February 02, 2009
Old Style Business Models Slow Medical Advances

A couple of articles in the New York Times draw attention to business models in medicine that slow the rate of improvement in medical service delivery.

Two main causes of the system’s ills are century-old business models, for the general hospital and the physician’s practice, both of which are based on treating illness, not promoting wellness. Hospitals and doctors are paid by insurers and the government for the health care equivalent of piecework: hospitals profit from full beds and doctors profit from repeat visits. There is no financial incentive to keep patients healthy.

“The business models were all created decades ago, and acute disease drove those costs at the time,” says Steve Wunker, a senior partner at the consulting firm Innosight. “Most businesses in this industry are looking at their business model as entirely immutable. They’re looking for innovative offerings that fit this frozen model.”

Why have old business models lasted so long in medicine? It seems hard to price wellness maintenance as compared to pricing procedures and consultations. How to incentivize individual doctors to keep patients healthy? It is a lot easier to say it is a worthy goal than to describe a system for doing it that would work financially. Anyone have suggestions along these lines?

I would like to see far more automation of diagnosis. This requires wider spread use of electronic medical records so that the data which medical expert systems need will exist in electronic form. It also requires an economic model for medical care that provides incentives for automation. Medical expert systems can make better diagnostic decisions because the huge and growing quantity of medical test results and the large number of diseases and treatments really test the limits of the human mind to process all that information. Medical expert systems can free up smart doctors to do more original creative work such as medical research and product development.

Most doctors in private practice still do not use electronic medical records systems, making them outliers in a world where a very large fraction of all high information work is done using electronic information systems. Digital medical records make the discovery of better medical practices possible.

The Marshfield Clinic, a large doctors’ group in Wisconsin, shows that computerized records can indeed improve the quality and efficiency of medicine. Yet the Marshfield experience suggests that the digital record becomes truly useful only when patient information is mined to find patterns and answer questions: What treatments work best for particular categories of patients? What practices or procedures yield the best outcome?

This group of doctors have used their medical software system to help cut total costs by allowing them to manage diabetic care more efficiently.

From mid-2004 through the third quarter of this year, the percentage of the clinic’s diabetic patients with blood cholesterol at or below the recommended level rose to 61 percent, from 40 percent earlier. The percentage with satisfactory blood pressure increased to 52 percent, up from 32 percent.

Over the same span, hospital admissions among Marshfield’s diabetic population fell — to 311 per 1,000 patients a year, from 360. Because a hospital stay for a diabetes patient ranges from $8,000 to $22,500, according to national statistics, Marshfield’s results translate into an annual cost saving of $7.3 million to $20.5 million.

But for the average private practice a reduction in the hospitalization rate of patients isn't going to boost revenues. The money saved probably all flows to insurers. The system lacks incentives for most medical providers to go after these forms of savings and care improvement.

A medical marketplace which rewards use of expert systems, electric patient records, and reduction of total costs by use of more effective and productive methods of purveying treatments is what we need. How do we get there?

Update: More use of computer systems in medicine improves outcomes.

A study published this week in the Archives of Internal Medicine suggests that broad adoption of IT systems may provide significant health benefits for patients. Researchers at the Johns Hopkins University School of Medicine, in Baltimore, rated clinical information technologies at 41 hospitals in Texas and compared those results with discharge information for more than 160,000 patients. Technologies recorded included electronic note taking, treatment records, test results, drugs orders, and decision-support systems that offer information concerning treatment options and drug interactions. The researchers found that hospitals that rated highly on automated note taking had a 15 percent decrease in the odds that a patient would die while hospitalized. Hospitals with highly rated decision-support systems also had 20 percent lower complication rates. Researchers found that electronic systems reduced costs by about $100 to $500 per admission.

Share |      Randall Parker, 2009 February 02 12:46 AM  Policy Medical


Comments
Ned said at February 2, 2009 5:57 AM:

How do we get there? William Shakespeare put it so well in "Henry VI, Part 2":

"The first thing we do, let's kill all the lawyers". - (Act IV, Scene II).

zylonet said at February 2, 2009 8:38 AM:

>>A medical marketplace which rewards use of expert systems, electric patient records, and reduction of total costs by use of more effective and productive methods of purveying treatments is what we need. How do we get there?

