acc. 2003.318
Placebo Effect
The placebo effect is a genuine improvement in symptoms produced by the psychosocial context of a treatment, expectation, conditioning, and the therapeutic ritual itself, rather than by any active ingredient.
Provenance
Henry K. Beecher, 1955 — The Powerful Placebo
A genuine placebo effect is robust for subjective pain and well replicated in experimental analgesia (Vase 2002: mean d=0.95 in mechanism studies vs ~0.15 in clinical-trial control arms). But its clinical magnitude is contested: Hróbjartsson & Gøtzsche's 2001 NEJM meta-analysis (114 trials) and 2004/2010 Cochrane updates found little effect on objective or binary outcomes and only small subjective benefit that shrank in larger trials. The open-label placebo finding has replicated in a larger IBS trial (Lembo et al. 2021, n=262), though with debate over controls. Net: real for some subjective outcomes, weak-to-absent for objective ones, and historically overstated since Beecher.
Inscription
“We found little evidence in general that placebos had powerful clinical effects... Outside the setting of clinical trials, there is no justification for the use of placebos.”
Specimen notes
Two non-exclusive mechanisms dominate: (1) expectation — conscious anticipation of benefit, supported by Levine et al.'s finding that the opioid antagonist naloxone reverses placebo analgesia, implicating endogenous opioid release; and (2) classical conditioning — prior pairing of treatment cues with real drug effects, which can produce placebo responses (e.g., in immune and hormonal systems) even without conscious expectation. The therapeutic ritual, provider warmth, and verbal suggestion modulate effect size. Critically, an apparent 'placebo response' in an arm of a trial is the sum of the true placebo effect PLUS regression to the mean, spontaneous remission/natural history, and Hawthorne-type effects; isolating the genuine effect requires a no-treatment control arm.
Conservator’s confidence
strong grounding
See also in this gallery
Comparative specimens
Levine, Gordon & Fields — naloxone reverses placebo analgesia. Naloxone increased pain in placebo responders, taken as evidence that placebo analgesia is mediated, at least in part, by endogenous opioids (endorphins).
Jon D. Levine, Newton C. Gordon, Howard L. Fields, 1978 · 51 patients (the design has been criticized as leaving room for alternative explanations)
Hróbjartsson & Gøtzsche — 'Is the Placebo Powerless?'. No significant effect on binary or objective outcomes; a small benefit on continuous subjective outcomes and pain, but the effect shrank as trials grew larger, suggesting small-trial bias.
Asbjørn Hróbjartsson, Peter C. Gøtzsche, 2001 · 32 trials with binary outcomes (3795 patients) and 82 with continuous outcomes (4730 patients) · SMD 0.27
Kaptchuk et al. — open-label ('honest') placebo in IBS. Open-label placebo significantly outperformed no-treatment control on global improvement, symptom severity, and adequate relief, challenging the assumption that deception is required.
Ted J. Kaptchuk, Elizabeth Friedlander, John M. Kelley, M. Norma Sanchez, Efi Kokkotou, Joyce P. Singer, Magda Kowalczykowski, Franklin G. Miller, Irving Kirsch, Anthony J. Lembo, 2010 · 80 patients (37 open-label placebo, 43 control)
Moseley et al. — sham (placebo) knee surgery. At no point did the real-surgery groups report less pain or better function than the placebo-surgery group, demonstrating a powerful procedural/placebo response to surgery.
J. Bruce Moseley, Kimberly O'Malley, Nancy J. Petersen, et al., 2002 · 180 patients
Kirsch et al. — antidepressants vs placebo (FDA data). The drug-placebo difference was below the NICE clinical-significance threshold except in the most severely depressed patients, and that gap was driven largely by reduced placebo responsiveness rather than greater drug response.
Irving Kirsch, Brett J. Deacon, Tania B. Huedo-Medina, Alan Scoboria, Thomas J. Moore, Blair T. Johnson, 2008 · 35 trials, 5133 patients (3292 drug, 1841 placebo) · d ≈ 3
On loan · documented cases
Sham knee surgery matched real arthroscopy for osteoarthritis · 2002
In a 180-patient randomized, double-blind trial, patients who received placebo knee surgery (skin incisions, no actual débridement) reported pain and function outcomes no worse than those who got real arthroscopic surgery, reshaping orthopedic practice.
Open-label placebo improved IBS even when patients knew the pills were inert · 2010
Kaptchuk et al. gave IBS patients sugar pills openly labeled as placebo; they still improved significantly more than no-treatment controls, a result later supported by the larger Lembo et al. 2021 trial (n=262), where 69% on open-label placebo reported clinically meaningful improvement.
Most antidepressant benefit is matched by placebo in FDA trial data · 2008
Kirsch et al.'s analysis of FDA-submitted antidepressant trials found the drug-placebo gap was below clinical-significance thresholds except in the most severe depression, with placebo arms capturing much of the measured response — fueling the 'Emperor's New Drugs' debate over how much of drug effect is placebo.
Handling instructions
Always interpret an inert-treatment arm against a no-treatment control, not against baseline alone; pre-specify objective outcomes; and quantify how much apparent benefit could come from regression to the mean and spontaneous remission before attributing it to the placebo.
Hróbjartsson & Gøtzsche (2001) showed that comparing placebo to no-treatment (rather than to baseline) collapses most of the supposed effect on objective and binary outcomes; Kienle & Kiene (1997) catalogued the confounds that made Beecher's estimate spurious.
Catch it in the act
Watch for claims that an inert treatment 'caused' improvement based only on before/after change in a single group with no untreated control. Genuine placebo effects show up mainly in subjective outcomes (pain, mood, nausea) and shrink under rigorous designs; if the 'effect' is on an objective measure (tumor size, blood chemistry, mortality) or the symptom would have improved on its own, suspect regression to the mean, natural history, or Hawthorne effects rather than a true placebo response.