I very much enjoyed Jerome Groopman's book, How Doctors Think. I love his writing at the New Yorker. But I am afraid I didn't think very much of the book's thesis, which is that doctors need to improve their clinical judgement rather than relying on evidence based medicines and statistics. "People are not statistics" is sloppy thinking; most of the time, we are. And there's substantial evidence that doctors do best when they treat their patients by the numbers.
Over at eSkeptic, Charles Lambdin voices the same criticism:
Groopman tells us he is troubled that new doctors seem to be trained to “think like computers,” that they rely on diagnostic decision aids and some seductive “boiler-plate scheme” called evidence-based medicine. Groopman’s position, when his various arguments are gathered and assembled, becomes untenable. He admits doctors suffer from innumerable biases that diminish the accuracy of diagnosis, reducing many diagnoses to idiosyncratic responses fueled by mood, whether the patient is liked or disliked, advertisements recently seen, etc. Thus Groopman agrees with decision scientists’ diagnosis of doctor decision making; but then he goes on to wantonly dismiss what many of the very same researchers claim is the best (and perhaps only) remedy, the way to “debias” diagnosis: evidence-based medicine and the use of decision aids. In place of statistics what does Groopman suggest doctors rely on? Clinical intuition of course, the very source of the cognitive biases he pays lip service to throughout his book.. . .
Most doctors do not like decision aids. They rob them of much of their power and prestige. Why go through medical school and accrue a six-figure debt if you’re simply going to use a computer to make diagnoses? One study famously showed that a successful predictive instrument for acute ischemic heart disease (which reduced the false positive rate from 71% to 0) was, after its use in randomized trials, all but discarded by doctors (only 2.8% of the sample continued to use it). It is no secret many doctors despise evidence-based medicine. It is impersonal “cookbook medicine.” It is “dehumanizing,” treating people like statistics. Patients do not like it either. They think less of doctors’ abilities who rely on such aids.
The problem is that it is usually in patients’ best interest to be treated like a “statistic.” Doctors cannot outperform mechanical diagnoses because their own diagnoses are inconsistent. An algorithm guarantees the same input results in the same output, and whether one likes this or not, this maximizes accuracy. If the exact same information results in variable and individual output, error will increase. However, the psychological baggage associated with the use of statistics in medicine (doctors’ pride and patients’ insistence on “certainty”) makes this a difficult issue to overcome.
The statistics vs. clinical intuition debate has ensued for decades in psychology. Where one sides in the debate is largely determined by what one makes of a single phrase: “Group statistics don’t apply to individuals.” This claim, widely believed, ignores many of the most basic concepts of probability and statistics, such as error. Yes, individuals possess unique qualities, but they also share many features that allow for predictive power.8 If 95% of a sample with quality X has quality Y, insisting that someone with quality X may not have Y because “statistics don’t apply to individuals” will only decrease accuracy. Insistence on certainty decreases accuracy. As Groopman himself says, the perfect is the enemy of the good.
. . .
Physicians who allow themselves to think in such discretionary ways can find “exceptions” everywhere they look, and, augmenting a decision aid as they see fit, will only end up lowering its overall diagnostic accuracy. Why? Because human beings do not apply rules consistently. Mechanical procedures always lead to the same conclusion from the same input. Doctors are subject to random fluctuations in diagnosis caused by judgmentally-irrelevant factors including availability, priming, recency effects, inconsistent weighting of information, fatigue, etc., all of which reduce accuracy. What leads to a correct decision for one case may not for another, and variables that contribute to the diagnosis made may actually be uncorrelated with it.
This is hardly restricted to doctors. Every profession resists being told that there is a standard way to do things, that a cookie cutter can cut better than their skilled hand. Journalists famously hate the "inverted U" style of writing a news story, even though it really does seem to work better than anything else; it's boring to write, and leaves no room for individual style. Teachers don't like "teaching to the test" or rigidly programmed phonics curricula, even though the latter produces measurably better results than all but the very best teachers. Unfortunately, for many of us, it may be time to welcome our new robot overlords.
For we non-journalists, what's the "inverted U style?"
There's another reason doctors aren't fond of evidence-based medicine; with the loss of discretionary authority, they also loose the ability to prescribe certain drugs or proceedures for which they're handsomely rewarded by drug and medical equipment makers.
