The effectiveness of placebos in the treatment of depression: A meta-analysis

Date of Completion

January 1995


Health Sciences, Mental Health|Health Sciences, Pharmacology|Psychology, Clinical




A meta-analysis was performed on 19 studies comparing antidepressant medication to placebos, and on 19 studies comparing psychotherapy to no-treatment or wait list control groups, to find the magnitude of the placebo effect. Mean weighted effect sizes, based on improvement from baseline, were computed for each of the four conditions. The magnitude of the placebo effect was determined by subtracting the effect size of the no-treatment control group from the effect size of the placebo group. The drug effect was determined by subtracting the placebo effect size from the drug effect size. Results indicated that the magnitude of the placebo effect is 0.77 standard deviations and is significantly greater than the effect size of spontaneous recovery or the passage of time (0.35 standard deviations). Moreover, the placebo effect accounted for 50% of the drug response, the drug effect accounted for 27% of the drug response, and spontaneous recovery accounted for 23% of the drug response. A search for moderating variables revealed that type of drug, length of treatment, method of calculating effect size, or rating source did not moderate the drug effect sizes. Two studies comparing monoamine oxidase inhibitors to placebo appeared to be responsible for the large variance within the drug group. Results were discussed in terms of the implications for the use of placebos in research, the notion of placebo responders, the probable underestimation of the placebo effect in the present study, and the ethics of prescribing placebo for depression. Placebo appears to still be considered an artifact rather than an autonomous effect, and many investigators are trying to "cope" with it rather than "exploit" it in research. ^