As someone who has been both seriously ill and provided management consulting services to medical practices, I have some insight. You cannot achieve efficiency in the medical realm because of government interventionism and because of the AMA. By artificially limiting the supply of Dr.s, the AMA has created a system were there is little incentive for efficiency. The pool of patients is far greater than the supply of Dr.s, leading to an incentive for profit taking. The massive government spending also reduces the need for efficiency because the money is cheap and easy, and because you must follow certain government standards to realize the medicare/medicaid payouts. Finally, the onerous regulations in terms of alternative care, ensure that competition will not arise.

Basically the medical field in the USA is a disaster because of government interventionism and by the deliberate choices of the AMA. Nothing can change the market for the better, except a roll-back of government spend and regulation. Obama's plan to, more or less, force electronic medical records is simply moronic. Private firms are selling these systems and the software is evolving. There is no single best solution to date, nor will there every be a consensus on basic issues such as platform. The government will completely destroy innovation if Obama has his way.

If you want good medical care, then get rich and find a private Dr.

Shannon Love said at February 2, 2009 9:05 AM:

You can't create an incentive system that focuses on keeping people healthy because people won't pay medical workers when they're not sick. Young people won't even buy minimal health insurance just to guard against the chance they might have an accident. On flip side, once people get sick, cost is no object to them. In short, people won't pay for health, they will only pay to return to health.

If you did create an incentive system to keep people well, this would cause the system to ignore people who were actually sick because they would be just a cost. Only insurance companies have an incentive to keep people healthy and they have no direct input on either medical care nor the behavior of individuals who are not yet sick. No one will trust the insurance provider to make the best decisions and no one will accept the authority of a corporation to mandate a healthy lifestyle.

The only way of creating such an incentive would be to have people control their own health care spending. We could use a voucher system combined with purchasing coops to accomplish this. Each individual would get a voucher to purchase health care but could only spend that voucher by joining a coop with several thousands of other people. The coop would hire health care administrators who would decide on standards and methods of care and who would contract with medical care providers. Any money the coop did not spend each year would go back to the individuals so people would have an incentive to seek efficient care. Since people could pick and choose their coops different coops could place obligations on members such as exercise and diet if they so choose. People would be directly responsible for their own health. We would socialize the cost of medical care but decentralize and de-politicize health care decision making.

Not a perfect system but it would: (1) cover everyone while (2) keeping the flexibility and innovation of a free market system and (3) create incentives to keep people healthy in the first place.

Wayne Conrad said at February 2, 2009 9:10 AM:

What I expect we will be hearing more of in coming days is this argument: "Look what the free market's done to health care! Only the government can fix it." However, the free market has not done this, health care being anything but a free market.

For one insider's view of what happens under nationalized health care, see http://nhsblogdoc.blogspot.com/ .

Ned said at February 2, 2009 11:27 AM:

The AMA does not control the number of physicians in this country. Anyone is free to start a medical school, and existing medical schools are free to expand their class sizes as they see fit, subject to approval by the Liason Committee on Medical Education (LCME). Physician licenure is controlled entirely by the states, with no input from the AMA. Graduates of domestic and foreign medical schools can obtain licenses. Any attempt to restrict the supply of practitioners in a licensed profession by a professional group is a violation of antitrust laws. The reason there aren't more medical schools is because they are so expensive to operate. Whether or not there's an actual physician shortage in the US (rather than a maldistribution) is debatable, but the AMA has nothing to do with it.

Mike Anderson said at February 2, 2009 2:36 PM:

Maybe we should send some health care administrators out to talk to their dentists. Dentists seem to have no problem selling preventive care.

Randall Parker said at February 2, 2009 7:40 PM:

Shannon asserts:

If you did create an incentive system to keep people well, this would cause the system to ignore people who were actually sick because they would be just a cost. Only insurance companies have an incentive to keep people healthy and they have no direct input on either medical care nor the behavior of individuals who are not yet sick. No one will trust the insurance provider to make the best decisions and no one will accept the authority of a corporation to mandate a healthy lifestyle.

I do not see why a system for keeping people healthy would necessarily displace a system that treats them once they are sick. Certainly the sick people would still be up for spending money to treat their illnesses. Surely an insurance company would be willing to sell disease-treating insurance for a lower cost to a person who is in a program designed to increase their chances of staying healthy.

Health care purchasing coops: I'm reminded of food coops which fall apart because volunteers do not do enough volunteer work. Why would you expect a coop to do a better job of choosing providers?

John Moore said at February 2, 2009 8:43 PM:

HMO's originally we designed based on the "keep people healthy" models (hence the name Health Maintenance Organizations). Significant cost reductions and health improvement were promised for clinic based medicine, early diagnosis and treatment and other "healthy" practices.

It didn't work.