The current US medical system is seriously broken because patients are generally not in a position to question their doctor's decisions. If you doctor says you need to take two of these pills every morning, how many of you would get a second opinion?
Drug and medical equipment salesmen take advantage of this knowledge gap by offering lavish gifts and outright bribes to doctors for proscribing their products. And why not? Most medical salesmen are compensated handsomely on commissions, and these prescriptions can sometimes bring tens of thousands of dollars per patient. And doctors are often enlisted in this scheme for their own person enrichment. Insurance companies are forever trying to prevent these practices by being scrutinizing every medical claim, which has the consequence of making it that much harder for all patients to get reimbursed.
This isn't only a matter of professional pride getting in the way of progress - it is also a medical system that is profoundly dysfunctional.
The current US medical system is seriously broken because patients are generally not in a position to question their doctor's decisions. If you doctor says you need to take two of these pills every morning, how many of you would get a second opinion?
Methinks this has everything to do with patient's laziness and nothing with lack of power to question the doctor. Seeing another one is just another $20 in copay and a week or six of waiting. So you won't do it because of a pill, but probably will do it if the doctor prescribes an operation. What medical system is likely to give you different results? I guess, the one where doctors are SO BAD you'd be ill advised to believe anything they say until you gather some statistics ;-)
At least decision aids offer the potential for vastly cheaper doctor's services. All we need is a way to change the structure of the whole system.
This certainly depends on the quality of the algorithm. Here's a simple algorithm. Say I'm a doctor. No matter why a patient comes to me, no matter what the symptoms are, I'll tell the patient there's nothing wrong with them. It's an algorithm! As it says, the same inputs always lead to the same output! This algorithm, however, is crap, and does not 'maximize accuracy.' Perhaps the quality of the currently available algorithms is quite good; the simple mechanical nature of them, however, is not a strong selling point.
The quality of the measurements here is the subject of much debate. People object to 'teaching to the test' even though they know that doing so can raise test scores. I think the primary objections have to do with (a) teaching only the material on the tests and (b) teaching students only to take these tests rather than to think critically in a variety of contexts.
Max wrote:
"Methinks this has everything to do with patient's laziness and nothing with lack of power to question the doctor."
Patients are lazy? That's just silly. We go to doctors because they know more and have been trained in a medical science that most of us don't know. This isn't like trying on a pair of shoes, where I can easily judge which ones will best fit my foot. It is precisely the (dare I say it?) asymmetric information that enables abuse in the system. Saying if a patient is too lazy to question their doctor who just told them to get an MRI is like saying we don't need meat inspectors because eventually people will stop buying the brands that kill them of food poisoning. It is true, but most of us would want to live in a world were a little bit of regulation would prevent a lot of harm.
And anyway, the medical system is generally not one in which price cues can help us, because as you point out, a $20 copayment gets you a visit to another doctor (never mind that the second visit is WAY more expensive than the $20 you pay, and possibily just a waste of everyone's time if you can't trust their opinion either). The cost is born by all of us in the form of higher insurance premiums.
Do you even understand how and why a free market works? It depends on information flowing among all parties so they can make rational decisions. When you deprive certain parties of relevant information, the system isn't going to create economically efficient outcomes. Certain relevant information like that your doctor will get a free vacation to Hawaii if he has 30 patients who get an MRI scan. The doctor isn't going to tell you this. The MRI equipment salesman isn't going to tell you this. And who else would know?
The "patients are lazy" mindset is empty libertarian posturing falls flat when confronted with reality.
Max wrote:
"What medical system is likely to give you different results? I guess, the one where doctors are SO BAD you'd be ill advised to believe anything they say until you gather some statistics ;-)"
How about a medical system in which doctors are not economically incentivized (corrupted) by drug peddlers? How about a medical system in which doctors are are compensated based on quality of care? Neither of these criteria applies to the US system. My own preference would be for a Singaporean-style mixed system.
