Re: Penalty for unblinding: confident in how big?
Post# of 148162
Nader briefly said the penalty was .005, that is, we'd have to meet p<.045 at full enrollment for significance. I wonder if that was "top of the head."
Say the DSMC sends the (brief) message: "continue - this size benefit at full enrollment would be significant." Making the details behind that statement public would vastly accelerate enrollment.
I haven't checked a lot of cases, but the penalty as stated looks like one less death in the 260 person treatment group. Alternatively, I'm guessing you could maintain full stat power by agreeing with the FDA to add another dozen people to the trial.