← The Casebook

Comparative specimens

01

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)

02

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

03

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)

04

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

05

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.

File your own case

Open the same case on your own draft.

Paste a memo, a research draft, or a strategy argument. It is scored against all 175 cards, and the strongest two or three risks come back with the evidence quoted and one practical next check.

Open a case on your draft →