George Rebane
Wired magazine has a very lame article on what should be an exciting report - the advent and functionality of clinical AI systems that are entering the healthcare service. ‘Paging Dr. Watson: Artificial Intelligence As a Prescription for Health Care’ authored by Brandon Keim paints an inconclusive ‘on the one hand, … but then on the other hand’ picture of how these AIs will perform. Reader motivation and understanding is tied to the now famous Jeopardy victories by IBM’s Watson supercomputer. But I’m sorry to report that Keim’s understanding of the power of these AIs and how they will serve is severely limited.
Once people become educated on the statistics of the complexity of the human system, its maladies, and medical errors encountered in human-only healthcare delivery (about 200K annual deaths attributed to ‘medical mistakes’), it will not take a leap of faith to conclude that physicians attending patients ‘a capella’ of AI will be rejected by those patients who have a choice. And as the Dr Watsons (equivalent and more powerful medical AIs) become available, if not ubiquitous in the industry, then insurance rates in great measure will reflect the advance.
Once people become educated on the statistics of the complexity of the human system, its maladies, and medical errors encountered in human-only healthcare delivery (about 200K annual deaths attributed to ‘medical mistakes’), it will not take a leap of faith to conclude that physicians attending patients ‘a capella’ of AI will be rejected by those patients who have a choice. And as the Dr Watsons (equivalent and more powerful medical AIs) become available, if not ubiquitous in the industry, then insurance rates in great measure will reflect the advance.
What Keim misses is the obvious power and glory such Dr Watsons will bring to the physician’s office and hospital bedsides. As an extension of the human clinician, Dr Watson will prevent the psychological pitfalls (e.g. anchoring, limited memory, poor deduction ability, …) which also afflict the physician, often with horrendous consequences for the patient.
Without making this a tutorial on how large data probabilistic reasoning systems work, let me just point out one obvious use of a Dr Watson. Working with the attending physician, the AI will be able to take existing diagnostics, ask the human more questions, prescribe more tests where necessary, and construct an extremely informative spectrum of potential maladies giving the probability values for each for various outcomes of the prescribed tests. Additionally, it will overlay the (Pareto) optimal trade-off of additional costs to uncertainty reductions in the most morbid/mortal diagnoses. This will allow the patient and the physician to make knowledgeable decisions on how best to proceed to a correct diagnosis.
Dr Watson will be able to explain itself in meticulous detail to the physician on the reasoning behind its assessments. In this the AI will use not only established clinical practice and published research, but also have the possibility of presenting original research done in the quest of the current case.
After the diagnosis, Dr Watson will present recommended prescriptions for treatment ranging from the established ‘standard of care’ to novel approaches buried in clinical literature and/or synthesized treatments never before encountered/considered. The informed alternatives will be presented in various formats to include normal English speech, printed reports, and interactive graphics. At this stage of introduction of these Dr Watsons, the human physician will mediate every step with the patient, therefore alleviating any and all concerns about the AI flying solo in the patient’s care.
I predict that it will not take long after that before the first stage mediator will be someone with less training than an MD. This will be possible because the AI will bring the world’s medical knowledge to that mediator who will be trained to bring in an MD should that be required. These machines will expand the availability of state-of-the-art healthcare at tremendously lowered costs to millions of people who today have few alternatives to waiting in line for an overworked and error prone physician or nurse practitioner.
What are the major hurdles to such a healthcare future? They are the AMA, Congress, FDA, and, of course, the entire legal profession – in short, government and the unions. But as far as the technology is concerned, fasten your seatbelts, you ain't seen nothing yet.
Without making this a tutorial on how large data probabilistic reasoning systems work, let me just point out one obvious use of a Dr Watson. Working with the attending physician, the AI will be able to take existing diagnostics, ask the human more questions, prescribe more tests where necessary, and construct an extremely informative spectrum of potential maladies giving the probability values for each for various outcomes of the prescribed tests. Additionally, it will overlay the (Pareto) optimal trade-off of additional costs to uncertainty reductions in the most morbid/mortal diagnoses. This will allow the patient and the physician to make knowledgeable decisions on how best to proceed to a correct diagnosis.
Dr Watson will be able to explain itself in meticulous detail to the physician on the reasoning behind its assessments. In this the AI will use not only established clinical practice and published research, but also have the possibility of presenting original research done in the quest of the current case.
After the diagnosis, Dr Watson will present recommended prescriptions for treatment ranging from the established ‘standard of care’ to novel approaches buried in clinical literature and/or synthesized treatments never before encountered/considered. The informed alternatives will be presented in various formats to include normal English speech, printed reports, and interactive graphics. At this stage of introduction of these Dr Watsons, the human physician will mediate every step with the patient, therefore alleviating any and all concerns about the AI flying solo in the patient’s care.
