KarinCA......According to the trial design the number taking the placebo is known. That is my point. It is 43% of those enrolled, and 57% was in the two drug arms of the trial. Now if i recall correctly there was 199 patients enrolled and 186 completed the trial. So how many were on placebo at the end? Well a range can be deduced by assuming the dropouts were all in the two drug arms.
So 199 x 43% is 86 of the 199 was originally assigned to placebo leaving 113 assigned to the two drug arms, and if 13 dropped out leaving 100 completing the two drug arms. So the final figures would be:
86 on placebo
100 drug arms
Now assume there were 45 of the above 186 by observation had 75% improvement at the end of the trial. By historical norms 10% [9 patients] of them could have been placebo leaving 36 out of 100 in drug arms having positive response.
Now about my assumptions:
A] Assumption is that all dropouts were in the drug arms. Is that realistic? Don't know, but it is being conservative.
B] Assuming total of 45 positive responses out of 186 patients which is about 25% of the total enrolled. Is this realistic? Don't know, but it is being conservative.
C] Assuming 9 positive from placebo group. Is this realistic? Historically from other studies it is realistic.
D] Assuming 36 out of 100 is that realistic? In my opinion it is conservative but i do not know.
The point is the total number of number of positive outcomes was knowable to the medical people involved in the trials by observing the results on the skin. And a very reasonable guess as to how many were on placebo would have been discernible. Thus enabling a very good estimate of the results even though the trial was double blinded.