an NHS National Analytics Institute where researchers are invited to use the vast NHS datasets but where all analysis happens in house and the data never leave the premises. In this way the huge data pot could be exploited without leakage and without the Googles and Facebooks benefitting. I imagine that monetising it could be very lucrative for the NHS too.

    1. Anonymous Coward
      Anonymous Coward

      Re: Missing the point

      This is NHS Digital's job. The problem is your trusts aren't authorised to share those datasets for arbitrary research purposes.

      1. Chris the bean counter

        Re: Missing the point

        Then change the rules so the trusts are

        1. Anonymous Coward
          Anonymous Coward

          Re: Missing the point

          If you could articulate a ruleset that doesn't violate GDPR and lets the lovely wonderful, entirely nice research people you like have access to the data without big tech/big pharma/big advertising companies doing the same, I'd be all ears.

          1. Mike 137 Bronze badge

            Re: Missing the point

            It's already allowed for:

            Recital 10 "This Regulation also provides a margin of manoeuvre for Member States to specify its rules, including for the processing of special categories of personal data (‘sensitive data’). To that extent, this Regulation does not exclude Member State law that sets out the circumstances for specific processing situations, including determining more precisely the conditions under which the processing of personal data is lawful."

            Articles 9.2(g) and 9.2(i) implement this.

            Allowing that once the UK leaves the Union "member state law" becomes "domestic law" (see the Keeling Schedule to the GDPR), the powers can decide that almost any processing is "in the national interest".

      2. Adrian Midgley 1

        Best to send the programs to the data

        If you have a question, better to formulate it and send the program to the data, receiving back only the answer.

    2. Korev Silver badge
      Boffin

      Re: Missing the point

      I imagine that monetising it could be very lucrative for the NHS too.

      Like CPRD and similar datasets?

  11. Anonymous Coward
    Anonymous Coward

    Don't the NHS have a million PCs running Windows 7 still?

    1. Adrian 4 Silver badge

      Good point ! You could just turn them into a botnet and run the "AI" on that !

    2. Anonymous Coward
      Anonymous Coward

      Has Windows 7 become sentient?

      1. GreenReaper

        Should be! It's a decade old, and if you consider the eight years they spent creating it and Vista since XP was launched, it's basically an adult now.

  12. Adrian 4 Silver badge

    Where the money goes

    I was unfortunate enough to spend a week in hospital recently with 6 weeks of followup. It was immediately clear that hospitals waste a vast amount of money in poor management - for instance by putting me in a bed for 2 nights or transporting me in an ambulance every morning for 6 weeks because external support services that could have done it more cheaply were underresourced.

    1. Nick Kew Silver badge

      Re: Where the money goes

      I had a stay a few years ago. The inefficiency was staggering, not to mention Kafkaesque. They kept me four days blocking a bed, when I'd much rather have gone home and would've been perfectly prepared to come back in the morning and sit around all day waiting for a slot for the various tests they were doing. Not because I like sitting around, but because it would be less bad than being detained there and tying up a precious bed.

      1. CountCadaver

        Re: Where the money goes

        Worse my wife was admitted again post surgery a few years back as she was vomiting continuously and couldn't stop being sick (turned out it was a side effect of the drugs and inflationary gas and all she had to do was keeping drinking and puking and she would have been fine)

        instead she would have died of dehydration had myself and my mum not been there. why? because she was so dehydrated they couldn't get a line in for iv fluids but because they were so paranoid they wouldn't let her drink anything "just in case you need more surgery" and "your liver levels are abnormal"

        wouldn''t listen to me and I was strung right out anyway having not slept in days...my mum came in to see what was going on, marched all 5 feet of herself up to the nurses desk and told them in no uncertain terms that they were getting a line into her daughter in law sharpish and she wasn't leaving until they did

        nurse looked at her for a few seconds and then decided complying was the best option......

        Cue consultant hauled out of her office to have a go (why they didn't call on call phlebotomy is a mystery as they are the specialists in doing difficult iv lines) managed to get a line in

        liver surgeon turns up 2 days later, your levels were caused by dehydration and the drugs from surgery, drink plenty of fluids and you can go home later this afternoon

        wife was pissed as hell

        since then I've been an awkward sod and not averse to kicking up a royal stink, albeit the health board complaints department is more akin to "deflecting criticism department" and the regulator SPSO are a toothless tiger keen to avoid rocking the boat...

  13. Anonymous Coward
    Anonymous Coward

    re. improving cancer screening by speeding up the results of tests

    Hancock must have read it in the news recently, as I did. And being in a position of a skillful politician, he decided "AI" and "cancer screening" bring votes. Good lad! :(

  14. J.G.Harston Silver badge

    Isn't this proposal just an Expert System? Not AI at all.

    1. Adrian Midgley 1

      We can't do expert systems...

      But an opinionated system could be quite useful.

      Turn out some hinting engines for it.

  15. Anonymous Coward
    Anonymous Coward

    That £350 Million a week...

    ... isn't going to spend itself.

    More of this kind of investment!

POST COMMENT House rules

Not a member of The Register? Create a new account here.

  • Enter your comment

  • Add an icon

Anonymous cowards cannot choose their icon

Biting the hand that feeds IT © 1998–2019