I'm glad to hear someone make this important point: a lot of professional practice has more to do with the professionals' need to feel important and intellectually engaged than with the delivery of the best service. Across the board, individuation and novelty are prized in the face of overwhelming evidence that generalization based upon algorithm and already-existing tools are more successful. Some of this comes from a bad incentive structure (e.g. in academia, where professional advancement comes from "advancing knowledge" even if it's spurious, while there is no professional benefit to debunking that spurious knowledge), but a lot of it just comes from professionals' desire to lead interesting and well-paid lives. The fact that this desire is completely understandable and legitimate makes the problems that it spawns very hard to tackle.
Going off on the teaching/tests tangent. The key point there, and perhaps can be extrapolated to your other examples, is that when you devise a system that must be measurable, it's no wonder that you'd inherently pick the practice that produces the best measurable results.
What if you pick the wrong thing to measure? For example, there's been great emphasis on teaching children to read. Just read. They can read. As in recite the words on the page. But there are very very few students in schools that I know that actually comprehend what is on the page. The words just get regurgitated. Ask them what it means or to put it in their own words and chances are they just recite what they read only backwards. They didn't comprehend it. So here we are teaching and testing phonetics to make sure people can read and many don't even understand.
Trying to break down life and into little measurable parts does a great job (if you get everything right) of showing you a shadow of what success should look like. But it will never be true success, only a close-but-not-there-yet representation of it.
We're obsessed with measureable things but that reliance has led us astray. Measurable results are good. But they are not THE THING that is important. I'm struggling with words here because if I could do a better job it, then fixing the problem would obviously be easier than it is. But it's clear to myself and many others than the focus and emphasis that we place on so many things is awry.
Sorry for the long one..
I agree with the need for better statistics. But statistics only count if you have good data collection in place?
We have yet to take full advantage of all the data that hundreds of millions ill people in our society could potentially produce every day FREE of charge. Here a teaser:
______________________
Doctor: What do you eat?
Patient (overweight, diabetic): Oh.. only healthy things. Like whole-wead bread and cereals although they do not taste that great and organic lean meats and lots of salads. (But I do not mention all the café late insulin injection at Starbucks, it’s just coffee, the butter-fat insulin candy in between because I do hide myself when consuming that – and if nobody sees it – it does not happen… I also do not mention that I do NOT eat a lot of the healthy stuff because I do eat this in front of others and I do not like eating in front of others..).. oh – and the occasional Pizza slice with pepperoni and cheese – but only once or twice a week..
Doctor: Hmm… could be a genetic predisposition? Any family who has weight problems?
Patient: Yes – all of them!
Doctor: Hmm… we have recently identified two molecules in mice, known as GATA-2 and GATA-3, which appear to regulate the process by which fat cells are created. I think that in your family these molecules might be overproducing. We are currently working on a treatment and already have some great results in mice.. but I suggest you start exercising more.. That means – drive to gym and then sit on a machine there for hours – feeling rally bad compared to all the other healthy people around you – knowing that you can never be like that. Afterwards you can reward yourself with a treat. That will help you until we have a magic pill!
_____________________
My back-hand sucks on the tennis court so I take my son and lock him up on the toilette and force him to practice swings. What a great scientist I am – I get satisfaction knowing full well that by doing so I could potentially gain more insights than by doing nothing at all. Those few and expensive insights would be out of context to me in the real world – but who cares – at least I do not get my feet dirty and can cash in nicely, claiming I want to help improve the humans condition ...
No – this is too sad to spend a life on? Every day - millions of humans visit doctor's offices due to some health concern. Rarely are they asked to document their life-style and diet PROPERLY (they are asked, if at all, on a good faith basis which is always distorted). “Scientists” much rather lock themselves up and try to reengineer human ills in rats. As a result - relevant population studies are rare and expensive (Nurse studies etc). It is ONLY because such studies are currently very difficult that the tobacco industries were able to distort the connection between smoking and cancer for so many decades.
This could change easily thanks to modern technology. When it comes to e.g. nutrition - one can easily document everything one eats with tools such as www.nutritiondata.com?
(btw Turns out a Kiwi has more essential amino acids than a beef stake!!!)
