Listening to Prozac, but Hearing Placebo
The Emperor's New Drugs?
Posted Feb 20, 2010
Irving Kirsch is an expert in statistics and in clinical trial methodology. In The Emperor's New Drugs, he absolutely dismantles the case for antidepressants as a pharmacologically effective treatment.
The cornerstone of the book is a careful analysis of a vast database of drug company data. Using the Freedom of Information Act, Kirsch managed to get the data that the drug companies had sent to the FDA in the process of getting their medications approved.
Some of the key points include:
1) Nearly all the benefit of antidepressant medications can be attributed to the placebo effect. When all of the data are arrayed, Kirsch finds that common antidepressants barely beat dummy pills. More than half of the clinical trials sponsored by the pharmaceutical companies showed no significant difference at all between drug and placebo. As important, the benefit of the medication over the placebo is not clinically meaningful. The overall advantage amounts to 1.8 points on the 54-point scale used to gauge severity of depression (through questions about mood, sleep habits, and the like). For example, your mood could be just as rotten and you could be just as bothered by insomnia, poor concentration and all of the other symptoms of a depression, but if you're little less fidgety during the assessment, that's 1.8 points right there.
(2) When antidepressants do beat placebos in clinical trials, it is actually another version of the placebo effect. While the clinical trials are supposed to be double blind (neither patient nor evaluator knows who is getting drug and who is gettting placebo), patients in clinical trials can often correctly guess that they are receiving the investigational drug because they experience pronounced side effects. Knowledge that you may be receiving a potential cure enhances the placebo effect.
(3) Pharmaceutical companies and the FDA are not being intellectually honest. The pharmaceutical companies are given wide latitude concerning which clinical trials they want to publish and which data they want to include in the publications. This is important because many of the clinical trials yield negative findings (i.e., no differences between drug and the dummy pills). If a drug company does 10 studies and only 2 support the efficacy of the drug, they can elect to publish only the two. There is no law against cherry picking the data. The FDA does not compel disclosure of all of the unsuccessful trials. This leads to inflated perceptions about the drugs' efficacy from both doctors and the wider public.
Kirsch has written a book for grown-ups. He writes clearly and bends over backward to explain technical language. He refrains from heated rhetoric even when discussing the most controversial findings. Kirsch is a lover of data. If he ever goes beyond the data, he labels his statements as such. The data leads him to conclude that psychotherapies provide treatments for depression that are just as effective in the short-run and more cost-effective in the long run than antidepressants. I expect many readers will be convinced. The measured tone and laser-like focus of Kirsch's compact book are in some sense the exact opposite of Gary Greenberg's Manufacturing Depression, which I described as "a rambling, blistering, entertaining, attack on the biomedical disease model of depression." Kirsch is as sober as Greenberg is lyrical, but the net effect is not all that different. Read together, these new books are a powerful one-two punch against psychiatric and pharmaceutical orthodoxy.
Ironically, Kirsch's matter-of-fact argument for the power of the placebo effect may ultimately undermine the clinical effect of antidepressants. If we lose faith that they work, they won't work.