Senti, what I am telling you is the idea that this misallocation was spread out over 10 years is not possible. Do I need to describe simple block randomization? Picture the first 20 patients being skewed 15 vs 5, guess what happens during the next two blocks of 20.
The more complex acceptable randomization techniques only go further in making sure balance by all available parameters is maintained. A 32 patient skew by randomization is impossible.
Either FDA did not allow the final 17 patients randomized to placebo in the trial (highly, highly doubtful) or they removed placebo. These are your options to get to where we landed. The company, patients and physicians would be completely unaware of what took place, officially, thus integrity of the blind maintained. How regulators choose to view the population is up to them, but the Intent to Treat was DCVax alone for the final cohort.
Anyone believing that an error n the randomization schema occurred is not living in the 21st century of clinical research, no offense intended.
Market Data 
Markets 
AI