Why not link a computer system to your credit card and every time you buy something as Safeway, Wal-Mart or Wholefoods – bang – all your sugars, fats, carbs, amino acids, minerals, vitamins are stored and analyzed automatically. You do not have to think about it anymore. If you buy cigarettes it will be taken into account too…
Voila – after only a few years or months and the world changes from opinion to fact, from mice to men. Fact – that most who have been trying to lose weight for years simply eat more calories than they burn (90% of overweight people think that they are genetically predisposed to be fat which is true for maybe only 0.3%? It is an addiction or other psychological disease but not a bodily one except for some VERY few). Most humans eat too much saturated fat (butter fat in sweets, burgers, cheese, etc), too much sucrose or chemicals instead of fructose, consume too much insulin via dairy, too little vitamins and minerals (not enough uncooked or unprocessed fruits and veggies)… Vegans know this – but an average Joe must see it black on white… (and by average Joe I mean the average doctor who has, lets face the sad status quo, less brains and passion for what he does than a car mechanic?)
So – who wants to fund my new Consumption-Monitoring Software start-up. We will connect the retailers with the medical authorities and provide everybody with real-time accounting instead of guesses and estimates. What if Americans already consume 3 times more saturated fats than what the American Dietic Association claims is healthy? Should we really subsidize saturated fats agriculture with tax money (ca $20 billion per year)? What if it turns out that not a single school in the US offers a menu that is in line with the food-pyramid or any other dietary recommendations but is financed with tax money? Recently there was an interesting experiment with the short, yet dangerous, Atkins diet craze – let’s be the only ape on earth that eats meat and milk as a stable and avoids as many vitamins and minerals as possible…
There are only 5000 Bonobos left on the planet and not many more Gorillas and Chimps... But there are 6 billion humans with hundreds of millions suffering from some illness. This fact alone should lead us back to Hyppocrates pronto and to leave non-human animals alone? If anything we need apes in the wild and not the laboratory – because there we could observe the natural diet of apes better than anywhere else where culture has overtaken nature (for worse or better).
Stop bugging our phones and help us understand why our life-expectancy is about to decrease for the first time since decades – that would save more Americans?
Of course, treating patients like statistics has one major advantage to doctors... it reduces the likelihood that they'll be sued for malpractice. If you can show that you treated your patient according to the acceptable standard of care, that is a very strong defense. If you deviate from that however, you're opening yourself up to liability.
Well Sam, a whole lot of things are measurable. Leave education aside for the moment and consider the easy cases, like chess. Chess programs can beat Kasparov. Is that success? Or only a "close-but-not-there-yet representation of it"? Or to take the passage Megan quotes: if you are diagnosed, falsely, for acute ischemic heart disease, because your doctor refuses to use a machine that can outdiagnose him, perhaps you prefer "true" failure. Me, I'll take the close-but-not-there-yet-representation of success that the machine provides, and an accurate diagnosis.
In fact the bias runs against the measurable. People absolutely hate to regard themselves as anything but unique. Just look at the general hostility to insurance, which is the business of treating people as statistics. This blog's comments section provides no shortage of examples.
The term "robot overlords" grossly exaggerates the situation. Robot "assistants" is far closer to reality. Expert systems can very rapidly sort through all of the potential answers and provide guidance regarding all potential diagnoses, including assessing the probability of each potential diagnosis being correct. The expert then can rapidly proceed to determine the actual or most likely cause. Expert systems can greatly reduce the need for experts, since much of the diagnostic process is automated. The expert systems can also help assure that potential problem solutions, or critical steps in a process, do not get overlooked.
The important distinction is between acquiring skills and learning to use those skills. In the case of reading mentioned above, the ability to read the words on the page must generally come before the ability to understand the meaning of the message they convey as you read them. Similarly, applying logical processes to problem solving cannot proceed a knowledge and understanding of logical processes at some level.
Concerns about "teaching to the test" have always confused and frustrated me. Life is a test, every day; some pass, some fail, some excel. If the capacity to read at an 8th grade level is essential to success in life, testing to assure that this has been achieved seems perfectly reasonable. The same is true of basic mathematical skills. Word problems are reasonable in tests, since life is a series of word problems, frequently with multiple choice answers. Testing of each and every student will occur, either in school or in the workplace. The consequences of not testing well in the workplace are generally more serious and longer lasting.
The link below is to a physicians blog. He likes software diagnostic aids and the article explains why.
