Thursday, June 26, 2008

Conflicts, Bad Science, and Corlux

Recently, the watchful eyes of Charles Grassley have been peering into the bank accounts of big name psychiatrists. Melissa DelBello and Joe Biederman (1, 2) from the Wonderful World of Child Bipolar were first, and now Alan Schatzberg has been hit. Schatzberg is the Chair of Psychiatry at Stanford University. He is also the President of the American Psychiatric Association. In other words, he's kind of a big deal.

Pharmalot hits the details, but the gist is that Schatzberg is deeply involved at Corcept Therapeutics, a company for which he is chair of the scientific advisory board and holds a large amount of stock. According to Grassley, he did not disclose some of his stock sale profits or the magnitude of his multimillion dollar stock holdings in the company. Additionally, Schatzberg allegedly underreported income received from other drug companies. It appears that Schatzberg was not really required to disclose some of this information, so according to my brief review of the information, it is quite possible that he has broken no rules. Now, whether the rules need to be changed is a different story. No offense to Grassley, but I was well ahead of him on part of this story, noting in April 2007 that Schatzberg had a mega-conflict of interest going with Corcept. I also noted previously that Schatzberg was on the Zyprexa bandwagon, helping to "educate" fellow physicians about the Lilly wonder drug.

The Real Problem: But amidst all this discussion of conflicts of interest, I am afraid that we are getting a bit diverted from the main problem, that of shoddy science. It is admittedly interesting noting that Schatzberg is somehow supposed to be an independent, disinterested scientist while standing to make an absolute truckload of money if his sponsored product succeeds. But it runs deeper. While Schatzberg is a bigwig at Corcept, let's review how Corcept's main product mifepristone (RU-486; yes, the abortion pill) has done.

Mifepristone (aka Corlux) is intended to work as a treatment for psychotic depression. One main problem: It doesn't relieve depressive symptoms. In multiple trials, it has failed to demonstrate antidepressant properties. The CEO of Corcept and another member of their scientific advisory board have previously tried to spin away such inconvenient data by painting negative results as positive. To give Corcept credit, their scientists are consistent spinmeisters, seemingly always able to dredge a positive from obviously negative findings. Schatzberg has been an author on a couple Corlux-related papers that were shredded by independent analysts, who found statistical problems and overly optimistic interpretations of the study results. As the senior member of the Scientific Advisory Board, I assume that Schatzberg had some input on the other study reports that also overstated the efficacy of Corlux.

Could his millions of dollars in Corcept holdings bias Schatzberg, either subconsciously or overtly? You be the judge. But remember that this is not just about conflicts of interest -- this is about science. There is hard evidence that the research on Corlux, which is tightly linked to Schatzberg, has been misinterpreted for the sake of marketing. Conflicts of interest sometimes lead to bad science, but rather than focus just on conflicts of interest, we need to dig a layer deeper and see the poor science -- the shoddy evidence that is used as the foundation for "evidence based medicine" in many cases.

Note also that David Healy has written an interesting piece on the topic of conflicts of interest and bad science, pointing out that a larger problem is lack of access to company-owned data. Think Paxil and suicide. He concludes:
If I were employed in a company marketing department I would much prefer to have the field think that all that is wrong is that a few corrupt academics fail to declare competing interests than to have the field think that company practices that restrict access to data while still claiming the moral high ground of science are the real source of the problem.
I'd love to know what American Psychiatric Association members think about this. The news had already broken about Schatzberg overstating the efficacy of Corlux before he was elected APA president. Do APA members not care that their president has a documented record of putting product promotion before scientific evidence?

11 comments:

Daniel Carlat said...

Great post--I just posted about the 6 million dollar man as well: http://carlatpsychiatry.blogspot.com/2008/06/six-million-dollar-man-witch-hunt-or.html

Radagast said...

"...I'd love to know what American Psychiatric Association members think about this..."

Regrettably, you'll probably never find out. And I imagine the reason you'll never find out is that those members are probably doing something similar, in their way. Corruption: it IS the system.

Matt

Roy M. Poses MD said...

We discussed some of theses issues, and did a brief review of the 2002 Biological Psychiatry paper that Schatzberg authored, in this post back in July, 2006 on Health Care Renewal:
http://hcrenewal.blogspot.com/2006/07/conflicts-of-interest-at-stanford.html
My conclusion then was that the paper provided extremely weak evidence, to be charitable, about the efficacy of mifespristone

Anonymous said...

Very good post with some fair assessments on Corlux. This blog post won an honorable mention on the Redscrubs.com weekly wrap up.

Keep up the good work!

Anonymous said...

I've had some private conversations with Aubrey Blumsohn on this point, and I tend to disagree that the issue is purely bad science. Here's a copy of some of the reasons for my disagreement:

"Science is not an artificial process composed of methods performed automatically, by rote, unthinking machines. Science is a praxis, an irreducibly human endeavor, and human foibles, biases, and predispositions affect the process at all levels and all times, starting with the questions we think are important enough to try to answer, down through the methods, and through the interpretation of raw data. (This is bedrock philosophy of science -- Kuhn, Feyerabend, Hacking, etc.)

