« Direct Talks | Main | Slate Turns 10 »

June 19, 2006

The Problem With Probabilities

Nice illustration of a constant medical dilemma by Ogged, who happily appears to be stomach-cancer free:

the debate hadn't been between those recommending a gastrectomy and those favoring a follow-up endoscopy, but between those favoring a follow-up endoscopy and those who wanted to send me home and tell me to forget the whole thing. The latter group was convinced that the original finding were just a strange anomaly--not a mistake, exactly, but not worth worrying about. Older, more cautious doctors eventually brought everyone around to the consensus that they couldn't take even the small chance that I do have cancer lightly, so a follow-up is warranted. But even the doctor I talked to "officially" said "we don't expect to find anything."

So much of medicine is probabilistic. If you wanted to really cut costs, you'd take a coldly statistical view of the whole thing, with those who ended up on the wrong side of the numbers regrettable sacrifices. As a society, we're not ready or willing to do that -- and rightly so. But this is the essential conflict: politicians and hospital administrators look at the global budget, while doctors and patients look at the individual's health. The latter militates for constantly seeking the lowest possible error, the former for going with the statistics and saving money where you can.

And speaking of gastrectomies (the removal of the stomach, usually to stop stomach cancer), here's a set of 11 cousins, all with high genetic likelihood of developing stomach cancer, who've all removed their vulnerable stomachs.

June 19, 2006 | Permalink


TrackBack URL for this entry:

Listed below are links to weblogs that reference The Problem With Probabilities:


tinyurl makes homeless orphans cry.

Your probability point is a good one, though. It's something to keep in mind with the whole health savings accounts thing--diagnostic and preventative care, basically hedging strategies against somewhat long-shot illnesses, become luxuries for those without a lot of income, since there's a good chance any money they spend to that end is "wasted" in a sense (if they turn out to be healthy).

Posted by: Matt F | Jun 19, 2006 3:28:31 PM

As Matt F points out, their is a quite clear moral dilemma here(in the case of HSA's, but also for the MDs working under conventional insurance).

The moral imperative for the individual involved is between what amounts to suicide (betting that you won't die through a undetected, untested disease) or choosing a course that results in higher aggregate costs (someday) for everyone insured.

I've been recently watching the TV program called "House" (on Fox, reruns on USA network) about a team in a hospital (led by Dr. House, a 'difficult' person) that performs diagnostic medicine on difficult cases with unknown causes. It's a great way to understand the hard choices MDs have to make when the facts are not clear. Repairing a automobile and repairing a human body are not in the same universe of choices, but sometimes people seem to confuse medicine with collision damage insurance for cars.

Posted by: JimPortlandOR | Jun 19, 2006 3:55:40 PM

"led by Dr. House, a 'difficult' person"

What were the probabilities that Hugh Laurie would one day portray a genius doctor on American TV?

Posted by: bob mcmanus | Jun 19, 2006 5:50:04 PM

I think that as powerful as modern medicine has become, it is based on the wrong fundamental premise. Modern medicine seeks to treat specific problems, and does that rather well. What it does not do is seek to improve the general resiliency of the human body so that these problems come up less frequently.

Not enough emphasis is placed on the role of muscular strength as a protection against both joint damage and obesity. Bodyparts are viewed in isolation, not as part of an entire system that can be affected by small changes in situ. Doctors too frequently resort to drugs that may control symtoms, but do nothing to their cause.

More than that, little effort is made to harness the mind's power to exert control over the body's healing. The "placebo effect" is seen as an impediment to "real" medicine, not as something that should be cultivated on its own. This is incredibly wrongheaded.

I'm not disparaging modern medicine. I'm criticizing the way in which its application has displaced techniques designed to build systemic robustness and resilience in the body, which could dramatically increase the general level of health in society.

Posted by: Mastiff | Jun 19, 2006 6:17:14 PM

probability is tied to consumer demand as a driver for medical care: it is worth noting that some of the most effective care, in terms of resource cost and probability of enhancing outcomes are, essentially, clever public health interventions.

Posted by: Bruce Wilder | Jun 19, 2006 6:24:14 PM


We already know what the "secrets to maintaining health" are. Exercise, lose weight, eat healthy. You dont need a doctor, someone with a minimum of 12 years of training after college to tell you this. Every person over the age of 15 knows these "secrets" to maintaining health.

