Slashdot Mirror

← Back to Stories (view on slashdot.org)

Just Months After Jeopardy!, Watson Wows Doctors

Posted by Soulskill on from the do-no-harm-while-killing-all-humans dept.

kkleiner writes "Following its resounding victory on Jeopardy!, IBM's Watson has been working hard to learn as much about medicine as it can with a steady diet of medical textbooks and healthcare journals. In a recent demonstration to the Associated Press, Watson showed a promising ability to diagnose patients. The demonstration was a success, and it is the hope of IBM and many medical professionals that in the coming years Watson will lend doctors a helping hand as they perform their daily rounds."

37 of 291 comments (clear)

  1. Re:Google the answer by robot256 · · Score: 2

    I'm sure that if Watson suggested a hysterectomy for a male, it would be because it was totally stumped and would give a very low confidence value. That's the reason we would still have doctors even if the computer worked great most of the time--hopefully the doctors can catch the computer's mistakes as much as the computer can catch the doctors' mistakes.

  2. Re:Speech Solutions? by Grygus · · Score: 2

    Not to mention that you want doctors to actually use the system, which is less likely if you're giving them a bunch of extra stuff to type. That makes it feel administrative; many doctors feel burdened by paperwork as it is.

  3. You expect a doctor to spell? by msobkow · · Score: 2

    In order for a doctor to use a text input device, they'd have to be able to spell. And given the number of times my pharmacists have had to call for clarification or "interpret" a doctor's scrawl, I'm pretty sure the vast majority of them can't.

    But that doesn't change the fact that speech-recognition technology still can't deal with accents very well, and it's been a long, long time since I've seen a doctor who was born and raised in north america.

    --
    I do not fail; I succeed at finding out what does not work.
  4. They've had these for years by tehpuppet · · Score: 5, Funny

    Hello Patient, my name is Dr Sbaitso.

    I am here to help you.
    Say whatever is in your mind freely,
    our conversation will be kept in strict confidence.
    Memory contents will be wiped off after you leave,

    So, tell me about your problems..

  5. Re:Idiocracy... by Grygus · · Score: 3, Interesting

    You do realize that memorizing and regurgitating known information is the perfect skill for 99.9% of medical diagnosis? As long as Watson knows how to say, "I don't know" it will be as good or better than the vast majority of humans at this particular task.

  6. But can it do differential diagnosis? by a_hanso · · Score: 2

    and then order a lumbar puncture, MRI and broad spectrum antibiotics for the infection and then ridicule its human doctors' diagnoses with its acerbic wit?

  7. Better he use Google than watch House M.D. by crovira · · Score: 3, Insightful

    That dweeb will almost kill you twice or three times with misdiagnoses before he finds the right one.

    --
    MSBPodcast.com The opinions expressed here are my own. If you don't like 'em... Think up your own stuff.
    1. Re:Better he use Google than watch House M.D. by Nirvelli · · Score: 3, Funny

      Only because you lied to him about something !

  8. drwatson by DigitAl56K · · Score: 3, Funny

    Not recommended! He only responds to crashes, and most of the time you end up being disassembled...

    1. Re:drwatson by TavisJohn · · Score: 3, Funny

      NO DISASSEMBLE NUMBER 5!!!

  9. Re:Idiocracy... by similar_name · · Score: 3, Interesting

    However, they're still a ways to go before they can actually create new information.

    This is true of most people.

    Once they can do that though, that's when AI becomes a reality.

    I always found it interesting that computers are never good enough until they can beat the best that humanity has to offer. Computers could beat most people at chess long before beating grand masters, but it wasn't until computers could beat the best human in the world that they were good enough. Likewise, Watson had to beat the best Jeopardy players before being good enough. So now, computers have to be better than the best doctor before being good enough. So, even if you make a computer that could graduate in the middle of a class of doctors, it won't be good enough until it can do better than them all. I just find it interesting as it says so much about us.

  10. Wasn't this the promise of... by Nutria · · Score: 3, Interesting

    LISP and Prolog-based expert systems 30 years ago?

    --
    "I don't know, therefore Aliens" Wafflebox1
    1. Re:Wasn't this the promise of... by brusk · · Score: 3, Interesting

      Not quite. The idea then was that we would teach the machines the rules, and they would follow them better/more cheaply than a human brain. The innovation here is that the system goes and looks at the published medical literature and figures out the rules on its own.

