>…likely noise…given all of the other factors that affect whether the trial will be successful in the real world.<
This is a non sequitur, IMO. The fact that it’s extraordinarily difficult to predict this trial’s chance of success does not make a stated success differential more important or less important than its numerical value.
Perhaps you intended to say that DNDN’s power calculation may not reflect the true efficacy of Provenge, which is a valid and germane argument. As has already been noted, if the true efficacy of Provenge were considerably less than DNDN’s modeled efficacy, then the true power of 9902b at the final analysis would be lower than DNDN’s stated power figures (both before or after the SPA change). If this were the case, the delta power from the SPA change would be magnified by a multiplier effect.
(Naturally, it was to DNDN’s advantage to ignore this line of reasoning during the CC and, unsurprisingly, none of the analysts on the CC thought to go there or wanted to go there.)
Finally, since you say the 90%->88% loss of power from the SPA is “noise,” I would like to know how large the swing would have to be for you to characterize it as something more. Please bear in mind that virtually all phase-3 trials are powered in the 80-90% range. Regards, Dew
“The efficient-market hypothesis may be
the foremost piece of B.S. ever promulgated
in any area of human knowledge!”
