Doc says 30% chance of survival

[jason1990]

Imagine my wife is about to have surgery. The doctor comes and tells me that there is a 30% chance she will survive the operation. Which of the following short paragraphs best describes what the doctor most likely means by this statement?

A. I have looked at 1000 women whose circumstances are all very similar to your wife's and who have elected to have the surgery. Of them, about 300 have survived. To the best of my knowledge and judgment in the field of medicine, I have no reason to believe that your wife bears any significant differences from the other women in this group.

B. I am 30% confident your wife will survive the surgery. Although this is only my opinion, it is based on my expertise in the field of medicine and is consistent with all of the known information about your wife's circumstances.

[BruceZ]

[ QUOTE ]
Imagine my wife is about to have surgery. The doctor comes and tells me that there is a 30% chance she will survive the operation. Which of the following short paragraphs best describes what the doctor most likely means by this statement?

A. I have looked at 1000 women whose circumstances are all very similar to your wife's and who have elected to have the surgery. Of them, about 300 have survived. To the best of my knowledge and judgment in the field of medicine, I have no reason to believe that your wife bears any significant differences from the other women in this group.

B. I am 30% confident your wife will survive the surgery. Although this is only my opinion, it is based on my expertise in the field of medicine and is consistent with all of the known information about your wife's circumstances.

[/ QUOTE ]

You probably aren't using the word "confident" in B to refer to a statistical confidence, though this could lead to a 3rd interpretation. That is, 30% of total survivors would have circumstances as bad or worse than your wife's.

I've wondered about what the weathermen mean by a "30% chance of rain". Sometimes that might just mean that the showers are so scattered that there is a 30% chance of being under one. Otherwise, I think it is like choice A, meaning that observed conditions have known to lead to rain with a probability of 30%. Not a 30% confidence of rain, which would mean that 30% of rainy days would have conditions less indicative of rain than today's conditions.

I hope your wife is OK.

[jason1990]

I hope I didn't give you the wrong impression about my wife. She is just fine and this question is entirely hypothetical.

[ QUOTE ]
You probably aren't using the word "confident" in B to refer to a statistical confidence, though this could lead to a 3rd interpretation.

[/ QUOTE ]
That's right. Here, his being 30% confident is meant simply as a subjective measure of his belief that my wife will survive. 100% would mean he's certain she'll live, 0% means he's certain she'll die, and 50% means he believes both outcomes are equally likely.

[MoreWineII]

B. It's his best guess.

btw, subtract about 15-20 percentage points from any doctor's projection on chance of living, that'll be more accurate.

[jason1990]

[ QUOTE ]
I've wondered about what the weathermen mean by a "30% chance of rain". Sometimes that might just mean that the showers are so scattered that there is a 30% chance of being under one. Otherwise, I think it is like choice A, meaning that observed conditions have known to lead to rain with a probability of 30%. Not a 30% confidence of rain, which would mean that 30% of rainy days would have conditions less indicative of rain than today's conditions.

[/ QUOTE ]
I've often wondered this as well. I don't know how meteorologists make their predictions, but I know a little about how some of the finance people make theirs. They have models that make predictions for them and, in a very simplistic and watered-down sense, their job is to estimate the parameters in the model. So when they say there is a 30% chance you will go broke, it may often mean this: "The model says there is a 30% chance you will go broke. I believe the model is accurate. And I have a certain level of statistical confidence that I have accurately determined the values of the parameters in the model." I imagine the meteorologists do something similar.

[[censored]]

It should probably be A but is most likey B.

[TStoneMBD]

i find it unlikely that your wife will survive the operation precisely 30% of the time so its clearly an educated guess based on his expertise. i dont think there is any arguement here.

[drudman]

I understand that when a meterologist says there is a 30% chance of rain, it means that 30% of the area for which the prediction has been made will definitely recieve rain.

[kuro]

When a doctor is talking about survival rates they are either making an educated guess or they are quoting a number from some study they read about in a journal. Most doctors avoid talking about percentages like the plague. People are completely results oriented when it comes to medical decision making. It simply doesn't matter to them or their family if a procedure is 99.9999% safe if they're the unlucky 0.0001%.

[fnord_too]

[ QUOTE ]
B. It's his best guess.

btw, subtract about 15-20 percentage points from any doctor's projection on chance of living, that'll be more accurate.

[/ QUOTE ]

That really depends. With a procedure or treatment that has is used frequently, say chemotherapy for a specific cancer, you will get the straight statistics I think if they give you real numbers. On the real numbers bit, my wife had Hodgkins Disease (one of, if not the, most beatable forms of cancer). We asked about the chances the chemo would render her infertile and got something like "there is a very small chance it will leave her sterile." Later I asked just what the chances were and he said something along the lines of "very small, there is a 66% chance she will be able to still have children." She was able to still have children, but not everyone's idea of "very small" is the same.