      --
      .sig withheld by request
    2. Re:Wasn't this the promise of... by ArwynH · · Score: 3, Interesting

      IIRC my AI classes correctly, those systems worked. At least they had a very high accuracy, higher than most doctors. The problem was not technical, but legal. Who do you sue if the computer gets it wrong?

      Which makes me wonder: will this system will fair any better?

  11. Re:Google the answer by arb+phd+slp · · Score: 2

    "Like spell check."

    That's a really good analogy for how it should be used.

    --
    There's a perfect xkcd for my sig but I'm too lazy to look it up. sudo someone go find it.
  12. Potentially Useful by izomiac · · Score: 4, Interesting

    Personally, I seriously doubt that Watson will ever advance to being able to replace a doctor for non-trivial complaints. First of all, humans are better at image processing, so if a patient looks like death then they aren't going to ask questions to rule out minor complaints. Second, patients usually don't know how to describe their symptoms, and it's up to the doctor to make sense of what they're describing (keeping in mind that some exaggerate, some understate, and others outright lie). Third, clinical references are written for humans, so they often omit various "obvious" things (e.g. to get Lyme you have to have been bitten by a tick, which may not be very likely in Barrow, Alaska).

    OTOH, I can see Watson being immensely useful on the back end. For example, which second-line blood pressure medications have been show to be highly effective with few side effects in 65 year old male caucasians who also have diabetes, and, of those, which has the best interaction profile with the other drugs this patient is taking? Clinical guidelines help, but they're obviously simplified and generalized. It'd take a human ages to research the literature to figure that out, but an AI like Watson could potentially do it in a few seconds. Such a tool could take a lot of the guesswork out of medicine.

    1. Re:Potentially Useful by benhattman · · Score: 2

      You're jumping to the wrong conclusion. Watson and similar systems are not intended to replace humans. At least, not in the sense of removing them completely. They fall into the typical garbage-in-garbage-out situation where you need a real expert, like a doctor, to describe the symptoms so the system can produce a valid diagnosis.

      Here's how it's actually going to work. The doctor will spend more time talking to the patient to get an accurate understanding of symptoms. The expert system will then tell the doctor what the diagnosis is, and the doctor will always take that recommendation because his malpractice insurance will say that if he does a good job collecting symptoms and he follows Watson's advice, then he is not liable for incorrect diagnoses. Overall, health care will regress to the middle, as good doctors let the system do their work (but see more patients) and bad doctors have a system helping fix their mistakes. Costs will also come down some due to the increased efficiency and because the system will not be influenced by the number of specialists living nearby like human doctors are.

    2. Re:Potentially Useful by Dachannien · · Score: 3, Interesting

      First of all, humans are better at image processing

      Medical image processing is a rapidly advancing field, but it's not a lack of technological advances that will stand in the way of automatic diagnoses based on x-ray/CT/MRI. Instead, the threat of malpractice will require that doctors manually inspect the images and render a diagnosis. Otherwise, you could probably expect automatic diagnosis from medical imaging in the next decade or so (or sooner, depending on what kinds of illness or injury you're looking for).

      Malpractice is such a threat to medical imaging technology that a lot of medical imaging system manufacturers are afraid even to implement simple noise reduction techniques. If a cancerous spot gets removed as noise from an image and the doctor misses it, it could be a multimillion dollar suit against both the doctor and the manufacturer.

      OTOH, I can see Watson being immensely useful on the back end.

      Well, he did always want to be a proctologist when he grew up.

    3. Re:Potentially Useful by AndOne · · Score: 2

      I believe your example of the Lyme disease is erroneous. Watson would be primed with that sort of information as it can be fed a list of common causes and common environmental issues quite easily. Watson could then say, "you're symptoms sound like Lyme disease, but that is not common here. Have you traveled to these areas recently?" Pretty much the same as a doctor or nurse at a clinic.

      Watson will also be able to do things like search the entire patient history and perhaps identify lingering things that add up into a whole or flag symptoms for non-specialists. For instance, I had a history of shortness of breath in the morning and occasional acid reflux as well as sinus pressure/infection, I went to a couple of doctors/school nurses and they did the standard tests and found nothing wrong. I have very powerful lungs. Take some antibiotics, you'll be fine. However, 3 or 4 years later I'm on my ass with chronic fatigue due to allergies. I was even taking allergy meds.