Medicine really doesn't know much about how to keep people healthy. For example, it knows that obesity causes increased morbidity and mortality, but so far has found NO sucessful intervention other than drastic and dangerous gastic surgery.

It is very comforting psychologically to believe that if we do the right things we will be healthy. Experience has shown that so far, the gains aren't as much as expected, and the behavior modifications are very difficult even for well motivated individuals.

-------------

As for automation, it is clear that the current system is pathetic. We have a Mayo Clinic and Hospital here in Scottsdale, AZ, and it is highly automated. The result is dramatically improved care at significantly lower overhead, and with a much happier workforce.

In said at February 4, 2009 9:18 AM:

"It is very comforting psychologically to believe that if we do the right things we will be healthy."

With enough intellect and will, the information is now available for most people to be capable of greatly increasing their health through their choices.

The reasons people don't include:

1) that they can't assess health information critically, i.e. they take bad advice (which is everywhere), this is probably the most common reason
2) they don't see value in it
3) "and the behavior modifications are very difficult even for well motivated individuals", the changes neccessary can seem to be relatively radical compared to how one has always lived
4) probably other reasons

John Moore said at February 4, 2009 11:18 AM:

Fat people do not want to be fat. Our culture is strongly against fat. Many fat people understand the health consequences of obesity. Many fat people understand that excess calories make you fat and exercise can help reduce that.

In spite of all of this, these people remain fat.

Behavior modification is simply not a very successful approach, even if it is the only one available.

In other words, how do you CREATE and MAINTAIN the will, given the already strong incentives?

If you can solve that problem, then perhaps keep-healthy approaches will have some significant impact. Of course, if you can solve that problem, you can also improve human behavior and society in a very wide variety of ways.

The problem has not been solved, and to believe otherwise is to ignore this critical issue.

Dan Hamilton said at February 4, 2009 12:07 PM:

Weight loss programs are the only 15% successful treatment that Droctors would ever accept.

If the weight loss program fails then the patient is worse off then if he had not dieted at all.

85% of all weight loss programs fail.

So much to first do no harm!

CS said at February 4, 2009 12:47 PM:

"Automation in diagnosis?" Jeepers. Stick with areas you know something about, whatever those may be. And the "AMA limits the numbers of MDs" stuff is an old misunderstanding. There obviously aren't enough physicians of some types, or we wouldn't have to import them. But the AMA's sins are limited to being the least effective trade association in Christendom. We don't have enough competent specialists, but that supply is limited by the work being fairly hard, and paying less than investment banking and big-firm law. Family practitioners? There are too many, and the job description is overtaken by events. Almost everything they do is well within the scope of a decent PA or NP.

Dentists? Sure, let MDs charge patients thousands out of pocket if they get sick, and some will work harder at staying well. For the rest, just look in some mouths in free dental clinics.

On the more general question, most people manage to stay pretty healthy most of the time without much help from physicians. Their health care consists largely of acute fixes. Those who think, or claim to think, that those at risk for chronic illness can be cajoled or reasoned into lifestyle changes are dreaming. All that's possible is to refuse to treat such disease pour encourager les autres. Sound likely here?

Oh, if the Gummint wants to improve health care and stimulate a couple of sectors, let 'em fund a system of centralized medical record storage and retrieval. Patients and their providers log in to read or note their charts. No more providers working at cross purposes due to incomplete records, and far greater efficiency in seeing patients. This kind of system works privately under HIPAA for coverage info, and should for clinical stuff. There's just an obvious commons problem here, which only govt could solve.

Dana H. said at February 4, 2009 1:16 PM:

Do people say the system of automobile repair is broken because mechanics don't have proper incentives to keep vehicles healthy? Do we have a plumbing crisis because you only call the plumber when the toilet backs up, but not to preventively clear out the roots that cause the problem? Should we appoint a commission of learned scholars to look at these problems? Sheesh.

The only problem with the business model of hospitals and doctors is that government has taken over whole sectors of medicine and regulated to death what remains -- from licensing requirements, to tax treatment of medical insurance, to mandatory coverage requirements, to Medicare/Medicaid, to the ease of malpractice suits. In short, the problem with the business model is too much government and not enough business. And this is hardly an old model -- it is one that has steadily been brought about over decades from the New Deal, to the Great Society, to G. W. Bush's prescription drug plan.

What we need is a MORE antiquated business model, like the one we had in medicine 100 years ago.

