Placebos have powerful effects on the mind and the body. They can produce improvements, cause side effects, and reverse the actions of medications. Indeed, placebos are so powerful that medications are approved by the FDA only if they are proven better than placebos in clinical trials. For example, Namenda was not approved for treatment of mild to moderate Alzheimer’s disease because it did not do better than placebo. But treatment versus placebo effects is complicated as is demonstrated by teasing out whether antidepressants work better than placebos in treating depression.

Irving Kirsch (2008) completed a technical review (called a metaanalysis) of antidepressants and placebos as they are compared to no treatment. Among his findings were that some get better by treatment with active medications. Some get better with placebos (as many as 30% of study participants by some estimates). The difference between the effects of placebos and antidepressants is very small.

The effects of antidepressants are often below standards of clinical significance (make you actually feel better) when compared to placebos for those with moderate levels of depression. Clinically significant differences are only found in smaller studies and only for those with severe depression. On average, placebos produce effects when compared to no treatment that are twice the size of effects of the active medications compared to placebo effects.

The issue is that for any drug when there is a response to a drug it may be due to the placebo effect and not a result of the chemical composition of the drug. All treatments need to be proven better than both no treatment and a placebo to be considered effective. This is true of medications as well as supplements (where the FDA has no control). But there’s more to the story.

As Kirsch points out his analysis does not prove that antidepressants don’t work. His review only proves that the effects of the medication compared to placebo are small. This is not an inherent problem with the medications but rather a limitation of clinical trials. They are flawed by several factors such as short duration, unrepresentative samples, and breaking “blind” as researchers and subjects can tell active treatment by the side effects. We need better designed studies that explore who benefits (e.g., women versus men, older versus younger) and then be able to target medications.

So should you take an antidepressant? Unless your depression is severe, there is no clear answer. Does it really matter if it is a placebo effect or a treatment effect if you feel better? We need to better understand and use the power of placebos.