>>> You're statements were vague enough that I'm not quite sure what you mean by 10 mths more survival 80% <<<
You're right , it was a poorly constructed example. I was trying to present a choice in a simple way that might be the way it would be explained to patients , without getting into statistical detail.
My point was only to say that many terminal patients might look more favorably on treatments that showed a larger average effect size , though with more variability , smaller trials , etc. , and thus a higher probability that they were due to chance , than on treatments that showed a small effect size but with a high reliability that the effect is real , if this treatment choice could be explained to them in easily understood terms. Something like a toteboard at the racetrack , showing continuously updated odds and payoffs , might be useful. ;)
PGS said that docs will often disregard other studies if they have doubts about them , in favor of treatment strategies they have confidence in and are more familiar with. The current flood of support from patient groups in favor of Provenge approval is evidence that patients often have their own ideas about such things , and it seems their ideas are at odds with at least a few of those docs , and more than a few of the statistical purists.
Market Data 
Markets 
AI
