Program Survival Bias - why it's really just BULLSHIT.
I thought of an interesting vignette to prove it so.
Supposing you are such a financial genius that through mergers and acquisitions you suddenly owned a collossal clinical trial program that included ALL ongoing biotech P3 clinical trials.
What to expect ? There is uncertainty about these drugs, as only P1s and P2s have happened, as no P2 is designed to give certainty (because a trial designed large enough to give certainty is not a P2 but is already a P3).
So, prospectively, you expect some P3's to fail, and you dont know which ones. If the Program Survival Bias theory is any good, can it help you at this point ? No, of course not, you will always have uncertainty until you do the P3s. Even if you resigned, or bought in extra consultants, etc - the P3s still have to be done, and sure as the sun will rise tomorrow in the east, some will fail.
Now let's look at the program results in the rear-view mirror. Some of the P3's will have succeeded, and some will have failed.
So, post-hoc, is it legitimate to single out the P3 failures, and calling them errors in judgement owing to a Program Survival Bias "phenomenon" infecting human CEOs ???
Of course not. looking only at the each failure is just a post-hoc, selective, result - and the other sub-group of successful trials is ignored.
Remember, prospectively there was only ONE large clinical trial program, only ONE prospective "group" of trials.
And so, it is just a post-hoc, sub-group analysis to dress up a "theory".
ps: .......ocyanblue I remembered you from DNDN days recently on another board, and remembered also your contempt for this Program Survival Bias nonsense. I thought you might like this "proof". It works. Take care everyone with DNDN, risky days ahead IMHO.....but you never know I suppose.....
"....on the biotech battle-field, you need some élan...."