Its a huge waste of resources to train doctors for 12 years, only to have them waste their education by informing patients what any 15 year old teenager knows.

Now if you are talking about things BEYOND that level of simplicity, you are getting into untested, unscientific modalities that have no place in modern healthcare. Many quacks claim that that they can "maintain health" by stuff like spinal alignment, aroma therapy, or acupuncture when they have zero data to support their case.

As for drugs, you misunderstand how scientific drug research works. And its absolutely not true that drugs ONLY treat symptoms. TAke antibiotics for example. They dont treat symptoms, they cure the disease. Now would it be nice if all disease could be reduced simply to a very basic cause and effect, give this drug and it will cure you paradigm? Sure, and modern scientific research is working towards that.

All this nonsense about "treating the whole person" is just feelgood hype. There is no scientifically tested paradigm for this model, in fact nobody really agrees on what "treating the whole person" means in the first place.

If an overweight 15 year old has introverted right hip, nonweight bearing, intense pain, and X-ray shows inferiorly displaced right femoral head, the MD is going to diagnose SCFE syndrome. He's going to send the kid to a surgeon who is going to drill a screw to hold the femoral head in place. He's then going to tell the kid that he needs to lose weight, diet, exercise to prevent this from happening to the other leg. The surgery is curative, it doesnt just mask the symptoms, it fixes the problem. So stating like a broekn record that MDs only mask symptoms and never fix the problem is just an untrue myth. Thats also as close as "treating the whole person" as you can get with still maintaining scientific integrity.

Now the fat kid could also go to a quack mind/body "expert", who says that he needs to realign his "chakra" and meditate so that his body will heal itself. Is that what you consider more of "treating the whole person?" This has quickly become a euphemism for exploring feelgood nonsense that has zero basis in scientific evidence.

Posted by: joe blow | Jun 19, 2006 8:41:19 PM

"doctors and patients look at the individual's health. The latter militates for constantly seeking the lowest possible error..."

No. Doctors do not seek the lowest possible error. Doctors have a bias in favor of intervention. When there is uncertainty, doctors tend to want to do SOMETHING, even when the odds are high that the right thing to do is nothing.

Now, as Ezra points out, one problem with doing something is that it costs more than doing nothing.

But there another problem with doing something. There are risks associated with treatments, and some of these risks are very substantial. Unnecessary surgeries kill or cripple many people- simply due to the error rates and unavoidable risks of all surgeries. So it is important that people are not subjected to unnecessary treatment - and this is just as important as making sure that they receive necessary treatment.

Posted by: JR | Jun 20, 2006 9:10:30 AM

"Doctors have a bias in favor of intervention. When there is uncertainty, doctors tend to want to do SOMETHING, even when the odds are high that the right thing to do is nothing."

Uh well its not that clearcut. Patients also have a bias in favor of intervention. Very few patients go to the doctor and are satisfied with a "this will get better on its own" or "there's nothing I can do about this" type of response from the doctor.

Teh calculus for deciding whether to intervene or not can be quite complex, and is extremely patient-dependent, therefore its hard to study in aggregate. One patient may be fine with a capillary hemangioma on her face. Hemangiomas usually resolve by adolescence with no intervention. Another young girl may be absolutely mortified by the facial hemangioma and demand laser treatment, which is somewhat effective, but does carry side fx. Who are you to say the second girl is wrong for choosing to intervene? The decision to treat with laser, although hemangiomas will resolve on their own with no laser, can still be a justifiable medical decision, if the patient has deemed that the hemangioma is so dastardly to her appearance that she's desperate for a treatment. The responsibility then falls to the doc to explain the relative success/risk of lasers.

Posted by: joe blow | Jun 20, 2006 7:55:15 PM


I know a little something about doctors and bone surgery, okay? I have seen children with neuromuscular conditions who have been under the knife dozens — dozens — of times, where the net effect is to weaken the body with minimal gains to mobility.

Some surgeries can be extremely useful; I know from personal experience. But others are useless because they don't go far enough up the causality chain. The problems they supposedly corrected return in spite of the surgery. I know that from personal experience too.

And as far as no clinical evidence for acupuncture et cetera, you may want to check out the medical literature before you spout off.

Posted by: Mastiff | Jun 21, 2006 9:49:46 PM

This sounds all too painful and disturbing...

Posted by: Burn the fat feed the muscle | Jul 22, 2007 8:58:47 PM

The comments to this entry are closed.