As the author of the story, please allow me to point out that I interviewed many more people than were quoted in the final version of the piece, as any serious journalist would, including NIH staffers, the long-time head of placebo policy at the FDA, several experts on the history of placebos in medicine and drug development (including Anne Harrington, author of the most widely-respected books on placebo, and Ellen Langer, author of a new and relevant book called Counterclockwise), pharma anthropologist Andrew Lakoff, as well as trial managers, doctors, and others. I also disagree with you that the academic researchers I spoke to are “in the tank” for Big Pharma. Most of them felt quite ignored by Big Pharma until recently, and financed their own experiments through their own universities.

If publication bias fully accounted for the phenomenon I describe in my article, why would Lilly, Pfizer, Sanofi-Aventis, Johnson & Johnson et. al. be investing millions in the data-mining project I describe in the final section of the piece? Do you think they don’t know they have a publication bias? Do you think that they or I don’t read the NEJM?

The problem, as I make clear in the piece, is not restricted to antidepressants, but is affecting many drugs for many kids of disorders, including Parkinson’s disease.

I do appreciate your interest in the subject and my article.

Steve, thanks for posting. Your article addresses a really interesting topic in medicine and there’s a lot of very good material in it. Your discussion of Benedetti’s work on the effect of Alzheimer’s on the placebo effect was especially good. Also, I apologise for writing that you’d only interviewed three people. I didn’t mean to imply that you hadn’t worked hard on the article. It was a very poor choice of words on my part and I’ve corrected it. Sorry.

However…

(1) The evidence for increasing placebo effect in depression is not nearly as strong as presented. You mentioned two studies showing an increase in placebo effect over time. I assume you are talking about Rief 2009 and Walsh 2002 (or possibly even Kirschner 1998 which did not directly address the subject but included data that matches). These papers are controversial and have many opposing papers, such as Hróbjartsson & Gøtzsche 2001 who concluded that the placebo effect is negligible for all practical purposes (I think H&G are wrong, btw). Even within Rief 2009, there is a great deal of reason to doubt the finding. Some placebo effects increased over time, but so did the drug effects. And some placebo effects, most notably on self-reported depression, did not change. Since self-reporting is arguably the most important scale there is for depression, this would tend to indicate that the placebo effect has not changed all that much in 25 years. Rief herself says, “The effect sizes in placebo groups in 2005 were more than twice as great as those in 1980, but only for observer ratings, not for patient self-ratings. The result was partly due to increased homogeneity of samples of recently published trials.”

(2) I could not find from Rief’s article any discussion on the different placebo arms. Most modern trials would compare drug + psychotherapy to placebo + psychotherapy. Since we now know a lot more about which forms of psychotherapy are effective, one would expect the placebo arms of trials to show an increase in effect size over time even if the specific placebo effect is constant.

(3) Even if one accepts that there is an increase in the placebo effect, Rief’s paper provides several plausible explanations that have do not require a mysterious society-wide placebo enhancement over time. Bob Park provided another. I provided a third (although Rief disputes my hypothesis). Your article presented no alternative explanations.

(4) I wouldn’t accuse these researchers of being “in the tank” as it were. But the fact is they are being paid by pharmaceutical companies. And we know that this has an effect on research findings. One of your interviewees is now a VP at Merck, which means not only is he being paid by Merck, but he will be receiving stock options as well. I have worked for Merck myself and I never met anyone there whom I would accuse of being manipulative or deceitful. And yet Merck was behind the Vioxx debacle. You don’t need to be “in the tank” to have your research influenced by your sponsors.

(5) Having said that, I have no sympathy whatsoever for the drug companies in the matter of publication bias. If they want to reduce publication bias, they can publish their freakin’ data.

(6) I have no idea what the drug companies hope to achieve with their data-mining project. I very much doubt, however, that is has much to do with preventing publication bias which the pharmcos could fix by the much cheaper option of open disclosure. To quote your article, “In typically secretive industry fashion, the existence of the project itself is being kept under wraps. NIH staffers are willing to talk about it only anonymously, concerned about offending the companies paying for it.” Is this seriously meant to convince me of their noble intentions?

Best wishes, Chris