Blinding, unblinding, and the placebo effect: an analysis of patients' guesses of treatment assignment in a double-blind clinical trial.

We administered a questionnaire to assess maintenance of patients' blindness at the end of a double-blind clinical trial of Osmotic Release Oral System phenylpropanolamine (PPA) vs. placebo in mild obesity. Seventy-four percent of placebo participants and 43% of PPA participants guessed their treatment correctly. Appetite control was the most frequently reported basis for guessing PPA, even by placebo participants. Lack of adverse drug reactions was the most frequently reported basis for guessing placebo, even by PPA participants. Participants receiving either PPA or placebo and guessing PPA lost more weight, had less diet difficulty, and had more adverse drug reactions than had participants receiving either PPA or placebo and guessing placebo. Although blindness was probably maintained in the PPA group, the placebo group seems to have been, at least at the study's end, unblinded. These results suggest that in double-blind studies, differences in outcome or incidence of adverse drug reactions may act as unblinding factors.