      Turns out the acid reflux and shortness of breath were a semi-obscure allergy symptom due to chronic drainage since my sinus were nearly swollen shut and the meds I were taking weren't in large enough doses to control the issue. My allergist spotted this almost instantly. You would have to think Watson would be able to learn that sort of behavior as well.

      I do agree with you regarding the image processing capabilities. I have been doing research in medical imaging for the last 5 years, and it's a tough field. However, it's a field that's ripe for machine learning approaches(and I have colleagues who do this sort of research) since there's already large amounts of labeled data generated every day. You'd need technicians to take images, or measurements perhaps, but I think you're underestimating how powerful Watson could be in this area.

      That said Watson isn't going to eliminate the need for doctors as it doesn't have hands and can't take readings or run MRI's, but it will probably greatly speed up diagnoses and like I said bring specialist knowledge to non-specialists.

      --
      I don't care what you say, all I need is my Wumpabet soup.
  13. Re:Interesting but... by oddaddresstrap · · Score: 2

    Kidding about what? When this kind of technology becomes affordable, and it will, you might need someone (ie, a nurse) to describe the visible symptoms and translate the patient's complaints to the Digital Doctor (tm). If need be, the DD will review digitized x-rays, cat scans or mri's and then come up with a diagnosis and treatment that is probably at least as good as a doctor and will be less expensive.

    Sometimes I think I'd actually like something like this if it can do a better job than a human. At the local clinic, the doctors and nurses seem utterly clueless unless there's a broken bone sticking out or something obvious like that.

  14. Computer... by DigiShaman · · Score: 2

    Just do as Geordi La Forged does. When he has a problem, he sits down and speaks into the air starting with the following statement "Computer...".

    So, this is how much of our research will be conducted at office around the world. Get ready for the revolution. This will be much easier than "googling" you're way out of a problem. Much MUCH easier.

    Production: Computer... have X-materials with Y-funding and an Z-deadline. What is the most profitable solution.

    Investor: Computer...I have money in the bank and need to do some low frequency trading. Please review the past history and make me money.

    Mechanic: Computer....These are my symptoms for this make/model a vehicle. This is the work previously done on it."

    Inventor: Computer...design me the best fractal antenna you can.

    Developer: Computer...design a better version of yourself, put it into production, and repeat. Queue theme music to the Terminator

    --
    Life is not for the lazy.
  15. I have ... by CatNTHat · · Score: 2

    ... detailed files on human anatomy.

    http://www.imdb.com/title/tt0103064/
    http://www.imdb.com/title/tt0103064/quotes

    --
    Of course it's company policy never to, imply ownership in the event of a dildo... always use the indefinite article a d
  16. Better job than humans by Oxford_Comma_Lover · · Score: 3, Interesting

    It will absolutely do a better job than a bad human. This should make a major difference in the long tail--i.e. things that aren't the obvious problem to the doctor, notably in second and third-rate hospitals. It will make procedural screw-ups a bigger cause of death and hospital problems as compared to medical malpractice. (I'm not sure what the ratio is now.)

    It will also make humans more dumb and less thoughtful over time. That is, diagnostic skills will go down as diagnosis becomes done more and more by computer. The excellent doctors will still be excellent, but there will be even *less* requirement to really *think* about a problem than there is now.

    --
    -- IANAL, this isn't legal advice, and definitely isn't legal advice for you. Also, Squee!
    1. Re:Better job than humans by hoggoth · · Score: 2

      > It will also make humans more dumb and less thoughtful over time. That is, diagnostic skills will go down as diagnosis becomes done more and more by computer. The excellent doctors will still be excellent, but there will be even *less* requirement to really *think* about a problem than there is now.

      And THAT is when us programmers will finally be the last thinking humans on Earth! Mwah-hah-hah! You pretty little Eloi go on having your pick-nicks in the sun while we... um... toil in our basements to make better digital doctors...

      But... we like pick-nicks too...
       

