JAMA Editors Need to Come Down off their High Horse--and 'Fess Up

The story about JAMA has been brewing for some days now, and every time I go to write about it, a new chapter emerges. Let me try to walk you through what's happened so far.


The story goes back to this article (note: subscription may be required to access some JAMA links):


http://jama.ama-assn.org/cgi/content/full/299/20/2391


Dr. Robert Robertson (Univ. of Iowa) and colleagues studied two treatments for patients who had just had a stroke (and who were therefore known to be at high risk for depression)--counseling and escitolapram (Lexapro). As the study showed that Lexapro was superior to placebo, Robinson gave a number of interviews and made several statements urging that all patients post-stroke be considered for drug therapy.


Enter Jonathan Leo and Jeffrey Lacasse, a couple of PhD's who had previously published very nice work about the excesses of the serotonin theory of depression. They first published a letter in JAMA pointing out a problem in interpreting the original study:


http://jama.ama-assn.org/cgi/content/full/300/15/1757-b


In reply Robinson admitted their main point:


http://jama.ama-assn.org/cgi/content/full/300/15/1758-a


That is, while the investigators had shown that the drug treatment was better than placebo, they did not show that it was better than counseling. On review of the data Robinson admitted that the drug and counseling showed no difference. This of course was hardly the pro-drug message that had been put out through the media.


Leo and Lacasse then dug a bit deeper and wondered whether Robinson had received financial support from Lexapro's makers. No such conflict had been revealed in the original JAMA paper, but it took them a few nanoseconds on the internet to discover numerous financial links between Robinson and Forest Labs. Leo notified JAMA of this omission in the stated conflicts.


After time had gone by and no response had been received from JAMA, Leo and Lacasse decided to write a letter to the "Rapid Response" section of BMJ describing the incident and raising their concerns (in notably temperate language that did not accuse JAMA of anything bad and that indeed recognized JAMA for its strong editorial stance in favor of disclosure and avoidance of conflicts). Leo later reported that they had sent JAMA copies of their letter to BMJ and generally had checked with JAMA about the progress of the matter, without initially receiving any reply. The "Rapid Response" appeared in BMJ on March 5:


http://www.bmj.com/cgi/eletters/338/feb05_1/b463#208503


And then the stuff started to hit the fan.


The same week, JAMA published a letter from Robinson admitting that there had been incomplete disclosure of conflicts, and citing "erroneous recollection" as the excuse:


http://jama.ama-assn.org/cgi/content/full/301/10/1023-a


We next heard about the fallout from the BMJ letter in the Wall Street Journal health blog, in a posting by their veteran reporter, David Armstrong:


http://blogs.wsj.com/health/2009/03/13/jama-editor-calls-critic-a-nobody-and-a-nothing/


JAMA told Armstrong that the appearance of the letter from Robinson in that particular issue had been planned for some time previous, and that JAMA felt it a major violation of good editorial practice that Leo and Lacasse had gone to another journal to air their charges instead of letting JAMA complete its internal process. Armstrong reported further that JAMA deputy editor Phil Fontanarosa had called Leo and basically threatened him with permanent banishment from ever having an article published in JAMA. Editor Catherine DeAngelis then called Leo's dean to complain about his unprofessional behavior. Things got really juicy, as far as journalism is concerned, when DeAngelis told Armstrong that he should not waste time on this story because Leo is "a nobody and a nothing."


As this story gained legs in the blogosphere (though the regular media seems to have been pretty quiet about the whole thing, not surprisingly--who really cares about a spat between journal editors and academics who wrote a letter?), JAMA apparently felt mistreated, and so this rather odd editorial appeared on its website:


http://jama.ama-assn.org/misc/jed90012pap_E1_E3.pdf


In the editorial the editors basically deny Armstrong's account of their phone conversations and deny that anyone was threatened. DeAngelis said flat out that she never said that Leo was "a nobody and a nothing"--we are waiting for the next WSJ blog to see what Armstrong thinks about being called a liar. The main point of the editorial is that JAMA has a perfectly good internal system for dealing with allegations of incomplete disclosure of conflicts. Leo and Lacasse, once they notified JAMA, should have sat by quietly and let that process work. Above all else they had no business writing to BMJ, talking to any of the media, or in any way raising their voices. Doing so constituted a serious breach of confidentiality and was therefore something to be complained of to Leo's dean, for example.