John S said at February 4, 2009 1:50 PM:

Three more places to read:
http://thehappyhospitalist.blogspot.com/ -- Dr Happy blogs powerfully on the "Medicare National Bank"

http://www.kevinmd.com/blog/ -- Dr Pho

http://www.healthbeatblog.org/ -- Principally Maggie Mahar, a healthcare policy analyst

Shannon Love said at February 4, 2009 2:04 PM:

Randal Parker,

Health care purchasing coops: I'm reminded of food coops which fall apart because volunteers do not do enough volunteer work. Why would you expect a coop to do a better job of choosing providers?

I had in mind a legal framework in which people would join together to hire medical experts of their own to supervise their own care and set their own standards. Medical suffers badly from the agent problem. No one has an incentive to look after you health except you. We need to create a system in which people directly choose the people that oversee their medical care.

It would be something like being partner in your own insurance company and medical standards board.

MF DDS said at February 4, 2009 2:08 PM:

A couple of things: Yes, dentistry has done a good job with prevention, but medicine and dentistry differ greatly in the number, type and mortality of the diseases they treat. It would be nice to have the equivalent of brushing, flossing and fluoride for general health, but this won't happen anytime soon.

Second, Shannon Love was spot on above mentioning incentives and health care spending. Helping people become better consumers of health care services will take time. Convincing people to become better stewards of their own health is much more difficult because of human nature and free will. A funding system that understood and respected incentives would go far to correct our problems.

Finally, government interference through mandates and burdensome rules are the source of the problem, with the tort system in second place. Mandates have an inverse relationship to health plan affordability.

Kyle said at February 4, 2009 3:13 PM:

For example, it knows that obesity causes increased morbidity and mortality, but so far has found NO sucessful intervention other than drastic and dangerous gastic surgery.

You mean that there is NO "EASY" intervention. I'm 40yrs old and have lost 60lbs over the last 8 months (280 to 220), and it is anything but easy. However, it is much lower cost than going on blood pressure medication and a statin which is where I was headed. In my case it means watching what I eat and getting up and working out 5-6 days/wk (planning my first ever 10k this spring).

I could have easily made excuses why this would have been impossible for me. I have a ruptured disk at L4-L5 and my knees complained a lot, but I started first with watching my food and then slowly (very important) added the exercise. Now I'm having much less pain than before I started this process and now that I may burn up to 1000 calories solely by exercise, I don't have to limit my calories nearly to the degree I once did.

I'm a little over 6'3" so I'm still in the overweight area of the BMI chart but my goal is the 200lb range which will put me at the high end of the normal range. That should be about right for me since I do a fair amout of resistance training as well.

Had my insurer charged me more for being obese (or better yet, paid me to lose), I probably would have been motivated much sooner.

TMLutas said at February 4, 2009 3:21 PM:

As IT support for a 1 doctor practice using a radical low-overhead medicine business model I think I can provide some real world input. The CPT/ICD-10 code combos are the basis for all payment. CPT codes for wellness either do not exist or are compensated at joke rates. Fix compensation and you'll get both more primary care physicians and ones that do the work necessary for better health outcomes. It really is that simple.

Automation of diagnosis is difficult. At best we've achieved a system that automates the up front grunt work and gives doctors a list of differential diagnoses to start at. That's no mean feat but this reduces doctor time to completion in a good patient encounter. It does not eliminate it. Doctors are diagnosing you from the moment they see you, they smell you, and they touch you. Their questions and your answers are only part of the process. It's not something that can be handled by any machine ready for use anytime soon.

Nick G said at February 4, 2009 4:16 PM:

"Automation of diagnosis is difficult. "

Not really. 30 years ago published studies found that a simple decision-tree based computerized patient history/complaint program could do as well as the average doctor in diagnosis (half of all doctors are below average...). Automated diagnosis should have been widespread decades ago.

"Doctors are diagnosing you from the moment they see you, they smell you, and they touch you. "

Really good ones are. OTOH, if we can match the 75th percentile, we'll have improved care for 75% of patients.

Obviously, a computer-assisted diagnosis would be the very best, getting the best of both worlds. Also, such a system would help with self-care enormously.

The bottom line is that the quantity of medical information outstripped the ability of humans to process it many decades ago - that's why we have specialists. Right now they're doing quite badly: repeated studies find that common ailments don't get the standard of care from 40-80% of the time. We desperately need to automate as much as possible.

In said at February 4, 2009 6:01 PM:

John Moore

"Many fat people understand that excess calories make you fat and exercise can help reduce that"

This is what I mean by bad advice. That is an oversimplification. The human body is a complex adaptive system; calories in does not equal calories out. You tell someone this and they don't have the intellect part down (and most dieters don't).