COIs in and of themselves obviously do not mean the scientific data in question is ipso facto invalid. I don't think anyone seriously believes that. The problem, as Howard Brody and many others have pointed out, is that the presence of extensive COIs creates a much higher likelihood that in the long run of iterations, the biases and predispositions -- many of which affect behavior on the subconscious level -- generated by the relevant enmeshment will affect the science. And indeed, the overwhelming body of evidence suggests this is exactly what has happened.

Science, like any other human practice, is unavoidably interpretive. We are interpretive creatures. This cannot be eliminated, nor should it be, IMO. The problem is that extensive COIs absolutely color and shape the features of those interpretations. Of course, in any given case, irreproachable scientific methods, procedures, and inferences are certainly possible, regardless of COIs. But over the long run of cases, such behavior is far less likely than it would be in the absence of the obscene sums of money exchanging hands.

As Howard Brody also points out, it's not simply an issue of trust. It is also an issue of trustworthiness. And Brody is right on, in my view, that an investigator who knows the likelihood that these COIs influence behavior, and continues to insist that they will not be affected because their methods are so pure, is less than completely trustworthy."

CL Psych said...

I've yet to see anyone who is not financially tied to Corlux say anything positive about it. Yet trials of Corlux continue to recruit patients.

Roy -- excellent work on your prior post on the topic.

Danny -- thanks for your post on Schatzberg's latest. I agree that he really should choose to either be a tycoon or an "independent" academic, but he can't have it both ways.

Brandon -- thanks for noticing my post; I'm honored by your nomination.

Matt -- No doubt there is much corruption throughout the system, as has been noted here and in many other outlets. I hope that many American Psychiatric Association members are doing a bit of soul searching over their election of Dr. Schatzberg. He may be a nice guy for all I know. But the sponsorship of bad science -- that's not becoming of a leader in medicine or any other field.

Anonymous said...

it just occurred to me: the scary name of ru486 always catches me more than mifep, mifes, whatever; ru486 the abortifacient sounds like: are you for 86ing, or ending, this fetus?"

Some don't know this but '86' means 'to end '(wikipedia notes this, as well as being common knowledge).

So that got me thinking: if ru486 gets approved for depression, then a physician could 'off-label' prescribe this abortifacient for a pregnant woman for the purpose of morning-after abortion, but document 'depression' as the reason for visit and the reason for rx ru486. Since all communication is confidential, the doc would have a generally easy time in court if a prosecutor tried to demand that the doc tell whether in session the issue of the abortifacient property was discussed at all. Wait: it would necessarily have to be discussed because if ru486 is considered for a woman as tx for depression, the doc would HAVE to note any potential teratogenic effects, as done with some antimanic meds.

Genius.

And generally a pregnant woman considering abortion has some handful of symptoms: distress, worry, depressed mood, crying, disrupted sleep, etc. to be documeted as supporting the dx of 'depression.'

Genius: A work-around to address legal access for abortion, and/or barriers to abortion associated with clinic access (i.e., med record would exist, third-party payor issues, distance to facility, stigma, protesters outside the abortion clinic, etc.).

Genius. I am almost convinced enough to buy stock. If only it worked for depression.

Anonymous said...

this is my first time to your site, as I found it through Furious Seasons, who I like his reporting.

This matter is damning for my field, as I am a psychiatrist, but not a member of the APA. I left it in 1995 when we had the election between Eist and Sharfstein, and even though Dr Eist won, that 48% voted for Dr Sharfstien's unstated premise that "managed care is here to stay, deal with it" spoke volumes to me how clueless too many colleagues had become.

Managed care was, still is, and will continue to be as much if not more THE PROBLEM than what big pharma has done of late. Dr S is representative of what is wrong with psychiatry. As long as too many colleagues say nothing, their silence is deafening to the public.

This is wrong. To say anything less is equally wrong!

Radagast said...

CL Psych wrote:
"...But the sponsorship of bad science -- that's not becoming of a leader in medicine or any other field."

Quite. Do you know? I'm extremely skeptical about a branch of "science" that retains unto itself to determine when a person is "defective" and also when they are not defective, as psychiatry does.

Call me paranoid, if you like, but if one has a vested interest in diagnosing people as mentally ill, then won't one be tempted to do that, irrespective of whether they are, or not? I don't say that that happens routinely, you understand, but given the way the system is currently set up, that could happen. I don't think "professionals" are any more trustworthy than anybody else, frankly; certainly not where there own interests are concerned, and I don't care how they dress up their Great Wisdom.

Matt

Anonymous said...

If you would review the literature, you might find that there actually were some positive results with RU-486 by someone other than Schatzberg.

Crwno.

CL Psych said...

Perhaps you could enlighten me about the positive results on Corlux. I'd be very thankful if you could point me to the study or studies of which you speak.

As you've perhaps read in my prior posts, much of the Corlux research is not positive, but I'm excited to learn about these positive results that you claim I have somehow missed.