I predict that it will not take long after that before the first stage mediator will be someone with less training than an MD. This will be possible because the AI will bring the world’s medical knowledge to that mediator who will be trained to bring in an MD should that be required. These machines will expand the availability of state-of-the-art healthcare at tremendously lowered costs to millions of people who today have few alternatives to waiting in line for an overworked and error prone physician or nurse practitioner.
What are the major hurdles to such a healthcare future? They are the AMA, Congress, FDA, and, of course, the entire legal profession – in short, government and the unions. But as far as the technology is concerned, fasten your seatbelts, you ain't seen nothing yet.
First off, I'm not sure there are going to be any "hurdles" as you state. What would these hurdles look like? Neo-Luddite laws and regulations that ban, say, Watson-esque technologies from the workplace?
You can't put the cat back in the bag once he's out.
> I predict that it will not take long after that before the first stage mediator will be someone with less training than an MD.
That's the idea George. In fact, there might come a time when your average citizen might be able to reasonably self-diagnosis him/herself. This would be quantitatively different than the kind of WebMD.com diagnosis we currently have to suffer through from our neurotic family and friends.
The real news here is that people already expect this kind of service, even though the current slew of toy technologies come no where close to being as useful as Watson. The dazzling, yet frankly useless (and stupid) iPhone Siri comes to mind. Imagine a cloud-based (multi-tenant/shared) version of Watson compared to Siri.
Posted by: Ryan Mount | 19 October 2012 at 07:12 AM
Studies in the late 90s discovered that patients were more honest answering medical questions on the computer than they were to a person, in this case the Doctor. Adding the AI component should improve patient care.
Posted by: Russ Steele | 19 October 2012 at 08:47 AM
RyanM 712am - From your mouth to God's ear. I may be wrong in my apprehensions, but over many decades I have seen lawyers and unions successfully suppress technology advances for their own narrow self-interest. In today's world of growing structural unemployment, this tendency will be stronger than ever. Call it the 'John Henry syndrome'; opposing the machine has an established constituency.
http://rebaneruminations.typepad.com/rebanes_ruminations/2012/06/the-john-henry-law.html
You (and Russ 847am?) need not labor excessively to educate me on the advantages of new technology in our lives. I have spent my entire life on the bleeding edge, and have been rewarded handsomely, if not uniformly ;-), for my efforts. There is no doubt that an educated cohort of us will demand AI-augmented medicine, including the ability for 'self-diagnosis' and certain kinds of self-care. It would be utterly foolish not to do so.
But I admonish all to never underestimate the power, persistence, and perfidy of collectivists to have their way with us - aka, to f@%k us over - especially when they marshal wholesale ignorance in their support.
Posted by: George Rebane | 19 October 2012 at 09:11 AM
A local company developed the instrumentation and computer software to monitor sewer and water plants for cities and towns enabling the pumps to send message they were about to fail and let the maintenance team to take corrective action before the plant failed. It had some AI components to analysis the vibration signatures of the pumps. The Unions raised such a fuss that the cities would not buy and install the product. It would have eliminated a lot of monitoring jobs. The only market the company could find was overseas making portable sewer plants that were automated. UC Davis developed an automated pot hole detection machine that could find and fill the void in the concrete, before the pot hole became a road hazard. But the Unions did every thing possible to prevent the purchase of these machines, which would significantly reduce the road maintenance work force. Evidence the George's concerns are well founded.
Posted by: Russ Steele | 19 October 2012 at 09:33 AM
George, I certainly wasn't questioning your qualifications on the topic. I was just speaking as somewhat of an insider.
IBM is targeting two specific markets for WATSON:
1) Heathcare, which you mention here
2) Financial Services
The current deployments of WATSON are essentially single-threaded: one huge hardware deployment that can accommodate one question at a time. Although that's changing quickly even as I type this.
It's no surprise to anyone that a good chunk of IBM's business is in the public sector. Most of it has aided in reducing the costs of government through increased automation. As I've stated here before, it would be a curious experiment to point WATSON at something like Medicare Fraud auditing to see what it comes up with. But perhaps that's a ways off because our government is currently in the "dig a hole so someone can fill it up" mode.
Posted by: Ryan Mount | 19 October 2012 at 10:55 AM
RyanM 1055am - gratefully understood. And I consider your sharing IBM's latest announced penetrations of Watson and Watson++ into markets a boon to these pages.
Nevertheless, identifying and highlighting the 'John Henry syndrome' (my 911am) is a worthy item of discussion which I don't think that many liberal readers will participate in. And as Russ (933am) underlines, the examples of lawyers and unions impeding the adoption of new technologies are legion.
Posted by: George Rebane | 19 October 2012 at 11:18 AM
On the science front, Some interesting developments have been occurring
at the Hawaiian volcano. The "big" crater is coming back to life.
