Totally OT: Surgery

[Scuba Chuck]

[ QUOTE ]
Chris,

Please update us as you can. I wish you the best for a best case procedure and a speedy recovery.

scuba

[/ QUOTE ]

[AleoMagus]

As I'm not the type of guy who gets all weepy and cuddly when it comes to this sort of thing, but who still wants to give you something to be optimistic about and wish you well, let me enlighten you somewhat about the shoddy and self-serving nature of medical statistics.

First, when doctors tell you there is a 15% chance of something bad happening, that really means there is a much smaller chance. This makes good sense to the medical profession, because 15% happens often enough to make a lot of people wonder if their doctor did something wrong, etc... So if 15% actually meant 15%, and doctors told the whole truth, more people would tend to be upset at the results of a given medical procedure. In truth, when doctors know that something has (say) a 3-5% chance of happening, they can then say 15% and the 3/100 that it happens to will think "well, It did have a 15% chance of happening". The effect of this is more noticeable in more extreme cases, like when a doctor says that a person has (say) a 15% chance of a full recovery. The odds of this happening are usually much higher, but in most cases, people tend to be happier with their particular result. If they don't have a full recovery, they think "well, I only had a 15% shot anyways", but if they do, they think "wow, I must have had a good doctor and must be a super-survivor, because I only had a 15% chance and yet still made a complete recovery!"

Secondly, Doctors are very bad at applying bayesian implications to the stats even when they do tell it like they think it is. Example:

hypothetically, a medical test is 95% accurate. 1% of people who undergo a medical test have the medical condition being tested for. 95% of people with the medical condition will get a test that indicates they have the condition. 95% of people without the medical condition will get a test which indicates they do not have the medical condition (and thus, 5% will get a test which indicates they do have it). A random person undergoes testing and receives a result that indicates he has the serious medical condition. What is the probability that he ACTUALLY has the serious medical condition?

Most doctors, when faced with this kind of info, give a very incorrect answer. Most guess that the person has a 95% (or close to 95%) chance of having the medical condition. (The test is 95% accurate, right?!) Apparently only 10 to 20% of doctors can answer this question correctly. Various studies have been done with similar problems involving all kinds of real medical conditions to confirm this. And it has been confirmed. Etensively.

The correct answer is that a person who tests positive on such a test will actually only have about a 16% chance of having the serious medical condition, despite the test being 95% accurate.

What does this have to do with you? Well, medical tests do tend to operate this way, and this kind of testing happens throughout the diagnostic process, not just with regards to whether or not you have a given medical condition, but with regards to how severe it actually is (which in turn affects subsequent judgements about chances for complication, etc...)

Bottom line - Don't put too much stock in those numbers. I'm confident you will pull through just fine.

Best wishes
Brad S

PS - To all the doctors out there: I have more respect for your profession than perhaps any other. You aren't statisticians, and that's just as well. After all, if you have to choose between studying stats tonight, or the best way to fix Chris's leg, stick with the surgery stuff. I'll understand.

[raptor517]

all the best chris. we never talked or anything, but ive always respected yer input. everyone else is right, modern medicine is amazing. ull be fine bro. take care. holla

[clutch]

Good luck and godspeed.

[yid3655]

Best of luck with your surgery Chris, try to stay positive and I hope everything will turn out well for you

[Irieguy]

[ QUOTE ]


PS - To all the doctors out there: I have more respect for your profession than perhaps any other. You aren't statisticians, and that's just as well. After all, if you have to choose between studying stats tonight, or the best way to fix Chris's leg, stick with the surgery stuff. I'll understand.

[/ QUOTE ]

Most of the doctors I know understand bayesian implications.

Surgeons are trained to counsel about the risks of surgery. What we do is quote the published risk from research trials involving the procedure being performed.

It has become a type of folklore that only 10-20% of doctors understand specificity and sensitivity in medical tests. Now, all biostatistics classes start with the story about how doctors don't understand biostatistics. Then, most medical students pass the exam and proceed on to residency training, where there is daily review of medical literature and the statistical merit of this literature.

So, Chris is likely to do well and my sincerest well wishes are extended until he completely recovers from this ordeal.

As a bonus, it is likely that his surgeon tells the truth and understands statistics. Which is nice.

Aleo, you know I like and respect you. I have also learned a lot from you, and have publicly thanked you for that. But this post blows. It is not an accurate description of how most doctors think, nor is it an accurate assessment of the statistical sophistication of physicians trained in this era of evidence-based medicine.

Irieguy

[SuitedSixes]

AM-
I think your sample size is too small.

[The Yugoslavian]

Chris,

You will have to pwn Monday!!

I like your attitude:

[ QUOTE ]

Where I work, when we go out to deploy programs, we’re always happy when something minor goes wrong early, because the rule is that if it doesn’t, then something major always happens later. Hopefully that will apply here as well.


[/ QUOTE ]

[img]/images/graemlins/smile.gif[/img]

Get well soon!

Yugoslav

[AleoMagus]

Well, I did realy mean my medical statistics bashing in the best possible regard as far as Chris's surgery is concerned.

...And, I think that my points have at least at some point, been very sound ones. I didn't make this stuff up. As I said, it has been confirmed, repeatedly.

It has a lot to do with the counter-intuitive nature of the statistical implications themselves, and the mistakes that I speak of are often easy to fall into unsuspecting, even for a trained statistician.

The most famous example providing this result was with regards to breast cancer mommography. This example showed doctors getting it right only about 15% of the time. As I felt pretty confident my inductive logic and Phil of sci classes didn't teach me urban mythology, I did a quick google search for some sources. It was relatively easy to find quite a few. Just one quote:

[ QUOTE ]
Next, suppose I told you that most doctors get the same wrong answer on this problem - usually, only around 15% of doctors get it right. ("Really? 15%? Is that a real number, or an urban legend based on an Internet poll?" It's a real number. See Casscells, Schoenberger, and Grayboys 1978; Eddy 1982; Gigerenzer and Hoffrage 1995; and many other studies. It's a surprising result which is easy to replicate, so it's been extensively replicated.)



[/ QUOTE ]

Again, trust me when I say that there is no offense intended, and I do not make these comments as any kind of disparagement to the medical profession. I really meant it when I said I respect it more than any other. It's just hard to get the stats correct in a field like this, and they tend to be skewed one way for a variety of reasons.

And I think that many would agree that they are in fact, skewed in the opposite way than that which is most conducive to the healing process.

Geez, I personally know so many people who made a 'miraculous recovery' of one sort or other that it gives me enough reason to doubt the unlikeliness of it all. But I suppose on that count, maybe I have just been lucky, and my sample size is too small.

Regards
Brad S