Thursday, May 21, 2015

Say It Ain't So: Logical Fallacies in Defense of Conflicts of Interest ... in the New England Journal of Medicine?

Introduction

We have been viewing with alarm the web of conflicts of interest draped over medicine and health care since we started Health Care Renewal.  We have been particularly concerned about how conflicts of interest may have led to threats to the integrity of clinical research, especially due to manipulation and suppression of clinical research studies.  We have also been concerned about how COIs have led to threats to the integrity of medical education, especially given how health care corporate marketers have paid influential health care professionals and academics to be "key opinion leaders," mainly to act as salespeople in disguise.  We have discussed individual and institutional conflicts of interest involving all sorts of health care organizations.

When we started writing about these issues, we did not find many who shared our concerns, but the topics have become better known.  The Institute of Medicine wrote an apparently authoritative report in 2009 on conflicts of interest which got some notice, but attracted few adherents.  There have been few changes on the policy front in the US regarding conflicts of interest, with the notable exception of the Sunshine Act incorporated into the Affordable Care Act which required increased disclosure of payments made to health professionals and organizations. 

So it was surprising that the New England Journal of Medicine, probably the most influential and important English language medical journal, recently published an editorial by Drazen (1) and three commentaries by Rosenbaum(2-4) about conflicts of interest, all suggesting that concerns about COIs are overblown, and that excess attention to COIs may be inhibiting medical progress.

It was more surprising, given the reach of this journal, that these articles featured a catalog of logical fallacies in support of their arguments.  We have noted that logical fallacies have been a stock in trade of those who actively defend laissez faire policies about conflicts of interest, and other kinds of interactions among health professionals and industry.  However, I would not have believed that the New England Journal of Medicine would go along with this sort of thing.

However, they did, and so we will endeavor to sort out their catalog, noting the most important uses of logical fallacies, in order of the chronological sequence of the publications....

Burden of Proof Fallacy: That All Physician - Industry Collaborations are Beneficial is Assumed, but Contentions that Financial Conflicts of Interest Affecting Physicians Must be Disclosed, Regulated or Banned Require Rigorous Proof  

"The burden of proof is a fallacy in which the burden of proof is placed on the wrong side," per the Nizkor Project definition.

The Assumption that All Physician-Industry Interactions are Good

The Drazen and Rosenbaum articles assert that the burden of proof rests on those who assert that conflicts of interest ought to be disclosed, regulated or restricted.  However, they take the benefits of all physician-industry interactions as given.  For example,

This partnership between an academic researcher and a drug company went on to alleviate substantial human suffering and should be a model for current behavior. Unfortunately, it is not.(1)

Simply put, in no area of medicine are our diagnostics and therapeutics so good that we can call a halt to improvement, and true improvement can come only through collaboration.(1)

the benefits wrought by interactions between physician-scientists and industry are ... clear.(2)

[Physician-industry] interactions [are] characterized by a shared mission to fight disease.(4)

life-saving therapies ... development requires the combined talents of clinicians and industry scientists.... (4)

The series of articles includes multiple assertions that physician-industry collaboration, which is not further defined, is necessary for the advancement of medicine.  The articles never explicitly exclude various kinds of "collaborations" that others may question, including for example, corporate marketers paying well known, often senior academic physicians to be "key opinion leaders" and thus act as salespeople; or paying physicians to give "drug talks" that are clearly marketing exercises, (e.g., the case of "Dr Drug Rep.")

The NEJM articles only supply anecdotal data at best to support this broad assertion.  Of the two anecdotes used by Drazen(1), one was about collaboration between Selman Waksman and Merck during the 1940s in the development of streptomycin.  The applicability of this anecdote, from long ago, done under the pressures of wartime, and long before the era of "shareholder value" theories of management that put short-term revenue ahead of all else (look here), was unclear.   The other "cogent example has been a vaccine against Ebola virus disease."  However, no such vaccine has been licensed for use or accepted as effective, yet.  In fact, society's failure to develop such a vaccine up to now has been attributed to pharmaceutical industry management's emphasis on the preeminence of revenue.  Until the recent epidemic, Ebola vaccine was not seen as a big money maker (look here).