Just a few years ago web cams were installed for our viewing pleasure.
Things have taken an interesting turn in resent weeks. I do believe it will bubble over in the vary near future.
This can be observed from this link.
http://hvo.wr.usgs.gov/cams/
I sure wish I could be there to see this with my own eyes.
The ECO gang says that man can influence what this blue marble does.
Lets see if they can put a cork in this and make it bow to our will.
Posted by: Walt | 19 October 2012 at 12:31 PM
Never heard the word Luddite my whole life until yesterday and an old timer yesterday dropped the word on me explaining he would rather just smash his gadgets and new contraptions. But, his wife likes the gadgets and he is a self proclaimed Luddite: Here is a repost from Dr. Rebane's Luddite piece.
http://money.msn.com/technology-investment/c_gallerysidebyside.aspx?cp-documentid=250221329
Posted by: Billy T | 19 October 2012 at 08:00 PM
Brother Billy:
I have this hanging in my backyard next to my Weber:
http://images6.fanpop.com/image/photos/32300000/never-underestimate-the-power-of-stupid-people-in-large-groups-random-32353939-468-312.jpg
Posted by: Ryan Mount | 19 October 2012 at 08:12 PM
Daughter has been in favor of more P.A.'s armed with AI for years. There are only 5,000 or so folks in her specialty, and they do a lot of very, very, delicate sewing, which takes years to learn, and I rather doubt she'll be replaced in her lifetime by a machine. Now wasn't it Buckminster Fuller who pointed out that indeed technical advances could indeed allow for unprecedented leisure time for many, but silly him, he thought such breakthroughs would be shared by all, like Salk and the polio vaccine.
Posted by: Jesus Betterman | 19 October 2012 at 08:32 PM
Mr. Mount, good one. Correction: I meant to say I have never heard of the the Luddites until RECENTLY. Yesterday I met one and now that word seems to pop out of sundry and divers places. BTW, don't can me bro, bro in law :) Love, Brother Billy and the little High Steppers.
Posted by: Billy T | 19 October 2012 at 09:04 PM
Here is some insight in to the current AI thinking:
SAN FRANCISCO – “We don’t die because the laws of physics require us to die – we die because we’re not currently smart enough not to die ... Why settle for predicting human behavior when we can re-engineer the human genome? ... An exponential function is a multiplicative derivative ... Solving the problem of friendly artificial intelligence is the key to saving the world.”
Humanity’s inevitable future, according to the speakers at the 7th Annual Singularity Summit held this past weekend in San Francisco, is one in which everyone lives forever, poverty and violence are relics of the past, and the intelligence of man and machine fuse into one to usher in a new “human-machine civilization.”
Link for more here: http://www.policymic.com/articles/16546/human-immortality-singularity-summit-looks-forward-to-the-day-that-humans-can-live-forever
Posted by: Russ Steele | 20 October 2012 at 08:22 AM
RussS 822am - thanks Russ for that article. I want to add that Singularity Summit12 that was held last weekend has a website where more illuminating discussion is available.
http://singularitysummit.com/
Regular RR readers should be familiar with the main themes which have not changed much since the first Singularity Summit in 2006 which I attended. The major predictions then, as now, are all coming to pass.
Posted by: George Rebane | 20 October 2012 at 08:52 AM
PHIL BOWERMASTER: Robots Will Steal Your Job.
Some will argue that, based on history, only luddites and the technologically (or economically) ignorant fear a machine takeover of the full human employment space. The “threat” of automation has never panned out. So far. Then again, machines are getting smarter and faster at an exponential rate; we are not. So it could be that we will all end up being yoga instructors, experiencing a transition to a whole new economic model, one wherein we are increasingly employed in the lucrative field of value-added intermediation — filling the gaps the machines can’t.
On the other hand, machine capabilities may improve so fast that they leave very little gap for us to intermediate. Or there might be limited economic value in the gaps we can fill. The possibility of a truly painful transition is out there and it’s very real.
In the long run, however, there is an exponential trend that favors us. The overall human condition is improving exponentially, and has been for some time. Vast improvements in machine capability are just a subset of, and contributing factor towards, that overall trend. In fact, it may turn out that we need these machines — and their ability to do everything we can do faster and better — to keep that trend going.
Some interesting times lie ahead. Everything may work out great, but I think Pistono’s advice about having a backup plan makes a lot of sense. In fact, maybe a couple or three backup plans wouldn’t be out of the question.
The book is Federico Pistono’s Robots Will Steal Your Job, But That’s Okay.
Posted by: Russ Steele | 24 October 2012 at 10:57 AM
Analysis in the late 90s discovered that patients were more honest giving answers to medical care issues on the pc than they were to a person, in this scenario the Doctor. Such as the AI factor should improve personal care.
Posted by: BarryV | 18 November 2012 at 10:56 PM