.
Unsolicited Endorsement: Pepid
szr wrote:
How about a medical system in which doctors are are compensated based on quality of care?
This need not have a government component to it. In virtually all markets, payment is withheld if the service is incompetently performed. I would never buy rotten produce from my grocer, or pay a mechanic who made my car worse.
Unfortunately, I wouldn't bet on being able to talk a doctor into treating me on the condition that if he doesn't fix my problem he won't get paid. I suspect one big reason for this is the doctor's monopoly power (via the AMA and licensing). Wherever you find monopolies you usually find producers unresponsive to their consumers.
Megan, as a good sort-of libertarian, I am surprised you don't observe the role that compulsory licensure laws play in all this. If somebody wants to start a hospital which uses nurse practiioners and expert systems applying mechanical statistics-based approaches, well, they can't.
Interfering with the market prevents people from using science-based approaches to compete with the High Priests.
In my experience, many doctors are very poor at diagnosing problems because they are not taught to be very rational, systematic and objective. This is a real failing of medical education,and is exactly what a diagnostic aid could provide.
I have had two cases in my life in which doctors badly misdiagnosed problems that I had. In both cases, I found out later that they "diagnosed" a problem that they themselves had had, and so were rather fixated on. When I investigated myself, I found that the symptoms did not even come close to squaring with the diagnosis.
What doctors should be very good at providing is accurate "data" to feed to the diagnostic aid. (Where does your foot hurt? What time of day does it hurt? What is the pain like: sharp, dull, stabbing, throbbing? What makes it worse? What makes it better?) But, doctors seem to be very reluctant to do this amount of digging. If my doctor had asked these questions, and been objective about the answers, he would have known that my problem was not plantar faciitis.
Some areas of medicine also suffer from serious inertia which could be helped by decision aids. We knew log ago that clot busters like TPA or streptokinase were very important treatments for heart attack - but getting to the point where a patient who needs them could be reasonably certain of getting them in a timely fashion has been the work of decades. I don't know why there aren't more malpractice suits on this kind of thing. Other specialties seem to do better.
I love all of the socio-political one-up-man-ship that substitutes for debate.
Evidence based medicine and statistics are useful tools when employed using common sense gained from clinical experience. To argue either-or is nonsensical.
Models, based on mathematics, physics, chemistry, biology and statistics are incredibly useful tools only if the Doctor using this diagnostic model knows exactly how it works and more importantly knows when it does not work.
If you think docs should only rely on robot diagnoses, then go fly in an airplane flown by a computer. Of course, we all enjoy the safety of computer assisted flight, however, no one would get on an airliner with an IPOD at the controls.
The best docs (and other technical professionals) use all of the best tools in combination with their clinical (or "field") experience derived common sense.
Uncle Bill - Amen!
The USDA is also not innocent? How come that "farmers" are allowed to determine what the food-pyramid should look like in the US? There is a conflict of interest? (Same with the school-program?)
"inverted U"? Wouldn't that be an "n"?
Journalists - always trying to make their job more impressive-looking!
"we are statistics" just insofar as we are; and, on many dimensions, we aren't. The evidence that "doctors do best when they treat patients by the numbers" is just that those particular numbers improve when doctors focus upon them. The extrapolation to "health care quality" in a comprehensive sense is a non sequitur.
Decision aids are used in treatment, not diagnosis (in general) and there are many cases when individual patients don't fit. What we need is not ebm/decision aids rather than physician discretion but physician discretion making judicious use of all available decision aids/ebm-type imperatives judged to apply to the individual case. i suspect that this is what Groopman says (havent read the book).
The chorus calling for objective measures as a sufficient (not just necessary) indication of quality in any human activity gets louder and louder. Im sorry to see Megan adding her voice to it.
I think it's important to point out one significant difference between human doctors and machines: the doctors have the capacity for narrative. I don't know much about the algorithms that are used for mechanical diagnosis, but I suspect that many of them are based on quite simple statistical regression techniques -- "In the presence of symptoms x and y, but absent symptom z, the correct diagnosis is [blah] with probability 0.78." This is the sort of statement you can make after collecting a bunch of data and mining it for patterns among variables.
