Novel approaches to drug-placebo difference calculation: evidence from short-term antidepressant drug-trials

Abstract

The calculation of antidepressant-placebo difference (50% - 30% = 20%) in drug trials is based on the postulate that all placebo responders should be 'automatically' antidepressant responders, an assumption that has been never been specifically investigated and substantiated. However, some studies show that a clinically significant part of placebo responders are also antidepressant nonresponders. The traditional calculation of antidepressant-placebo difference seems, therefore, to be wrong because of an inherent fundamental bias resulting in a marked overestimation of the placebo effect. If the mechanism of action of antidepressant and placebo are independent (unrelated) and the randomization results in two identical (homogenous) groups of patients, then the two basic principles by which the evaluation of potentially useful drugs are based on, such as the relationship between antidepressant response and placebo response rates, would also be independent. In this case, only 50% of placebo responders are antidepressant responders (and another 50% of them are antidepressant nonresponders) and the antidepressant-placebo difference would be 50% - 15% = 35%, instead of 50% - 30% = 20%, as calculated by the traditional method. For real interpretation, decisions that have been made on traditional drug-placebo difference evaluation should be recalculated and reviewed, not only in major depression but also in other psychiatric and medical disorders where the drug-placebo difference is in the same magnitude.