What I'm arguing is that due to the internet and emerging theories of metabolism, information is readily available to anyone that is sufficiently motivated and capable of disseminating that information to greatly improve health. Many people have done so. The people giving health advice that make the most sense are all saying the same things. The basic insight they start with is that we are well adapted to a more paleolithic lifestyle.

"Behavior modification is simply not a very successful approach, even if it is the only one available. In other words, how do you CREATE and MAINTAIN the will, given the already strong incentives?"

I don't know, but the public health recommendations are not helping. Most "Experts" recommend a carb based diet. Research has shown that carbs can damage appetite control: http://www.marksdailyapple.com/carbs-the-gift-that-keeps-on-giving/

Consider that for most of human history grains did not make up a significant portion of the human diet.

There are reasons to doubt the lipid hypothesis as well: http://www.amazon.com/Fat-Head-Tom-Naughton/dp/B001NRY6R2/ref=pd_bbs_sr_1?ie=UTF8&s=dvd&qid=1233701058&sr=8-1

Randall Parker said at February 4, 2009 6:42 PM:

Kyle,

Most people who lose a lot of weight go on to regaining all of it. Diets mostly fail even when they succeed in the short run. I hope for your sake that you are an exception. But the odds are against you.

Nick G,

While writing this post I went googling to find some studies on expert systems that out-perform doctors. But my searches turned up too many irrelevant pages. I know I've read the same claims as you are making about the efficacy of expert systems for medical diagnosis. But I can't find the research. You have any useful leads on this?

Anon said at February 4, 2009 6:51 PM:

Kyle, give yourselves five years and come back and report. As RP says, the odds are strongly against you.

I speak from personal experience. I lost 97 lbs and did lots of exercise and only succeeded in resetting my metabolism so that 1400 kcal/day is all I can eat 10 years later without gaining weight. Did I mention I'm 6'2"?

John Moore said at February 4, 2009 6:56 PM:

In,

This is what I mean by bad advice. That is an oversimplification. The human body is a complex adaptive system; calories in does not equal calories out. You tell someone this and they don't have the intellect part down (and most dieters don't).

No, it is a simplification and a very useful one. Yes, it appears that lower carb diets may increase the amount of calories on can ingest to some extent, and have some other benefits.

But fundamentally, calorie balance is a pretty close approximation.

Kyle said at February 4, 2009 7:40 PM:

On reporting back in 5 years. Yes, the odds are against my not gaining the weight back in 5 years. I'm doing what I can to make this a new part of my life. I'm not following some fad diet and not even counting calories. I have tried to move my nutrient intake much more like what my grandparents used to eat, whole grains, much more fruits and vegetables and cut way back on sugar. I eat fats, but the right kind (olive oil, flax seeds, etc). I don't go hungry. When I do want something sweet (I love chocolate) I try to eat a couple of pieces of dark chocolate, not that many calories and it is satisfying.

The reason I exercise first thing in the morning is it's a conscious decision to make it a priority. If I wait until the end of the day, I'm too tired or there is always something else to do.

What got me motivated was that I was finally seeing blood chemistry changes that I didn't like and I don't want to spend the money or deal with the side effects of the recommended medications. And those changes hit home much harder with young children at home, I intend to be around for them.

Again, I would very likely taken these steps sooner if there were some economic penalty/reward set by my employer or insurer. Some employers are getting it, like IBM - http://wjz.com/health/wellness.programs.company.2.278462.html

For companies, it's all about the money. Studies show $3 in health care savings for every for every $1 spent on wellness.

Of course this article shows why it can be hard on employers to make common sense decisions related to employee health: http://www.cfodailynews.com/4-gotchas-hidden-in-wellness-programs/

In said at February 5, 2009 6:44 AM:

John Moore
You are missing my point. I'm saying (you seem to be saying this in your second post as well), that it doesn't help fat people to know "that excess calories make you fat". I'm arguing that there are things that fat people can know that can help them lose weight much more easily and sustainably. I agree with you that many lack the motivation, but a huge part of the problem is misinformation. The conventional wisdom works against the dieter's purpose.

John Moore said at February 5, 2009 8:58 AM:

I think the fundamental problem is that even adequate information does not provide adequate motivation. I'd bet most overweight people are motivated to change - look at the huge industry in diet products, ranging from fraudulent to serious medically supervised programs. Even the latter, which have required class attendance, have terrible outcome rates.

I think the reason for this is the extremely powerful homeostatic mechanisms of the human system, which seeks an equilibrium, and not necessarily the best equilibrium. There are limits to how much people are able to change even with motivation. Of course, like everything about complex biological systems, this varies widely.