These sorts of approaches definitely have their strengths. For example, one area in which they are likely to outperform many humans is in assessing the significance of correlated variables. A human doctor might observe five or six different symptoms, all of which point to a certain diagnosis, and treat this collective evidence as an overwhelming indication that the suggested diagnosis is correct. If those symptoms are highly correlated, however, we shouldn't consider them as independent data points -- in the extreme case of perfect correlation, we've actually only seen one "metasymptom", and this should affect the way that we treat any other symptoms that contradict the diagnosis for the five or six that agree. It can be tricky to correctly account for such correlations among many observed variables, and this is definitely a strength of statistical models.
BUT -- There are substantial weaknesses in the diagnosis-by-correlation approach. Do the symptoms make sense in the broader context of the patient's medical history and in light of any recent lifestyle changes? If some symptoms are correlated empirically, *why* are they correlated? The statistical model probably doesn't have a clue, yet this sort of reasoning might lead an experienced doctor to identify the underlying mechanism for a disorder, rather than just reach a boilerplate diagnosis.
In the end, I think we have a false dichotomy between "letting doctors rely solely upon their natural intuitions" and "letting smart statistical algorithms replace human doctors". Any enlightened form of modern medicine ought to be informed by statistical inference, because of the acknowledged biases in human judgment and because large, representative datasets offer an unprecedented opportunity for us to learn from the past. But the machines will never be perfect, and I don't think we want to cede decision-making power to them.
The real goal should be to treat statistical algorithms as important tools that can inform, but not replace, doctors' broad interpretive abilities. If a doctor knows how the statistical model works, she can effectively consult it: learn how the variables are correlated, be presented with alternative diagnoses that might not have sprung to mind, and have immediate access to a concise summary of many lifetimes' worth of hard data. And, if the doctor knows the assumptions that the model is making, she can confidently reject counterintuitive statistical diagnoses where it is clear that the model has missed some key aspect of reality.
Maybe this is just a utopian fantasy, and we cannot hope to produce statistically competent doctors who synthesize algorithmic output with the mechanistic knowledge and human empathy gained through medical education and practice. I think this is the best possible approach, however, and our debate should focus on how to get as close to this ideal as we can, rather than whether we should replace our doctors with robots.
This seems a little silly. It’s only a shoot-from-the-hip guess, but I would think most people in most professions would do better by following preordained decision trees. The few pros who beat the algorithm (because they recognize some incompatibility between the situation and the model) will be eliminated, as well the many who would do worse.
I think there is a Malcolm Gladwell thinkpiece somewhere regarding an algorithm for choosing hit movies and popsongs.
The solution, which would appeal to a libertarian-minded journalist, is to collect and publicize the success rates of doctors. By way of illustration, even if one posited that the vast majority of stock-pickers would do better by using some algorithm, we would be worse off if all investors used the same model. Areas of imperfect information wouldn’t get flattened out if people weren’t always trying to exploit them.
It’s a minor paradox that the algorithm in these cases can’t improve unless doctors depart from it from time to time. Patients may prefer a doctor with the ineffable power of a magic white coat to someone following a rigid decision tree; but I’d wager a great many patients would choose a doctor with consistently better outcome stats even if he “cheated” using a computer to get them.
Megan, I'm sure you meant the inverted pyramid, not the inverted U. Right?
I've done a fair amount of work in artificial pattern recognition. You take a digital image or signal, feed it to a computer and then, via nested decision making or correlations a decision is made as to what it is. It is pathetic compared to the human eye and brain in most cases. You show me a picture of a bike, and I know what it is because I've seen a million of them. There are exceptions.
I think the parameter space of medical conditions is too broad for analytical tools to come close to replacing doctors, but that doesn't mean they are not useful. Strictly speaking, for an algorithm to diagnose a patient with strep throat, it would need to rule out a broken leg, leprosy and athlete's foot. The doctor is way ahead before the algorithm starts. A competent doctor starts out 15 questions ahead in a game of 20 questions. They just might have trouble formulating questions 19 and 20 sometimes.
