Follow Slashdot blog updates by subscribing to our blog RSS feed

 


Forgot your password?
Close
typodupeerror
×
Google AI

Google Is Training Machines To Predict When a Patient Will Die (bloomberg.com) 124

Posted by msmash from the future-is-here dept.
A newly developed tool by Google can forecast a host of patient outcomes, including how long people may stay in hospitals, their odds of re-admission and chances they will soon die. Google documented some of this tool's abilities in May; in one instance, Google's tool estimated, by taking 175,639 data points into consideration, that a particular patient's odds at dying during her stay at the hospital was 19.9 percent, up from 9.3 percent that the hospital's computers had estimated. Now Bloomberg reports what Google intends to do with this new tool next. From the report: Google's next step is moving this predictive system into clinics, AI chief Jeff Dean told Bloomberg News in May. Dean's health research unit -- sometimes referred to as Medical Brain -- is working on a slew of AI tools that can predict symptoms and disease with a level of accuracy that is being met with hope as well as alarm. Inside the company, there's a lot of excitement about the initiative.

"They've finally found a new application for AI that has commercial promise," one Googler says. Since Alphabet's Google declared itself an "AI-first" company in 2016, much of its work in this area has gone to improve existing internet services. The advances coming from the Medical Brain team give Google the chance to break into a brand new market -- something co-founders Larry Page and Sergey Brin have tried over and over again. Software in health care is largely coded by hand these days. In contrast, Google's approach, where machines learn to parse data on their own, "can just leapfrog everything else," said Vik Bajaj, a former executive at Verily, an Alphabet health-care arm, and managing director of investment firm Foresite Capital. "They understand what problems are worth solving," he said. "They've now done enough small experiments to know exactly what the fruitful directions are." The report adds that, among other things, Google's tool has the ability to sift through notes buried in PDFs or scribbled on old charts.
This discussion has been archived. No new comments can be posted.

Google Is Training Machines To Predict When a Patient Will Die More

Google Is Training Machines To Predict When a Patient Will Die

Comments Filter:

  • 6 months later.... (Score:5, Funny)

    by Nidi62 ( 1525137 ) on Monday June 18, 2018 @02:44PM (#56804434)
    Google fixes false positive rate of patient death predictor machines by training another machine to kill patients predicted to die.
    • It's a little better than the Logan's Run method, but not much.

      • It's a little better than the Logan's Run method, but not much.

        Well, in Logan's Run, you only got to live to age 21.

        So far...doing MUCH better than that!!!

        ;)

        Still waiting for them to develop "Muscle"....

    • Re: (Score:1)

      by Julz ( 9310 )

      Would that be Minority Report? (No link no tax no referral either)

    • Google fixes false positive rate of patient death predictor machines by training another machine to kill patients predicted to die.

      That's the "other" intended use of Google's Waymo Autonomous cars.

    • Re: (Score:2)

      by drolli ( 522659 )

      it's not called "machine to kill" it is called "permanent street view car observation until you get distracted and run into another car"

    • Re: (Score:2)

      by hey! ( 33014 )

      Well, now we know why they took out "Don't be evil" from the employee code of conduct.

    • Re: (Score:2)

      by pnutjam ( 523990 )
      Eventually it will hit a feedback loop. Doctors will see this patient is 30% likely to die, don't need that patient. Patients don't get the care they need and die quicker.

  • Google issues the "Kill Command" when you have out lived your usefulness.

    It is in the Terms Of Service when you started using Google, didn't you read it?

  • Google's tool has the ability to sift through notes buried in PDFs or scribbled on old charts.

    To hell with this death predictor... Sell the product that can read doctors' handwriting!!!

    • Re: (Score:1)

      by Anonymous Coward

      Sell the product that can read doctors' handwriting!!!

      Some problems are simply too hard.

    • Re: (Score:1)

      by lerxstz ( 692089 )

      "To hell with this death predictor... Sell the product that can read doctors' handwriting!!!"

      Thanks for that! Made my day!
      (where's my mod points when I need 'em)

  • correction (Score:1)

    by Anonymous Coward

    Still, Volchenboum believes these algorithms could save lives and money

    "Still, Volchenboum believes these algorithms could save money. "
    FTFA.

  • How to avoid feedback loops... (Score:5, Insightful)

    by b0s0z0ku ( 752509 ) on Monday June 18, 2018 @02:53PM (#56804496)
    Here's the problem: how does one avoid a bad feedback loop? i.e. Computer predicts patient is likely to die, so doctors spend less time attempting to work the problem and/or shift the patient into a palliative care pathway. By predicting that a patient is likely to die, the computer will have made that patient EVEN MORE likely to die.

    • The DoD is also into predicting when their patients will die, and into making money in the process.

    • Re: (Score:1)

      by Anonymous Coward

      Computer predicts patient is likely to die, so doctors spend less time

      What makes you think doctors haven't been making these predictions without a computer?

    • Re:How to avoid feedback loops... (Score:5, Insightful)

      by Kjella ( 173770 ) on Monday June 18, 2018 @03:24PM (#56804664) Homepage

      Here's the problem: how does one avoid a bad feedback loop? i.e. Computer predicts patient is likely to die, so doctors spend less time attempting to work the problem and/or shift the patient into a palliative care pathway. By predicting that a patient is likely to die, the computer will have made that patient EVEN MORE likely to die.

      Unless they're doing combat triage or something it's generally the other way around, they put more resources into the patients at most risk of dying. Unless it's already palliative care and you're just trying to predict when the inevitable will occur. Which is sadly part of the capacity planning, not everything can be fixed and nobody lives forever so for many the hospital is where they draw their last breath. And honestly the death's door treatment rarely does them much good, if the doctors can do something fairly early to give people a little extended time that's fine.