Consider drug addiction, alcoholism or tobacco. Most people who use strong opiates for pain treatment (morphine, fentanyl, oxycodone) will not get addicted. Some will.

Food, unlike opiates or nicotine, not only has similar neurochemical effects, but also cannot be avoided "cold turkey." Thus the operant conditioning effects remain present even in those who limit their caloric intake.

Statistically, it is easier to "kick" a nicotine or opiate habit than it is to curb over-eating (and easier to kick opiate addiction than nicotine, btw).

Last time I checked, the only bariatric treatment with significant success was gastric surgery of various sorts. It also is very dangerous and has unpleasant side effects. The drug combination Phen-Fen was actually pretty successful, but unfortunately also dangerous. I believe the best treatment for stubborn obesity will turn out to be medication.

Nick G said at February 5, 2009 9:17 AM:

Randall,

Sadly, I saw the research back in 1980, and didn't save it. It was in a peer-reviewed journal, and was based on nothing more than detailed interview of the patient - no electronic medical records required.

I've observed doctors quite a bit professionally and personally over the years in settings that range from primary care to tertiary care in the best teaching hospitals (including Mayo Clinic and the ER on which the television show was based), and I've found that most doctors do most diagnosis with very simple "cookbook" algorithms/decision trees. Any honest physician or participant in the health system will confirm this. This is not at all hard to duplicate (and then out-perform) in software. The doctor detectives you see on television, who diagnosis rare illnesses and track down problems with psychic accuracy....well, they may exist, but most patients don't see them.

The suggestion that software that out-performs most physician diagnosis couldn't have been developed decades ago seems laughably absurd to me. It's obviously physician resistance to competition.

BTW, I lost 30 lbs about 8 years ago by losing it very slowly (half pound per week, over more than a year), and making dietary changes, and have kept roughly half of it off (it fluctuates around the mid-point). The two keys: eliminating junkfood (sweets, refined bread, crackers, etc) from the house, and accepting a little bit of chronic hunger. We're really not designed for food abundance, and I'm ready for drugs that tone down the hunger signal.

In said at February 5, 2009 9:38 AM:

"serious medically supervised programs. Even the latter, which have required class attendance, have terrible outcome rates."

What you don't seem to understand is that even "serious medically supervised programs" are full of bad recommendations. There are various dogmas that are highly questionable but accepted by most. For example that low fat is healthy. These are the kind of ideas in the heads of an average dieter and there are a lot of societal forces that push these ideas, but they are ultimately pernicious.

There is a burgeoning community of paleo lifestyle bloggers that posit an alternative theory of human metabolism that makes more sense from so many perspectives. They generally advocate high fat and intermittent fasting, which is something paleolithic humans would have done as a matter of course. They pretty much all claim that cravings for carbs and other junk vanish in a relatively short order. I've experienced this myself.

I'm arguing that there is a different way of looking at metabolism that actually enables people to change how they are motivated to eat. Food addiction is more precisely carb addiction. It is the carbs that are making people fat.


John Moore said at February 5, 2009 10:07 AM:

In,

I have also experienced that sort of effect and it is real. There is no question that the medical establishment has been too consensus driven on the low-fat approach (the exact same phenomenon as happens with Climate Change).

I also experienced another effect once. I used a very restricted calorie diet, which was relatively low in fat. After about 2 weeks, I felt no need for further calories. Unfortunately, the willpower required to stay on that (damaging, as it turned out) diet was something I could only sustain a few years.

The high-fat (or high protein, or magic ratio, or whatever) insight is valid, but appears to be insufficient. There are a hell of a lot of people who have tried and failed at that approach, just as there are a lot of (highly motivated) people who have been able to maintain low weight on a low fat diet.

There is also the question of the impact on blood lipids, although my own experience with Atkins was that the overall lipid balance was beneficially altered.

On the other hand, I watched someone using Fen-Phen (or is it Phen-Fen) and no other modification, who lost weight effortlessly until the combination was no longer available.

CS said at February 5, 2009 11:46 AM:

Nick G, template-based diagnosis has the main virtue of forcing mediocre primary physicians to refer to roughly the right compass point instead of just punting. It's better for the patient even if the primary refers the wrong way, although those who care more about cost containment than patient care won't think so. If algorithmic systems (heuristics scare me in this area) can get 75 percent right, that's mostly the 75 percent that doesn't require referral. Effort to develop templates for specialist diagnosis are pretty miserable at this point. Thinking "automated diagnosis" will fix things is just another version of magical thinking, that clever newcomers can discover some hitherto unimagined solution to an intractable problem.

