Here are a couple of sentences I found surprising:
The studies examined were conducted between 1987 and 1999 and covered Prozac, Paxil, Zoloft, Celexa, Serzone, and Effexor. They found, on average, that placebos were 80 percent as effective as the drugs.
Update: I wanted to pull up a link from the comments to a blog post the Public Library of Science that sheds more light on this issue.
January 2, 2009 at 1:05 am
80% does sound high, but I don’t doubt the effectiveness of a strategic placebo where mental health is concerned. I luckily don’t suffer from depression, but I have struggled with ADHD for most of my life and although chronic depression is a much much more serious condition, I think there might be some parallels in their patterns.
I can often be highly focused and productive without medication, but when I have a few off days in a row, the frustration can overtake me and I might go for days or weeks without even really trying to get focused–giving up before I even start. Taking a pill that I believe* will help me can be enough of a boost to get me started on something, and once I start being productive, there’s a positive feedback loop that can keep me going for weeks–if I’m lucky.
January 2, 2009 at 9:19 am
That article by Ezra Klein is horribly researched and 6 years out of date on the state of this issue. I recommend instead reading http://www.plos.org/cms/node/424 for the current issues around this. Among the things you will discover is that you no longer would have to submit a FOIA application to get drug trial information. Additionally, you will find that this bias in publication isn’t always the fault of drug companies as many have submitted this data for publication, but been rejected because ambivalent results are not considered exciting.
Ezra Klein’s data point is horribly out of date and a misrepresentation of the true state of research. And the things I mention don’t even take into account the changes seen since thanks to modifications to the Declaration of Helsinki. Presenting 6 year old research on 10+ year old drug trials as current is hardly appropriate.