NorthWest Biotherapeutics Inc (NWBO)

[longfellow95]

I think you mean 'common treatment effect assumption'.

Which is not the same thing.

We have touched on this issue previously.
It's complicated and dry, and best left to statistical nerds, frankly.
RPSFT does make certain assumptions that makes it not ideal, it has to be said.

The RPSFT method estimates the difference in OS between treatment groups that would have been observed if crossover had not occurred. This method proportionally shrinks the estimated amount of additional survival conferred to patients who crossed over according to the treatment effect estimate for the experimental treatment. In its estimation procedure, the RPSFT method relies upon the randomization of the trial; it assumes that, in the absence of treatment in either of the randomized groups, average survival between groups would have been equal. However, some key assumptions, which are often unverifiable, need to be fulfilled[21]. First, the effect of experimental treatment on OS is the same irrespective of when treatment is given in a patient’s disease course, relative to the duration of time the treatment is received. Second, absolute OS benefit of an experimental treatment never exceeds actual treatment time. And third, all patients receive identical benefit from an experimental treatment. It is especially noteworthy that there cannot be subsets of patients particularly responsive to treatment.

http://eprints.whiterose.ac.uk/111326/1/LatimerN%20Renal%20Paper.pdf

Life will be easier if adjudicated PFS hits the target.
Which it should do, hopefully.