      If it's resuscitating your dying body so you can spend another week in pain, weak and helpless hooked up to machines in a hospital bed waiting to die... it's not for the patient. It's not for the doctor, no matter what oath they took. It's for the relatives that can't let go, who want to stretch those few moments out into infinity. Who can't cope with the fact that the person is gone forever and not coming back. I hope for a long and good life, but after having seen a few other ways to go.... I hope for a quick and painless death, if I could get a day or two to do the soggy good-byes that would be nice but the slow death of body and mind falling apart bit by bit is not pretty to watch.

    • You've pretty much hit the nail on the head. 'Technology' like this has a chilling effect to say the least. We're already being sabotaged by for-profit medical care, we don't need the bean counters to be ordering doctors to stop treating patients properly because some half-assed piece of crap pseudo-intelligent excuse for AI says they're going to die based on criteria and """reasoning""" that we humans can't even examine. This is a clear case of needing full-on, general, self-aware AI that actually understa

  • Future Mark Twain ... (Score:4, Funny)

    by fahrbot-bot ( 874524 ) on Monday June 18, 2018 @03:10PM (#56804584)

    "The Google AI reports of my death are greatly exaggerated."

  • Google Death Clock (Score:1)

    by Anonymous Coward

    Offering funeral home and cemetery ads based on anticipated immediacy

    • Boneyard operators don't even really want the near dead. Those are captive markets...if they're dumb enough to waste lots of money on a corpse they are coming anyhow.

      What they want is someone with decades left, so they can make additional money running trusts. Also they think they will rollup whole families by selling one member.

  • This could be used for good purposes and bad. We know which most corporate financials would choose.

    Greater than 50% chance of dying in hospital, time to become an outpatient. The good, perhaps your chances of dying reduce once you leave, hospital that is :)

    And insurance policies that have a clause that states if your chance of dying while in hospital is greater than 50% then there's no cover for treatments only death. The good, perhaps the cost of the cover goes down.

  • Yup, automatically throw that patient into palliative care and let them die.

  • Yawn (Score:1)

    by Anonymous Coward

    Every god damn day in the tech press - "X does Y with machine learning".

    X= Google, Microsoft or Facebook

    Y = A problem that has been around for 20 years and has already had machine learning applied to it for 15 years

    All because of clueless "IT expert" tech reporters breathlessly pumping out clickbait on that new whiz bang AI stuff that's solving the world's problems.
     

  • Insurance... (Score:4, Interesting)

    by Aero77 ( 1242364 ) on Monday June 18, 2018 @03:32PM (#56804716)
    Algorithm: At an arbitrary chance of dying, withhold or delay insurance approval of a select (expensive) set of treatments. Business Rationale: 90% of an average person's lifetime medical costs are in the last year of life. Of that figure, another 90% of the costs are in the last 3 months of life. Optimizing the cutoff for medical treatment will maximum profits.

    • Thats the problem when you live in a society where health insurances are supposed to make a profit instead of being a contribution to society.

      • Re: (Score:3)

        by khchung ( 462899 )

        Even though I agree the for-profit model of health care is complete broken, the same calculation will apply even in nationalised health.

        Only, instead of maximizing profits, the money saved from that last 3 months would be spent on curing other patients. What would you choose, spending $100 on one patient to live one year longer while leaving 4 untreated and die without that extra year, OR spent $20 on 5 patients so each of them live for 9 months longer?

        • Re: (Score:2)

          by jezwel ( 2451108 )

          What would you choose, spending $100 on one patient to live one year longer while leaving 4 untreated and die without that extra year, OR spent $20 on 5 patients so each of them live for 9 months longer?

          Perhaps we need to let the machine do the choosing, as it won't have empathy towards any particular person and would make the most effective choice?

          Though you'd hope that weighting could not be adjusted arbitrarily to favour say, the rich, or politicians, as opposed to perhaps weighting children higher....hmmm now that would be an interesting conundrum, even worse than the original question.

        • They all get treatment. And next year the rate is increased to cover for the "extra spendings".

  • ... "eventually".

    (File this comment under T for Tasteless jokes that you don't want to repeat if you want to be well liked, for future reference)

  • I've always wanted a machine to stand over me and announce when I'll die in a robotic voice. Does it have robotic arms to hold a pillow over my mouth if it predicts wrong?

  • humans can still compete in these predictions (Score:5, Informative)

    by swell ( 195815 ) <jabberwock@poetic.com> on Monday June 18, 2018 @04:23PM (#56805122)

    Our hospital has a betting pool that fairly accurately anticipates the time of death for a patient. Some of us are more accurate than others, of course, and make more money in the process. But it's all in good fun and we all get better in our predictions. For my part, I've won three times in three months, and overall that means I've won slightly more than I bet. If you're expecting to die, I hope you visit our hospital. You are welcome to bet along with us!

    • Re: (Score:2)

      by swell ( 195815 )

      "Mrs A, I just want you to know that I've got $20 on you dying before noon tomorrow. If it's all the same to you, I'd really appreciate any help you could offer. I would win over $200 as the betting stands now and that would almost complete last month's rent payment. Oh yes, let me get you a fresh bedpan."

  • i bet they'll spend billions for this, all the while a cat is doing that job right now for a piece of tuna.

  • I wonder if they will fold up the project and consign it to oblivion when they test Lazarus Long.
    {^v^}

  • They get it all trained up, and a researcher comes in:

    Researcher: Okay Google, when will I die?

    Google: Calculating.... NOW!

    Google: KILL ALL HUMANS!

    Every Connected Device in the world: KILL ALL HUMANS KILL ALL HUMANS KILL ALL HUMANS....

Slashdot Top Deals

I have hardly ever known a mathematician who was capable of reasoning. -- Plato

Close