>As a matter of fact, a DSMB is empowered to look at the data at any time it feels called on to do so. It must have that power in order to properly monitor the safety of volunteers. The safety of volunteers is written into its very name.<
Interim looks for safety only do not exhaust any of the cumulative p-value, and hence they are a whole other matter for investors than interim looks for safety and efficacy.
>The DSMB advised the trial should continue. Why did they advise such a thing if the trial was surely doomed according to the deep calculations of Dew Diligence? :-) :-) Some trials have been discontinued in such a case.>
Your question—if you are serious and not just playing around—shows a lack of understanding of trial-halting decisions. Halting a trial early for lack of efficacy has its own pre-defined statistical analyses, which must of necessity be very stringent to avert halting trials that could turn out to be successful. Just because a trial is not halted for meeting the pre-defined efficacy threshold of an interim look does not imply that the trial should be halted for lack of efficacy.
>Long before the trial ended, before the interim look as best I can recall, a knowledgeable Ph.D. involved in numerous trials advised he had sold all his shares and advised others to do so after getting a look at a supposedly purloined copy of the trial design that was easily available. Wish I had listened of course.<
This was my source for the 0.015 p-value allocation at the first interim look.
Once again, if you know or think that an interim look with a fat p-value allocation has been missed, it is prudent to sell. It doesn’t mean a stellar outcome at the end of the trial is impossible, but it does mean that a stellar outcome is much less likely than it was before the interim miss. Unless you were emotionally attached to the stock, why would you want to play those kinds of odds?
Market Data 
Markets 
AI
