Neurologist, Radiologist, What’s the Diff?
CodeBlueBlog offers easy money (h/t: Michelle Malkin) for any neurologist involved in the Schiavo case who “knows” from Terri’s CT scan that she is PVS. The bet is simple:
I will provide 100 single cuts from 100 different patient’s brain CT’s. All the neurologist has to do is say which ones represent patients with PVS and which do not.
If the neurologist can be right 6 out of 10 times he wins the $100,000.
Any takers?
March 30th, 2005 at 11:33 am
[...] $100,000. UPDATE: Michelle has picked up the CodeBlue offer, as have Jeff Blogworthy and Xrlq.
Posted by Eric | 6:57 am |
Trackback URI:
[...]
March 30th, 2005 at 1:34 pm
What, is this guy looking to lose money?
Admittedly with a binary choice (monkeys choosing), there’s only a 2.8% chance of getting at least 60 right.
But these guys aren’t monkeys, and for the bet to have the expectation of being profitable they’ll only have to be right 56.13% of the time.
6% better than a monkey? My money would be on the neurologists…
March 30th, 2005 at 3:00 pm
He ain’t gonna lose a single dime. I have seen enough of these between 2 classes taken and another taught to know that there’s virtually NO WAY to tell from a CT who is PVS and who isn’t, especially not from a single cut that’s above the eyeballs. I have seen a CT slice of a patient who was missing A WHOLE HEMISPHERE (one entire half) and the dude was conscious and normal except for a limp and decreased fine motor control.
There’s no way tot tell from a scan unless it’s a FUNCTIONAL scan, which as far as I know, Michael Schiavo refused to allow.
March 30th, 2005 at 4:09 pm
CalTechGirl:
I agree that judging PVS from a CT is mighty tough, the misdiagnosis rate is 37%.
While that’s stupendously bad, it’s good enough to take CodeBlue’s money and they only have to do just a smidgen better than blindly guessing to come out ahead.
March 30th, 2005 at 5:02 pm
Only if you assume that neurologists know how to read CT scans. CBB’s point is that they don’t.
March 30th, 2005 at 5:15 pm
I agree that a radiologist has more skill at interpreting a CT than a neurologist. But surely the neurologist is better than a monkey (who would get 50% right). And I’m fairly confident that a neurologist has enough skill to get up to 56%.
Imagine taking a true/false test when all you had to do to pass was to have a 56% of getting a question right…
There’s a lot of money at risk here and I think Code Blue is being unwise by setting the bar so very low and if he honestly believes that the opinion of a neurologist is no better than a monkey then he’s being doubly unwise.
March 30th, 2005 at 5:29 pm
Wy only single cuts?
March 30th, 2005 at 5:32 pm
b/c that’s all that’s available in the Schiavo case. People all over the place are purporting to make a diagnosis based on what’s not in that one slice.
Neurologist or radiologist, I still suspect none of them would be right with a diagnosis of PVS or not 60% of the time
March 30th, 2005 at 5:45 pm
In re the “monkey” or flipping of a coin: those are only true if the test sampling is either random or both types are deliberately equal. If weighted, and let us assume he weights fairly honestly with 3-10 of 100 being PVS, I don’t think he’ll lose dime one.
March 30th, 2005 at 6:13 pm
No John, it doesn’t matter how many CodeBlue chooses to be PVS as long as each decision is an independent event. While there’s a strong human tendency to “balance the stacks,” if you are truly coin flipping, the odds of guessing each one right does not depend on the actual value chosen.
Suppose I’m guessing trying to guess a sequence of 10 ones or zeros. You, being crafty have made 9 of them 0 and one of them 1 (perhaps the last one for the sequence 0000000001). By flipping a coin, my odds of guessing a zero the first time is 50%. I flip a coin the second time, the odds of guessing a zero in the second position is 50%. I flip a coin a third time, the odds of my guessing a zero for the third position is 50%. And so on. When I get to the last position, I flip a coin, the odds of me guessing a 1 is also 50%.
Let’s consider a different sequence, 1010101010. My odds of guessing the first 1 is 50%. The odds of guessing the second 0 is 50%. And so on.
As long as each coin flip (or monkey choice) is independent of the previous flips, the distribution of PVS cases doesn’t matter.