Nick G said at February 6, 2009 8:52 AM:

CS,

Keep in mind that most physicians (and vets, for that matter), are miserably bad at interviewing, getting and remembering complete, accurate patient histories, primary and specialty physicians alike. That's the very first thing that should be automated, and it forms the basis for good diagnosis.

Also, treatment needs to be computer-assisted. How many studies have we seen that say that common problems, for which standards of care exist, don't get the proper treatment 40-60% of the time??

I'm not suggesting that automated diagnosis & treatment can do miracles. But, I think we both agree that most people get sub-par diagnosis and treatment, and this can be changed. Also, such systems would help patient education, compliance and self-management.

Nick G said at February 6, 2009 8:55 AM:

CS,

I would also note that physician hostility to expert systems prevents them from improving. If such systems were welcomed, widespread and worked with by a very large number of physicians, they would begin to capture the very best approaches.

MDA said at February 7, 2009 8:15 PM:

I am a medical specialist who happens to do a lot of research on automated diagnosis and detection, e.g. http://care.diabetesjournals.org/cgi/content/full/31/8/e64
I would like to get back to the original post:
"I would like to see far more automation of diagnosis. This requires wider spread use of electronic medical records so that that the data which medical expert systems need will exist in electronic form. It also requires an economic model for medical care that provides incentives for automation. Medical expert systems can make better diagnostic decisions because the huge and growing quantity of medical test results and the large number of diseases and treatments really test the limits of the human mind to process all that information. Medical expert systems can free up smart doctors to do more original creative work such as medical research and product development."
If it were the case that a doctor, with a new patient, initially orders a bunch of tests, enters the results in a computer, integrates the results in his/her mind, and only then comes to a decision, it would be easy to 'automate the diagnosis'. However a) most doctor visits are follow up visits, not primary diagnosis, so though automation of this part of the process may improve accuracy, it will not improve productivity b) doctors 'think with their hands' mostly, i.e. talk with the patient, examine the patient etc, and only then order tests. A 'good doctor' will mostly order only those tests that confirm or reject his/her initial diagnosis that is sometimes made as soon as a patient walks in. Thus, automation would require this whole interactive, messy, sequence of listening, examining, integrating, and more examining to be automated. Right now, this is technologically impossible and far in the future. c) I will let the difference between surgical and medical specialties aside for now.
This interactive, exploratory mode explained above is primarily what people think of when they discuss automation of doctor's tasks'. However, a substantial part of doctors' time with patients - without references I will not go as far as saying *majority* of a doctor's time - is spent on other tasks. Especially, follow-up of patients once diagnosed with a disease, and after some decision with regard to management has been taken, for evaluation of the patient's management (~treatment), and development of complications or new problems.

And it is here that I see the primary application of automation, and where much of my and others' research focuses. In other words, use automation in a) selecting only those patients that need to be seen by the slow, not so efficient, but so far irreplaceable doctor b) to allow patients and the people close to them to monitor themselves, by automating. The last, allows patients for example to monitor their diabetes themselves, their eyes themselves, etc. and only visit a doctor when something needs to be done.

This is the goal, but let me not mention the political, legal, ethical and reimbursement entanglements of bringing such research into practice.

Bob Badour said at February 8, 2009 8:43 AM:
I would also note that physician hostility to expert systems prevents them from improving. If such systems were welcomed, widespread and worked with by a very large number of physicians, they would begin to capture the very best approaches.

It's perhaps somewhat ironic that I have seen such a system that appears welcomed, widespread and worked with by a large number of physicians: google.

I don't think google is necessarily the most efficient system, and I strongly suspect a different indexing algorithm focused on diagnosis would provide better results.


Thus far, my experience with the medical profession has been a curate's egg. The only things they seem to get right are simple things using old technologies, and even then, by breeding resistant bacteria strains, the medical profession has done some harm.

I have had several infections that required antibiotic treatment. Two of which required surgery.

Circa 1980 I had a cartilaginous tumor removed with a bone graft. Without surgery, the outcome could have been very bad.

More recently, though, when I described my autistic inertia to a doctor, she prescribed an anti-depressant, effexor, which did me no good and caused me great harm. She did this in spite of the fact I gave her information definitively ruling out a diagnosis of clinical depression. I now see effexor is not recommended as a primary drug, but only for treating the most severe, treatment resistant cases of acute depression.

