Re: Cox summation argument
>Ok, I think we are reaching the essence. If there existed a tool that incontrovertibly compensated for imbalance then there is no need to run a 'well-balanced trial'. Cox Regression is, to a large extent, that tool.<
Once again, you’re speaking theoretically while I’m speaking in terms of how Cox is seen by the FDA reviewers and advisory-panel members.
From a practical standpoint, Cox does not obviate the need for a sponsor to run a reasonably well-balanced trial. (Whether it ought to is a different question.)
Has the FDA or any regulatory body ever accepted Cox analysis as supportive of approval for marketing when the Cox adjustment lowered the p-value as colossally as it did in DNDN’s 9902a trial? If the answer is No, then DNDN will be blazing new ground statistically with its 9901/9902a BLA.
I admire trailblazing science but, when it comes to my investments, I do not necessarily desire to be associated with trailblazing use of statistics.
Market Data 
Markets 
AI
