>>> Not sure in this case, but don't noncrossovers often do worse than crossovers in trials because at the time they were allowed to crossover they were ... (i) dead or (ii) in very bad shape or (iii) simply not fighting as hard as the crossovers for a few extra months of life? <<<
micro,
Almost always , I think , those in group (i) do poorly.
For the other groups , I think it would depend a lot on the nature of the crossover treatment offered. A brutal tx., like most chemos , would probably cause someone who was already feeling very sick to think twice. On the other hand , a relatively benign tx. like Provenge might appeal to the sicker patients even more than those less sick , who might choose to go straight to chemo instead of Provenge.
At least walldiver's numbers stack up in favor of Provenge , whatever the confounding influences might be. Better than having to explain the reverse situation.
I wish all the DNDN longs luck tomorrow , as well as patients and others with an interest in seeing it succeed. My vote would be for a conditional approval like the USDA granted to the canine melanoma vaccine recently. That would serve all interests best , IMO , since Provenge availability would not be delayed , yet confirmatory efficacy data from 9902B would be required to get permanent approval. The FDA should look to the USDA for help on this , if needed. ;)
AI