March 30th, 2005 at 6:33 pm
According th this study (conducted by the Royal Hospital for Neurodisability in London) of 40 patients that had been diagnosed as being in a vegetative state, Neurologists diagnosed 17 (43%) incorrectly.
Among their conclusions:
# Many patients who are misdiagnosed as being in the vegetative state are blind or have severe visual handicap; thus lack of eye blink to threat or absence of visual tracking are not reliable signs for diagnosing the vegetative state
# Any motor activity, no matter how slight, that can be used for communication by the profoundly disabled patient should be identified at an early stage and repeated at regular intervals
# Identification of awareness in the presence of profound and complex neurological disabilities requires the skills of a multidisciplinary team experienced in long term management of disability due to brain damage
Obviously, they’ve never met the esteemed Dr. Cranford, who was able to diagnost Terri in 42 minutes.
March 30th, 2005 at 7:20 pm
“b/c that’s all that’s available in the Schiavo case.”
there’s only 1 cut from schiavo?
March 31st, 2005 at 1:09 am
OK Jody,. Yes, you may have a 50-50 chance. So, let’s say you get it exact – 50% – or get lucky with 56%. You lose, you need 60%. Now, CodeBlue is betting that the people he is challenging won’t do better than flipping a coin by more than something less than 10%, because to do so they would have to be able to “cheat” by knowing the outcome and that they cannot do so because they do not have the knowledge. I still think – and I am basing this on experiences as a patient – his money is safe. Get outside your area of expertise, and you’re a layman no matter how brilliant – think Einstein and quantum mechanics, which he refused to learn.
At the time of my original objection, I had thought the challenge was getting 60% of the PVS cases. I not only stand (well, sit) corrected, I had not thought that out since by saying ALL were PVS you would get 100% of the PVS cases… Oops…
March 31st, 2005 at 1:18 am
Oh, also, the BMJ recently published (see *point2point*
from the BMJ – “Results: Of the 40 patients referred as being in the vegetative state, 17 (43%) were considered as having been misdiagnosed; seven of these had been presumed to be vegetative for longer than one year, including three for over four years.”) what seems a schockingly high misdiagnosis rate of PVS. Heck, even Jody’s relatively modest 37% shocks me (sorry, I missed that before). More than one in three? Golly, wouldn’t flipping a coin to decide be not much (50% vs 40%) larger? Just what the heck is going on here?
March 31st, 2005 at 1:20 am
apologies to John from WuzzaDem. I have got to learn to carefully read comments before posting.
March 31st, 2005 at 2:26 am
I suppose in re “and that they cannot do so because they do not have the knowledge” CodeBlue doesn’t say he does – he says NOBODY does, including himself.
March 31st, 2005 at 8:06 am
Two quick notes:
I imagine I’m with you John in thinking that 1) an innocent woman was killed today 2) the courts were amazingly incompetent and downright hostile to the best interests of Terri, 3) the courts have a nasty case of judicial supremacy, and 4) the diagnosis of PVS is too flawed to put someone to death basedon that diagnosis. I just think that CodeBlue has put up a large chunk of money in what I think will be a losing proposition.
The 56% comes from the fact that CodeBlue pays out at 4:1. With a 56% chance of guessing right, 25% of the time you’ll get at least 60 right. So 56% is the break even point that CodeBlue should be taking into consideration with respect to his financial prospects out of this wager. And 56% is mighty close to 50%. If there are just 12 CT scans out of 100 that are unmistakeably not PVS, then they can blindly guess at 50% for the rest and get the expectation of 56% correct, 20 unmistakeable images and CodeBlue can expect to lose each trial.
Of course, if CodeBlue is willing to treat this as a loss leader to illustrate a point (I don’t see the neurologists doing much better than 65% which is mighty bad), that’s his call. I just felt compelled to inform him of his financial prospects so he was making an informed decision. (which is why I emailed him my post as soon as it was up)
December 12th, 2007 at 10:48 pm
I need to know the bare basics to read a CT scan.
The written reports sometimes leave things out. I’d like to see what they are talking about This is my body
and it’s my life at stake. So I have my CT and MRI s on disks. But, I’m needing some basic information to understand anything.
Please give me a hint where to look
December 16th, 2007 at 12:30 am
Six years of grade school. Six years of high school. Four years of college. Four years of medical school. Three years of residency. It’s a start. Good luck!