Dendreon Corp fka DNDNQ

[DewDiligence]

>I am actually thinking that it may be better for DNDN to allocate a very LOW alpha for this look, for example, a threashold of p=0.0001, thus keeping the entire 0.05 alpha for the final look. Obviously, in this case, the interim look will NOT be stat-sig, but the major merit here is that we minimize the risk that Provenge will not be stat-sig at the final look (2010) while we obtain an "official observation" at the data.<

Ordinarily, a trial protocol does not allow the DSMB to reveal the statistics of an interim analysis to the sponsor other than to say whether the analysis met the threshold for halting the study. However, if the interim look occurs after all patients have been treated and are merely being followed to measure survival, it’s more likely that the SPA would allow for the statistics to be disclosed.