Because colon cancer runs in my family, I go in for regular screening. Now it appears the fleet phospho soda prep may have damaged my kidneys, and I might not find out for several years. I have had double-doses of fleet phospho soda three times now, and all three times were AFTER doctors first identified the treatment as kidney damaging back in 2003.

Harumph

Nick G said at February 8, 2009 3:45 PM:

MDA,

Let me go through your narrative point by point:

"most doctor visits are follow up visits, not primary diagnosis, so though automation of this part of the process may improve accuracy, it will not improve productivity "

Reducing the doctor time spent on diagnosis is not what I would see as the goal. As I noted above, as far as I can tell, doctors spend far too little on diagnosis as it is. The goal is improved diagnosis, which will greatly reduce system costs and improve the quality of care.

"A 'good doctor' will mostly order only those tests that confirm or reject his/her initial diagnosis that is sometimes made as soon as a patient walks in"

A good doctor takes a really thorough history. As I noted above, in my observation very few do. Many times when doctors think they can make such snap evaluations they are mistaken. This initial preliminary diagnosis is crucial - as you note, the whole course of treatment is set at this point.

"automation would require this whole interactive, messy, sequence of listening, examining, integrating, and more examining to be automated...this is technologically impossible "

The patient interview/history is easy to automate - it was done 30 years ago. It wouldn't be that hard to enter the central results of the exam and followup questions into a system, and a good medical record really should have that info, rather than it staying in the doctor's head. Such entry might interfere with getting the patient out in 15 minutes, but that's ok.

" a substantial part of doctors' time... is spent on other tasks"

A crucial part of this is treatment - how many studies have we seen showing that 40-60% of patients with common problems, for which there are standards of care, don't get the proper care? Automated systems would help this greatly. Doctors resist this kind of thing very strongly. They hate anything which encroaches on their autonomy. But, it's desperately needed.

Finally, I agree that automated systems certainly could help patients with self-care.

Nick G said at February 8, 2009 4:32 PM:

Bob,

Thanks for your story. It illustrates nicely how badly doctors do at integrating large amounts of medical data, like medication side effects.

Randall Parker said at February 9, 2009 7:56 PM:

Nick G,

It helps to get specific (with URLs) when you say that for common problems patients do not get standards of care. I've read about, for example, hospital emergency rooms where heart attack patients do not get parts of a standard best practices list of treatments. Hospitals have to turn best practices duties over to nurses because the doctors can't be bothered to run down standard lists and make sure best practices are all followed.

Here's an example of a hospital struggling to implement best practices and coming up short:

Prior to the new system, Strong had a voluntary clinical guideline for preventing clots. Under a cooperative patient safety initiative, Strong Memorial Hospital, Highland Hospital, and three other Rochester-area hospitals put together a prevention program that included walking when possible, use of compression stockings or pads, and required early use of preventive medication, such as Heparin, at appropriate levels.

Although about 70% of people admitted to these hospitals have a high risk of clots, about 60% were not receiving preventive treatment. In the first three months trying the new guidelines, that level was cut to below 50%. Hospital leaders then re-emphasized the need and value of the initiative.

Here's one (and I've previously read better) about how doctors in hospitals fail to prescribe best practices drugs for patients getting checked out of hospitals after heart attacks:

The degree to which the care system was incorporated into each hospital varied. Some improvement was seen even in the hospitals that didn’t use the toolkit very often – for instance, an increase of about 7 percentage points was seen in prescriptions for aspirin and beta blockers that were written before patients left the hospital.

But in hospitals that consistently used the tools, the gains were much greater. Use of aspirin and beta blockers early in a patient’s hospital stay increased 6.6 points and 5.6 points, respectively. Pre-discharge prescriptions for the same drugs rose 12.4 points and 6.3 points, respectively. There was also a 7.7 percentage point increase in prescriptions for ACE inhibitor drugs given before patients went home. And a 9.6 percentage point jump in cholesterol tests was also seen.

I've read great research studies on doctor error rates in diagnosis and failures to follow best practices in hospitals. But I'm not having a lot of luck doing searches on either of these topics.

Nick G said at February 11, 2009 11:38 AM:

Thanks for looking.

Here's one discussion for heart attacks: http://www.wrongdiagnosis.com/h/heart_attack/intro.htm . This website appears to be helpful - I haven't reviewed it in detail yet.

Post a comment
Comments:
Name (not anon or anonymous):
Email Address:
URL:
Remember info?

                       
Go Read More Posts On FuturePundit
Site Traffic Info
The contents of this site are copyright ©