Tuesday, September 1, 2009

Fairness in Enforcing COI Regs by Journals, and Many Other Things Also

Warning: Long post; may be sleep-inducing. Read at bedtime if insomniac.

Laurence J. Hirsch, MD used to manage the Medical Communications Department for clinical research publications at Merck (2001-2006). He now works for a device company. This post concerns a paper he just published in the Mayo Clinic Proceedings:


To give some more background, I have previously blogged about Association of Clinical Researchers and Educators (ACRE), for instance, http://brodyhooked.blogspot.com/2009/07/announcing-fish-in-barrel-club-complete.html. The paper by Hirsch was presented at ACRE's inaugural meeting in July, so you can see it reached print in nearly record time given the glacial pace at which most medical journals operate.

The general gist of Dr. Hirsch's paper is that if medical journal editors are going to demand certain measures in response to financial conflicts of interest (COI) from people associated with the drug industry, they should demand equal obedience from those with other, non-pharmaceutical COIs. The specific grievance he seems to have in mind is the paper by Ross et al. on ghostwriting--which by the merest chance deals with events at Merck that were presumably under Dr. Hirsch's direct supervision when he worked there.

The Ross paper disclosed the fact that the information on ghostwriting related to Vioxx was obtained from the discovery process in lawsuits against Merck for heart disease deaths blamed by victims on taking Vioxx. A couple of co-authors admitted to receiving money from plaintiffs' attorneys for their work on these lawsuits.

Dr. Hirsch states that by the standards commonly applied to pharmaceutical COI, this is very inadequate disclosure. One co-author, David Egilman, has, according to Dr. Hirsch, earned $20-25M in his 20-year career of offering expert testimony on behalf of plaintiffs' attorneys. He allegedly reported that he had testified on matters as diverse as asbestosis, suicide risk with SSRIs, and silicone breast implants. This range of testimony certainly raises the possibility that Dr. Egilman would be considered what is contemptuously referred to in medical circles as a "hired gun."

I said: raises the possibility. The other possibility is that Dr. Egilman does in fact possess a wide enough range of expertise to legitimately be considered an expert witness on diverse drugs and risk factors. When we find out that he holds an MPH degree as well as an MD, and works on health policy, environmental health, and occupational health issues at Brown University, we can see that he might in fact qualify. That being admitted, I still think that $25M is a lot of bread.

Another co-author, Krumholz, had received reportedly some $300,000 for consulting with plaintiffs' attorneys on matters specifically related to Merck and Vioxx.

Now, so far, Dr. Hirsch is in my view on solid ground. I think it matters whether a person who consults with a drug company made $100 or $1M. The higher the number, the more likely it is given human psychology that the person will come to identify the company's (or the attorneys') interests with his own. If I have made $25M in my lifetime claiming that large corporations are evil, then I will certainly have a hard time dispassionately weighing new evidence that might show that they are not that bad after all. Sauce for the goose, sauce for the gander--if such disclosure is expected for Pharma hangers-on, same should be true of financial COI from other sources, including tort lawsuits.

You might notice that Dr. Hirsch's paper runs some 11 pages. You might wonder--the point I conceded above to be a solid point could be expressed very well in 2 pages. So what else does Dr. Hirsch have to say to illuminate the thorny issue of COI?

The fact is that the remainder of the paper, so far as I can discern, is a diffuse mishmash of non sequiturs and vague fingerpointing. I'd bore you to death with a complete rundown so just a couple of examples.

Early on Dr. Hirsch offers what appears to be the mantra of ACRE: the pharmaceutical industry has recently lost respect among the U.S. public due to "[a] small number of highly publicized incidents." In other words, COI is no big deal really; if you could just have weeded out a handful of bad apples, the rest of the industry has a spotless reputation, and academic docs could stuff their pockets with Pharma largesse guilt-free.

Two replies to the ACRE mantra--first, if true, it would undermine the one other part of Dr. Hirsch's article that makes some logical sense, the recommendations at the end where he proposes some Pharma business reforms. His proposed reforms would be completely unnecessary if the trouble was limited to a few bad actors.

Second, as I have detailed at length in HOOKED and later on this blog, the "few bad apples" defense simply rings false when compared with the facts. Rather than repeat all of HOOKED here, I will simply remind readers of one source that I cited--John Braithwaite, an Australian sociologist, who published a book, Corporate Crime in the Pharmaceutical Industry, way back in 1984, well before Vioxx and similar recent scandals. Consider the title of his book, which was a scholarly tome not intended for the mass paperback market. Not corporate unethical behavior or corporate misdeeds but crime. One of the revelations Braithwaite offered was that he had interviewed several corporate VPs at various firms who informally bore the title, "vice president in charge of going to jail." The "post" existed because the company had deliberately configured lines of authority and responsibility in such a way that the buck stopped with that VP, in the event the ongoing criminal activities of the firm should ever come to light and be investigated; and this VP received a compensation package that accommodated the extra personal liability he was placed under as a result. (Footnote: This "post" later disappeared as a result of major court rulings that held that it would be the CEO, not some underling, who would go to jail in such circumstances.)

