Psychiatric journal refused to retract flawed antidepressant study

The Journal of the American Academy of Child and Adolescent Psychiatry has refused to retract a study finding that the antidepressant Paxil was safe and effective in adolescents despite evidence of fabrication, plagiarism and failure to disclose conflicts of interest on the part of its authors, according to two academic researchers.

In a paper published yesterday in the journal Accountability in Research, Jon Jureidini and Leemon McHenry recount the myriad flaws of study 329, the same clinical trial that prompted a New York State Attorney General's lawsuit against Paxil's maker, GlaxoSmith Kline, in 2004. The story of that lawsuit and how it helped propel reforms in the way drugs are tested and marketed is the subject of Side Effects. In that book, I exposed how the authors of study 329 failed to fully disclose their lucrative financial ties with GlaxoSmithKline and how they misrepresented the data in the study to make Paxil look safer and more effective than it really was. I've also written about the study's extensive flaws in my blog here and here.

In this week's paper, Jureidini and McHenry concluded that study 329 violated JAACAP's written policy in several ways:

1. Failure to disclose conflicts of interest of "authors."
2. GSK concealed commercially damaging data (about the study).
3. Fabrication (creation of a strong false impression that one primary outcome measure was positive by deliberately confusing it with another measure).
4. Falsification (post hoc changes to secondary outcome measures and misrepresentation of severe adverse effects).
5. Plagiarism (submittting a ghostwritten manuscript).

The two researchers said they brought evidence of these clearcut violations to the attention of the journal's current editor, Andres Martin, but he has so far refused to retract the study. In their paper, Jureidini and McHenry said this case makes them wonder how many industry-sponsored publications of clinical research are equally flawed. As they note:

It is unclear to what extent one can extrapolate from Study 329 since relatively few industry-sponsored clinical trials have been exposed to this level of scrutiny. Given, however, the vast operations of "publication planning" conducted by pharmaceutical marketing, there is little doubt about the degree to which industry-sponsored research is suspect.

As a result, they reach the conclusion that other researchers have: that publication of industry trials should be banished from the journals. Instead, "pharmaceutical companies should be obliged to post results of their trials on websites, leaving journals to independently and critically examine the raw data and conclusions," they argue.

That's one solution. Here's another, long espoused by Dr. Marcia Angell, the author of The Truth about Drug Companies: prohibit pharmaceutical companies from funding randomized clinical drug trials and instead have those trials funded by the National Institute of Health and conducted by independent researchers who are not on the drug companies' payroll.