Re: Interim looks
>Few marketed drugs have had trials unblinded early.<
That’s because few marketed drugs had a pivotal trial in which as much as 0.015 of p-value was allocated to an interim look, the amount that was allocated in BIOM’s Theratope trial.
I cannot stress strongly enough that missing an interim look with a large p-value allocation (e.g. 0.015) is a disaster that virtually ensures the trial will not later have a stellar outcome. You simply must sell in such a case, and that’s what I recall posting on the BIOM board when this event happened.
Xigris, which you cited as a counter-example to my logic, is not a counter-example at all. A counter-example would be a drug that had a large p-value allocation for an interim look, missed its interim endpoint, and yet went on to hit the final endpoint and become a blockbuster drug. Perhaps there is even such a case, but there surely aren’t many of them. Successful investing is about gauging probabilities and having the odds on your side.
>BIOM's most calamitous blunder IMHO was "improving" the prospects for the pivotal trial by profiling "weller" patients than the ones they were successful with in the preliminary trials.<
I’ll give you this point because I do not contend that the large interim p-value allocation was BIOM’s biggest blunder. I merely contend that BIOM’s missing the interim endpoint that had a large p-value allocation was a strong sell signal.
To reiterate the point in my previous post: these days, most sponsors are smart enough to avert the kind of mistake that BIOM made: either they don’t schedule interim looks at all, or they allocate only tiny amounts of p-value to them. Regards, Dew
Market Data 
Markets 
AI