I tell that anecdote from Braithwaite to make a simple point. Even as far back as 1984, before the really cutthroat era of marketing competition we have seen since then, we did not have an industry that aimed to keep well within the bounds of law and ethics, and where only a few bad apples transgressed. Instead we had an industry, and apparently still do today, where skating as close to the edge of the ethical and legal thin ice as possible is the day-to-day business plan. When a skater falls through the ice we cannot dismiss it as an unfortunate aberration. We must see that this is a predictable outcome of deliberately chosen corporate strategy. (Hirsch insists in a couple of places that the rules on ghostwriting have substantially changed for the better in the last few years, and that we are now seeing scandals left over from 1995 and 2000, which are really old news and no longer pertinent. Maybe so. But how would we know, when the industry continues to handle its internal affairs in such a secretive way, that evidence of bad behavior, if there is any, only pops up when the firm is sued for damages and courts force the discovery of internal memos?)

Just to give one more example of a non sequitur, Hirsch mentions at one point that of about 20 Vioxx lawsuits against Merck, so far, plaintiffs have won only 3. Hirsch asks how this could have been the case if "Merck's conduct was anywhere near as egregious as depicted by ...authors". What does it take to win a tort lawsuit against Merk over Vioxx? One has to show that one's heart attack or stroke was caused specifically by Vioxx and could not have sprung from other causes. Anyone who has passed the most basic course in biostatistics knows that we can have highly certain knowledge that a drug (let's say) increases the risk of dying from a heart attack by 2, 3, or 10-fold, and yet never be able to say in any individual case that a drug was the cause of that patient's heart attack. Indeed if we are to be astounded at all, it is that even 3 plaintiffs won.

So anyway, if you are still awake, we have to explain why a journal such as the Mayo Clinic Proceedings published an 11-page paper that makes a couple of valid points that could have filled a long letter to the editor, and where the remainder is illogic and innuendo for the most part. This is especially odd when we learn from the editor of the Proceedings, Dr. William L. Lanier, that the journal has not solicited any papers specifically on the topic of COI, but has only published such papers when they have passed "the journal's stringent peer-review standards." I suggest that you read Dr. Lanier's editorial that accompanies Dr. Hirsch's paper:


In my humble opinion, you'll see where the journal's "stringent" standards come from. Dr. Lanier's editorial is 6 pages, relatively long for a typical medical journal editorial. It contains much of the same sort of rambling innuendo with a minimum of logical argument as does Dr. Hirsch's original paper, so that at several places I had to scratch my head to determine just what point Dr. Lanier was trying to make.

This post is already way too long, but as my charge against Dr. Lanier's editorial skills is serious, I feel obliged to offer at least one back-up point. Lanier says: "One wonders what the ghosts of Mayo Clinic greats Drs Edward Kendall and Philip Hench [winners of the 1950 Nobel Prize] would think of the current situation in which industrial affiliation in virtually any form places an investigator and his or her published material under suspicion. Working in concert with Merck & Co during World War II, Kendall and Hench's research eventually led to the discovery and clinical introductioon of cortisone."

Yes, indeed, one wonders; and so I for one would like to consult the historical record. Did Kendall and Hench put their names on articles actually written by hired PR staff at Merck? Did they conspire with Merck to suppress the publication of data that might later hurt cortisone sales? If they did none of those things then I would hope we'd say that collaborating with industry meant something different in the 1940s than it does today. If they did some of those things, I'd hope we'd put an asterisk next to our admiration for them (as I suggested in HOOKED we must do with Soma Weiss, another research great of that era).

But, geez Louise, guys, let's face up to the basic point we keep coming back to--if the drugs being discovered and aggressively marketed by Big Pharma today were cortisones rather than Vioxxes, we wouldn't be having this conversation at all.

Addendum 9/2/09: In the post above I tried to give Dr. Hirsch the benefit of the doubt by assuming that when he claimed that Drs. Krumholz and Egilman had received major sums from tort lawyers, at least he had his facts right. I was careful however to hedge my language as I had not had the opportunity to do any independent fact checking.

I have now been contacted by Dr. Krumholz who told me: "just FYI – the claim that I made over $300,000 is not accurate – and was never checked with me – and is not supported by the reference." I was also contacted by another party that wishes anonymity, but who offered some detailed background that suggested that the claims made about Dr. Egilman were also inaccurate.

Now, as I said in the post, I grant the basic conceptual point that Dr. Hirsch made, which is that a COI involving taking money from drug companies should be handled the same as a COI involving taking money from tort lawyers. But that assumes that Dr. Hirsch has his facts straight. If he does not, and he has simply maligned Drs. Krumholz and Egilman without just cause, then of course the sins of which he is guilty go well beyond writing a rambling article. If that should turn out to be the case--again I stress that I have done no independent fact checking--then I add my own apologies for being a vehicle of spreading the misinformation.


Roy M. Poses MD said...

Nice analysis of these very odd articles.

One does wonder why they were published (and so quickly publicized on the Evidence Based Health list, of all places.)

For a more logical discussion of the effects of conflict of interest, and its disclosure, on how one should evaluate clinical research, see this article:
Greenland S. Accounting for uncertainty about investigator bias: disclosure is informative. J Epidemiol Community Health 2009; 63: 593-598. Link here: http://jech.bmj.com/cgi/content/full/63/8/593, requires subscription.

For those without a subscription, I summarized its main points on the Health Care Renewal blog here:

InformaticsMD said...

Additional thoughts from first hand knowledge at Merck are here.

-- SS