A cheerfully depressing investigation: “Which Placebo to Cure Depression?”

This study pokes a pointed stick into lots of questions about medicine, science, and scholarship. The lead author writes (in a note to us) that

“It addresses a paradox in modern medicine: antidepressants are often considered to be mere placebos [1] despite the fact that meta-analyses are able to rank them [2]: it follows that it should also be possible to rank different placebos, which are all made of sucrose.”

The study:

naudetWhich Placebo to Cure Depression? A Thought-Provoking Network Meta-Analysis,” Florian Naudet [pictured here], Bruno Millet, Philippe Charlier, Jean Michel Reymann, Anne Solène Maria and Bruno Falissard, BMC Medicine, 2013, 11:230. The authors, at INSERM, Université de Rennes, and other institutions in France, report:

Antidepressants are often considered to be mere placebos despite the fact that meta-analyses are able to rank them. It follows that it should also be possible to rank different placebos, which are all made of sucrose. To explore this issue, which is rather more epistemological than clinical, we designed an unusual meta-analysis to investigate whether the effects of placebo in one situation are different from the effects of placebo in another situation….

Two network meta-analyses were run to indirectly compare response and remission rates among three different placebos: 1) fluoxetine placebo, 2) venlafaxine placebo, and 3) venlafaxine/fluoxetine placebo (that is, placebo compared to both venlafaxine and fluoxetine). Publication biases were assessed using funnel plots and statistically tested.

RESULTS. The three placebos were not significantly different in terms of response or remission. The antidepressant agents were significantly more efficacious than the placebos, and venlafaxine was more efficacious than fluoxetine. The funnel plots, however, showed a major publication bias.

CONCLUSION. The presence of significant levels of publication bias indicates that we cannot even be certain of the conclusion that sucrose equals sucrose in trials of major depressive disorder. This result should remind clinicians to step back to take a more objective view when interpreting a scientific result. It is of crucial importance for their practice, far more so than ranking antidepressant efficacy.

BONUS (from 2009) “Placebos Are Getting More Effective. Drugmakers Are Desperate to Know Why

BONUS: The 2008 Ig Nobel Prize for medicine was awarded to Dan Ariely of Duke University (USA), Rebecca L. Waber of MIT (USA), Baba Shiv of Stanford University (USA), and Ziv Carmon of INSEAD (Singapore) for demonstrating that high-priced fake medicine is more effective than low-priced fake medicine.

BONUS song (possibly related in some complex, yet simple, way): “A spoonful of sugar“:

UPDATE (June 30, 2014): Two updates, from some of the same authors:

Does reductio ad absurdum have a place in evidence-based medicine?” Florian Naudet and Bruno Falissard, BMC Medicine, 12:106, June 25, 2014.


Placebo for depression: we need to improve the quality of scientific information but also reject too simplistic approaches or ideological nihilism,”Andrea Cipriani, John R Geddes, BMC Medicine, 12, 105, June 25, 2014.