Three months ago, the University of Pennsylvania denied allegations made by one of its professors that several other academics – including his own department chair – allowed their names to be added to a medical journal manuscript, but gave control of the contents to GlaxoSmithKline. As it turns out, a medical communications company called Scientific Therapeutics Information played a large role in preparing the paper, which was published in 2001.
The study, which was funded by the drugmaker and the National Institutes of Health, looked at the impact of the Paxil antidepressant on patients with bipolar disorder. The university acknowledged the study was ghostwritten, but did not take any action, because such activity was standard practice between 1998 and 2001, when the study was prepared (back story). Also known as study 352, it was later published in the American Journal of Psychiatry, although STI was not mentioned.
Having filed a complaint about scientific misconduct a year ago with the federal Office of Research Integrity, Amsterdam is now charging that the university "went out of its way to find a path to whitewash the conduct of its own employees," instead of admonishing his colleagues for allowing their names to be appended to a manuscript that was drafted and revised by a ghostwriter, according to a letter sent yesterday to the ORI (here is the letter).
The letter, in which Amsterdam argues the episode is about "important issues of public health and scientific ethics," may raise the heat on a major university at a time when the controversy over ghostwriting continues unabated. Despite various efforts, including a set of recommendations released last month by a group of medical journal editors and pharma execs called the Medical Publishing Insights and Practices initiative (see here), a solution remains elusive.
Meanwhile, various episodes have stained the research associated with various drugs and, in the process, placed academics, medical journals and drugmakers on the defensive (read more here). This particular Paxil study, for instance, has been cited more than 250 times over the past decade, which is also problematic for another reason - Amsterdam claims the paper misrepresented post hoc analyses with a preordained outcome, made unsubstantiated efficacy claims, and failed to adequately report adverse events and safety information. In his complaint, he also claimed he was a co-principal investigator, but was excluded from the final data review, analysis and publication (here is the letter). He charges that his UPenn colleagues, Dwight Evans and Laszlo Gyulai, committed scientific misconduct by allowing their names to be placed on a ghostwritten paper. And in his missive to the ORI this week, Amsterdam adds that Evans may have withheld evidence from the UPenn investigating committee.
Last year, in a response to the charges, Gyulai wrote this to Amsterdam: "I was the primary investigator of the Penn site and, as you know, I worked on early drafts of the paper. I did not determine authorship, and as you know, the paper was taken away from me as first author during the writing process. However, I regret that I did not discuss the issue of authorship with you. I agree with you that SKB should have circulated the paper to all participants. I only saw the final draft shortly before it was submitted when only minor changes could be done" (see this).
In attempting to debunk the recent UPenn investigation, Amsterdam notes the committee contended that any study flaws, such as insufficient sample size that resulted in an underpowered study, was the responsbility of a Rosemary Oakes, a Glaxo biostatistician and a named author. But he maintains that every study author had equal responsibility to ensure the final paper was accurate. The committee, meanwhile, did not interview Oakes, according to the letter sent to ORI.
"The evidence appears to indicate that the academic authors were almost completely uninvolved in the data analysis and manuscript preparation, and were merely serving as key opinion leaders on the manuscript for commercial and marketing purposes," the letter states.
There is more. The letter goes on to say that, while Jack Gorman was serving as editor of the American Journal of Psychiatry and "closely involved in the editorial review process of the 352 study," he was also on the Glaxo speaker bureau and served as a consultant to Glaxo for promoting Paxil, "including participation in GSK-funded symposia, advertising videos, GSK-funded lectures, co-authoring GSK-funded articles in medical journals and book chapters" (see here).
We have asked UPenn and the ORI for comment, and will update you accordingly. UPDATE: A UPenn spokeswoman wrote us to say that nothing has changed since last March, when the university last responded to inquiries. So here is the same statement.
ghost pic thx to mattwi1s0n on flickr