I would have thought that when a placebo is given, it can't work if the subject knows.
Clinical studies use placebos as a 'control', but adjustments have to be made for the average group -- to compensate for various factors that the dummy drug causes.