One thing, not mentioned in this thread yet, unless I missed it, that occurs nearly always in *studies* by companies with a financial motive... is to design the study from before it starts, specifically to produce any favorable outcome\results that may be desired.
Just manipulate the placebo or some other vital part of the study and bingo... our *product* is the best thing since sliced bread even if it kills a few 10's of thousands.