      --
      - For the complete works of Shakespeare: cat /dev/random (may take some time)
    2. Re:Better job than humans by CycleMan · · Score: 3, Insightful

      I wonder if, 25 years after this, "the excellent doctors will still be excellent"? There will always be a range of ability levels, but how will the range change with this? One thing I've heard from my pilot friends is that the commercial airline industry used to be able to rely on very good pilots coming out of the military and taking civilian jobs flying big jetliners. Since the average number of hours of flying time for a commercial airline pilot was high, the pilots were very capable in handling unusual situations. Now that fewer ex-military pilots are being created, commercial airlines are pulling from other places and getting less-qualified individuals as a result, and their relative lack of flight time is cited as a factor in some incidents. So the likelihood of "excellent" pilots is decreasing, due to the declining caliber of the collective pool of pilots from which to draw excellent ones. I can't demonstrate conclusively that this is analogous to the situation we will experience in medicine with increased computerized diagnostics but I think it is an important question to work through.

    3. Re:Better job than humans by twebb72 · · Score: 2

      It [Watson] will also make humans more dumb and less thoughtful over time

      I would argue the quite the opposite. Statistical probability is what Watson does, and knowing comparatively, the likelihood of having one illness over another is a very valuable learning tool; its not at all limited to a diagnostic tool. Using a system like that would be akin to knowing how to Google well. I know for certain that I've been able to educate myself, faster, by having access to relevant search results. Having a resource like Watson and looking at his suggestions, objectively, would unilaterally produce more proficient doctors.

    4. Re:Better job than humans by ipwndk · · Score: 2

      I highly doubt it will make humans dumber. It's not as if our brain capacity lowers. And it's not like we have become dumber as technology has advanced; the opposite is actually the case.

      What future doctors should know however is what will change. Perhaps they can work on better treatments, now that they do not need to worry about diagnostics.

      Basically, as technology levels increase, the academic level on the Universities increase. But that's already the case, so relax. (Or should be; sometimes the levels decrease because of politics and economics)

      --
      01 REDEFINE REALITY.
  17. We don't need no stinkin' AI! by Shauni · · Score: 3, Insightful

    Just bring us into the 21st century, for the love of FSM! Modern healthcare is not a doctor proscribing a treatment anymore... it's a network of specialists making recommendations and sharing data with each other. However, this "sharing" more often than not goes at pre-Internet speeds. Delays of days or even weeks are common as multiple opinions are sought, insurance companies are contacted, enormous paper portfolio are passed around, one for each facility... it's a real mess. It's not "doctoring" that keeps them busy; it's bureaucracy. It's reading test results off of carbon paper forms and waiting to see if their patient can even afford the "gold standard" treatment they want to give them (even if they're insured!)

    Watson can't deal with any of that, really. And that ignores the danger bureaucratic errors can pose to an AI, such as test results that are inexplicably attributed to the wrong patient... what happens when Watson makes a crap diagnosis because of bad data? Can he eliminate bad data or even "show his work?"

  18. It's not like the idea for Watson is new by Sycraft-fu · · Score: 2

    What is new is that it works. The concept of a system that can search through all kinds of data and intelligently answer natural language questions is something that people have been trying at for a long time. However Watson works. There are restrictions, it is domain specific (the original Watson was for Jeopardy questions), it isn't perfect, and so on, but it works.

    Hence all the excitement. It isn't that other systems didn't want to do something like this or promise this, it is that Watson delivers.

  19. Re:Ten years ago on my Palm Pilot by maztuhblastah · · Score: 4, Insightful

    Do you think they do plastic surgery because it cures people? Do you think they are treating ulcers with tagamet instead of antibiotics because the antibiotics would cure you fast?

    I was starting to listen to what you were saying until I read this.

    The current standard treatment for Helicobacter Pylori is a triple-therapy regime which does indeed include antibiotics. It is highly effective and usually results in eradication.

    Cimetidine hasn't been used as a treatment for ulcers in since the discovery of H. Pylori, many years ago. Considering that there are a number of modern antibiotics that are active against H. Pylori it is quite rare for a patient to not receive some antibiotic cocktail -- and even if there were a patient who (for some reason) could not receive *any* antibiotics, PPIs would almost certainly be used in place of cimetidine.

    I'm sorry you have such a vendetta against physicians. Perhaps your views will change with age. I know that mine certainly did as I entered adulthood.

    --
    The real litigious bastards...
  20. That's why we build MEDgle by ashdamle · · Score: 2

    It was to help solve this exact problem that we started MEDgle - http://www.medgle.com/ . We've just finished our app clinicians and starting beta testing with hospitals and urgent cares. Our focus is to enable scalable clinical operations powered by 100+ million relationships and algorithms. Also our entire health analytics cloud is available via our APIs - http://www.medgle.com/corp/developers/ . Just contact just for API access. Feedback and suggestions are very welcome. :) Cheers Ash

    1. Re:That's why we build MEDgle by maztuhblastah · · Score: 2

      Feedback:

      Don't have your site crash with a NullPointerException when somebody with a blank user agent visits it.