"One study famously showed that a successful predictive instrument for acute ischemic heart disease (which reduced the false positive rate from 71% to 0) was, after its use in randomized trials, all but discarded by doctors (only 2.8% of the sample continued to use it"
This might be deceptive. What does this test do for false negatives? A 1% increase in false negatives might outweigh eliminating a 71% rate of false positives. Maybe the test decreases false negatives too, but we don't know.
why has this turned into a Big Blue vs Kasparov debate?
the problem with many medical authorities and universities is that they look at their field like an engineer would and not like an evolutionary biologists... If they were to think more like evolutionary biologists - there would be a natural tendency to document and analyse real life events and to put them into statistics and eventually context... more so than under the current world-view of "doctors"?
(not necessarily rigid decision trees..)
Such a paradigm shift is being attempted by some Harvard MDs and supported by eminent biologists such as E.O. Wilson and Richard Dawkins.
http://www.amazon.com/Why-We-Get-Sick-Darwinian/dp/0679746749
There would still be differences of interpretation among the medical community after such a paradigm shift - but we could nevertheless deploy technology better and therefore document events better that have nothing to do with technology per se... but this is a long battle.
As many have pointed out - there are no real rankings regarding quality of health care. And this is due to the transparency problem.. who and how could really judge how good a treatment is compared to all possible options?
I do recall a Chinese story about the medical adviser of Confucius. He was a very famous doctor and has cured a lot of royals. When asked if he considered himself the best doctor in the country
he replied: "No. That honour probably goes to my brother who lives in a small village. Nobody has ever heard of him because he has never cured anybody. People in his village never get sick!"
Before we get to the point of perfecting the expert systems that doctors would use to help diagnose patients, we need to get them to understand that it might be possible to better than they are.
When my wife was pregnant with her first child, I was concerned about episiotomies. After her OB blew off my concerns and said she'd do what she thought was necessary unless we specifically told her not to, I asked how many episiotomies she typically did. She had no idea.
I don't claim to know more than she does about delivering babies in general, but I'd read studies done by doctors at hospitals that concerned me. She was more interested in asserting her superiority than in discussing a real, controversial issue with a concerned husband of a future patient.
I have a friend who is an anasthesiologist and they are one of the few specialties that keep track of their outcomes in a central database.
Most specialties don't.
EI
I've found diagnose-me.com to be quite helpful. Click on my name to check out one version of what an expert system might look like.
Off topic to Uncle Bill: what was your foot diagnosis?
The arguments for doctors deviating from decision trees sounds just like the advocates of active management of investments.
We like to have control, especially over critical parts of life such as our health and wealth. Very few people like to think of themselves as below average, and for very intelligent and capable people the "acceptable" ranking is remarkably higher. You don't just want to be in the 99th percentile of the population, you want to be in the 99th percentile of doctors or money managers. In a world of 6 Billion, even those of us on the very far right hand edge of the bell curve aren't special flowers.
People still buy actively managed mutual funds instead of a few big, cheap, passive ETFs because that they believe that they can pick the "right ones" rather than relying on the math. Similarly you want your doctor to be hands on and a teacher that is creative and amazing. Unfortunately 75%+ of the time you get a better result by accepting your mediocrity and that of your service providers. It hurts cognitively and is damn hard for intelligent and successful people (especially) to accept, but it gives the best expected outcome. Of course if you get a crappy outcome, you'll blame the mechanism, rather than yourself or accepting that there is risk from all things. Someone must pay, and you will higher a lawyer to fight the laws of thermodynamics if you have to.
As to teaching to the test: it's worse if your child is in the top 5% and their teacher is in the top 5%. So you've got a 99% shot of a better outcome from teaching to the test. The other point is that teachers come from weak majors and low marks. Ed schools are abysmal and focus on PC pap rather than on any real utility. Expecting some 65% percentile who couldn't manage AP work or calculus in highschool to do wonders with your child is grounds for committal. This goes at least double if you abuse your children by sending them to public school.
hutkeeper listel protransfer transactional culmiferous nonfloriferous cankerworm effectful
http://www.toss.co.za >Two Oceans Slope Soarers
http://www.metalsoaps.com/
hutkeeper listel protransfer transactional culmiferous nonfloriferous cankerworm effectful
http://a.webring.com/hub?ring=fishgeekscom&id=1&hub >FishGeeks Web Ring
http://pilat-nai.com/