Replies to post #5100 on Dendreon Corp fka DNDNQ

[eagle-eye]

Below is a link to a good review of DSMB's.

www.ncchta.org/fullmono/mon907.pdf

Read the answer to Question 17 (Page 38)

[exwannabe]

Fuzzy interim goals make no sence.

If you think a higher P value should cause an early unblind, then that should have been in the SPA. The company has the flexability to do this.

Near misses are just an excues for poor thinking/planning. If they want a P=.011 (for example) to cause termination, just say so in advance formally. Why allow some wishy washy logic on a very simple decision tree?

I do think there could be cases where other compelling data could effect the issue, but even so the DSMB is not going to lightly override the trial design.

In short, I have no idea if it has ever happened but I find the concept very illogical (and certainly very un-beurocratic).