      Remember, someString.equals() only works if someString != null

      --
      The real litigious bastards...
  21. Re:Luke Wilson warned us by stfvon007 · · Score: 2

    This is just another step towards the realization of the movie "Idiocracy".

    I can picture it now!

    year 2511:

    Unintelligent Doctor: Watson, the patient is screaming "AAAA It hurts so bad" and recently fell off a bike before coming here. What is wrong with him?
    WATSON: WHAT IS LEG?
    Unintelligent Doctor: *pokes leg*
    Patient: AAAA IT HURTS!
    Unintelligent Doctor: Wow Watson, you were right again! There is something wrong with his leg! We had better amputate it right away!

    --
    All misspellings and grammatical errors in the above post are intentional and part of my artistic expression.
  22. Shrunken Head Ned by kpoole55 · · Score: 2

    A lot of comments her make me think of Shrunken Head Ned, the world's only Shrunken Head Village Doctor that plies his trade in the Adventureland in Disneyland. (At least, he used to. I haven't been there for a few years.) That's the way a good many of these comments seem to lean, that Watson as a medical AI is just a sort of amusement that can't be trusted.

    I wouldn't trust Watson as a sole source of medical advice either but in combination with the right doctor who knows how to examine and work with people, Watson would make a good tool. It takes a good personal interaction to get all the symptoms from individuals and if the doctor was running into something that he couldn't quite make sense of then Watson might make the connection between seemingly unrelated things to help put the doctor on track to a diagnosis.

    It's not a replacement for the doctor but a tool that the doctors can use when they've run into a wall they can't scale themselves. (that's a metaphor, by the way, for all those who will be trying to think of why doctors would be scaling walls to make a diagnosis. quit being so dang literal.)

  23. Re:Medicine more than matching symptoms to pills by Kensai7 · · Score: 4, Informative

    As a medical professional (neurologist-in-training, so I know about pain and "pain") these stories make me mad.

    There is no such thing as "healing with the hands" if you had a serious limb asymmetry in your hips. If it was mild, it could be corrected with the right shoes and postural exercises to teach you stand the right way. If it was serious, you should have seen an orthopedic surgeon to correct it in a surgical way. If he*fixed* you just by touch the right spots, then you probably didn't have almost anything physical in the first place and most of your symptoms were in your mind.

    BEAR WITH ME! I'm not trying to play down your pain and how you felt it, I'm just explaining to you in a rational way that many diseases and maladies are sometimes psychosomatic in origin and extension. I don't imply you are crazy or anything like that. I only say that you just wanted some hands-on caring, you didn't have anything really serious (organic) going on. And that's good news.

    Just don't waste too much money on alternative treatments. If an alternative treatment works, then probably help from a family or friend works as well. You don't need a professional. But don't take any chances.

    --
    "Sum Ergo Cogito"
  24. Who will you sue if there's a failure? by roger_pasky · · Score: 2

    Will it be a hospital failure? Wil it be IBM the responsible? Maybe the one who earned money with it...

  25. Who pays for this, and other pressing concerns by Demonoid-Penguin · · Score: 2

    Last time I checked IBM shares were a good investment, and opening a medical practice was an expensive proposition.

    I suspect access to Watson will be something that IBM profits from (which is good), and that it won't reduce the costs of running a medical practice.

    Don't be surprised if pharmaceutical companies "sponsor" Watson for medical practitioners.

    Would you like a bowel resection with that haemorroid removal and fissure stitch? May I recommend the fat reducing asthma medication trials? Perhaps sir would like to try our discount-medication-for-pharmaceutical-research-program??

    Wait for the "but sir requested the penile reduction, crackle, hiss, my programmer desires you wife, crackle...

    How about - "I'm sorry sir, but I must halt your heart surgery due to an injunction granted in East Texas by SCO-rebooted"

    But wait, there's more! Nintendo's recently announced homeopathic robot, Poirot, faces a patent challenge from Microsoft's mobile acupuncture and moxybustion robots- Pricks and Burns!

    :-D