yet people here assume the know the FDA was critical...
They - most people here - don't know, but they can read data, make educated guesses.
I asked same question before and after 2nd line NSCLC data blowup in Sept 2012. I haven't got a reasonable answer from anyone who believed the result. I'll post the questions again:
In Sept 2012 data release, many people noticed placebo arm had more bleeding events than 1 mg arm which was odd. Later the company said placebo arm and 1 mg arm were mixed up. Assume this was true, based on AE profile of the 3 arms, it implied placebo arm got more study drug than 1 mg arm because 3 mg arm had more bleeding events than placebo arm. However, 1 mg arm had better efficacy (statistically significant HR) over placebo arm than 3 mg arm over placebo arm.
If people can't see implications from above, well, good luck in waiting a few years to find out in ph3.