Biotech Values

[iwfal]

Give me one example of this happening from registration trials.

This is in response to my post that I agree generally that most DSMBs allow even very low p values to go to the end. Not halt unless there is an interim. So I think we're mostly in agreement. ? (And note this more conservative view of mine is a shift from my earlier view, and was caused by the discussion in this thread. So thanks.)

FWIW I think our only disagreement is that I am less absolutist? But that is based upon the fact that I've read a ton of statistical reviews, papers, ... and have bumped into more than one unplanned halt. But I've never made a study of them (e.g. Categorized them by NDA vs SNDA vs number of zeros in the p value vs whether the endpoint was OS or some lesser thing, ...)