The editorial has given rise to rebukes from other present and former journal editors:


http://online.wsj.com/article_email/SB123776823117709555-lMyQjAxMDI5MzI3MzcyNjM4Wj.html


In turn, Leo has put his own account of the timeline of events on line to compare with the allegations in the JAMA editorial:


http://online.wsj.com/public/resources/documents/leo_statement_for_WSJ.htm


Comment: In preparing their editorial response to the WSJ blog post, the editors seem to have consulted the well-known PR firm, When You're in a Hole, Grab a Shovel and Dig Yourself In Deeper. They started out in a pretty bad position. They appeared to be throwing around their weight and power to intimidate a couple of relatively minor academics at some minor institutions. Let's be really clear what it means to call up a professor's dean to complain--you are saying to the world, "I have no reasoned arguments or facts to offer against what this fellow has been saying. Instead I have power and muscle, and I will use that instead of reasoned argument, to show the guy that I am not the sort of person he can mess with." (To his credit the Dean seems to be backing up Leo.) Anyone who says this is really about professional standards and confidentiality is blowing smoke; it's about power.


The main proposal of the editorial, that authors should let JAMA carry out its supposedly infallible investigation process and impose a gag on themselves for however long JAMA chooses to carry it on, has correctly earned the hoots of other editors and academics, who refer instead to the value of academic freedom and freedom of speech.

It is also unfortunate that the editors have undone so much of their former good work. When you take a strong editorial stand against commercial conflicts of interest, you set for yourself a high standard. You proclaim to the world that you anticipate being brought to task if you fall from that standard in the future. If you act later on as if you deserve a free pass for having said all those nice things previously, and so expect that no one will beat you up if you later fall off your pedestal, you show that you never really had committed yourself to that high standard in the first place. No one ever recommended a life of virtue because it's easy.


I can relate to what Leo and Lacasse are going through because of the treatment I received at the hands of JAMA's editors a few years ago when I was doing the research for HOOKED. I became curious about the background to the CLASS study. Readers will recall that that study was notorious for claiming that Celebrex was highly effective in preventing serious gastrointestinal bleeds, when it was later disclosed that the authors had reported only on the first six months of data, and ignored the second six months of the study when the apparent benefit attributable to Celebrex was reversed. On the face of it, that sort of manipulation of the data amounts to presumptive research fraud. I was curious as to why JAMA, on finding out that these data were withheld (admitted by DeAngelis in an interview with the Washington Post), did not demand that the article be retracted.

I e-mailed Dr. DeAngelis for an explanation and, upon receiving no answer that I was aware of, I next tried contacting another member of the editorial board, Dr. Drummond Rennie. That brought a very angry e-mail from Dr. DeAngelis, noting (correctly) that she had indeed replied but that I had missed her e-mail, and that she was now accusing me of violating JAMA policy by going behind her back to quiz other editors and apparently trying to sow dissent among the editorial board. I apologized for having missed her e-mail and explained once again that all I wanted was some information and had no desire to violate board policies. The only explanation regarding CLASS that I received (and that I included in the endnotes in HOOKED) was that all such decisions are private and confidential within the editorial board.

No one, as I recall, told me that I was banned from JAMA. But the tone of Dr. DeAngelis's e-mail gave me the very strong impression that I should not be surprised if none of the papers I might submit to that journal in the future ever found their way into print.

I cannot help but note now that the full story on CLASS has never yet been revealed by JAMA and despite what would appear to be blatant research fraud, the original study has never been retracted--nor has any explanation been given for the non-retraction. (By contrast, JAMA has called for Leo and Lacasse to retract their BMJ letter, without offering any substantive reason why it is flawed.)

Accordingly my advice to JAMA's editors would be as follows:

1. Get off your high horse and stop acting like demigods.
2. Stop pretending as if JAMA's internal editorial policies are somehow sacrosanct and unchallengeable.
3. Set the record straight on CLASS, finally.
4. Stop beating up on Leo and Lacasse. Be big and apologize.