> As someone who has experience with antidepressants that goes beyond looking at numbers I can assure you that effect has been established very clearly. And it has nothing to do with placebo, only the second medication was the one that worked - it did more in three days than the first one after months on the highest dose.

I think this would still be consistent with it being a placebo. (Not saying it neccesarily is, just saying we would still expect to hear these types ancedotes even if it was a placebo)