How effective was patient blinding in the Aspirin Myocardial Infarction Study (AMIS)?
The study suggests that patient blinding in the AMIS trial was largely effective, as only a small minority of patients could make informed, correct guesses about their treatment assignment.
The Aspirin Myocardial Infarction Study (AMIS) was a double-blind placebo-controlled trial to test the effect of aspirin on the survival of 4524 people who had experienced a prior heart attack. Shortly before their closeout visits, 400 of the participants were randomly selected to be interviewed concerning their perceptions of their treatment assignments; 380 were actually interviewed. A bare majority (52)% correctly identified their study therapy, 28% mistakenly named the alternative treatment, 13% declined to guess, and 7% specified extraneous substances. According to the proposed formula for evaluating the patient blind, only 24% of the sample made "informed" guesses regarding their therapy, while the remainder guessed in an uninformed way or not at all. Those who tested their capsules (usually be taste) showed proportionately more correct responses than the nontesters. Correctness also varied with the reasons for the subjects' guesses (e.g., side effects). Among the sample as a whole, most people were only moderately or less than moderately certain their guess was correct. Even among those who were in fact correct, only 18% were absolutely certain of their choice.
Howard et al. (Mon,) studied this question.