In short, the series of articles accept the value of physician-industry collaboration, writ broadly, in the absence of clear evidence.

The Contention that the Burden of Proof is on Those Who Argue that COIs Should be Disclosed, Regulated or Restricted


On the other hand, regarding assertion that conflicts of interest ought to be disclosed, regulated, or restricted, Rosenbaum wrote

we still lack an empirical basis to guide effective conflict management.(3)

Equally unclear are the benefits and harms of regulations aimed at exposing or mitigating these conflicts.(3)

It remains unclear whether ... disclosures actually mitigate the risk of bias.(3)

conflict-of-interest policies have evolved not through careful data gathering and analysis.... (4)

In particular, most of Rosenbaum's three articles(2-4) focus on her general doubts about and perceptions of faults in the evidence-base about the harms of conflicts of interest, or the benefits of disclosing, regulating or restricting them. For example,

though considerable social science research suggests that even small gifts may influence physicians, it doesn't necessarily follow that greater financial stakes are more influential.(3)

Suggestive data may be worse than no data at all.(3)

It depends on how you define harm.  Consider pharmaceutical 'gifting,' a practice that smacks of bribery - which may be sufficient reason to prohibit it.  But does it actually harm patients?(4)

Furthermore, while decrying the lack of rigorous data in support of disclosing, regulating or restricting COIs, she raises doubts about such actions based on vague anecdotes and general, but unsubstantiated assertions, including

some of the young, talented physician-investigators I spoke with expressed worry about how any industry relationship would affect their careers.(3)

The proportion of physician-investigators who have such concerns was not stated.

A medical school dean probably won't lose her job if patents aren't produced under her tenure, but she will be taken to task if she appears to lax in regulating faculty-industry interactions.(4)

No further specifics about consequences to such academic leaders appeared. 

For many people, however, the medical-industrial complex elicits deeply negative feelings that make it tough to evaluate fairly any intervention aiming to mitigate industry influence.(4)

The evidence in support of this assertion was not apparent.

I think the desire for retribution against 'bad pharma' informs our management of industry interactions in a way that obscures the possibility that we are obstructing medical advances.(4)

The evidence in support of this thought was not apparent. 

Thus Drazen and Rosenbaum clearly believe that the burden of proof is entirely on those who advocate disclosing, regulating or restricting conflicts of interest.  Yet they never argue this point explicitly.  In my humble opinion, I see no reason that their beliefs should be considered a fundamental law of nature, while the beliefs of those who differ with them should be considered unproven hypotheses. The NEJM series of articles seem to be an extended exercise in the burden of proof fallacy.

Appeal to Authority: Important People and Organizations Agree with Us

The appeal to authority fallacy is that an argument supported by an authority must be true, as per Nizkor.  

Drazen and Rosenbaum corroborate their opinions with those of various authorities, but fail to identify any authorities who disagree with them.  In fact, as noted below, they often cite opinions with which they differ without noting who advanced them.  So, for example, 

The National Center for Advancing Translational Sciences of the National Institutes of Health, the President’s Council of Advisors on Science and Technology, the World Economic Forum, the Gates Foundation, the Wellcome Trust, and the Food and Drug Administration are but a few of the institutions encouraging greater interaction between academics and industry, to provide tangible value for patients.(1)

Shaywitz and Stossel, who have each written on the benefits of academic-industry collaboration and the challenges of bringing new products to market, are rare voices competing with a loud chorus of shaming.(3) 

Richard Epstein, a University of Chicago law professor who writes convincingly about the dangers of overregulating medical conflicts, questions certain limitations on the ties of FDA advisory-panel measures.(3)

Note that the authors of the NEJM articles do not discuss whether these authorities could have their own biases.  For example,while Drazen cited the support of the Gates Foundation above, Rosenbaum later acknowledged the current CEO of the Gates Foundation is a former Vice President of Genentech(4).  Neither noted that Dr Desmond-Hellmann was on record early as an apologist for the huge increases in drug prices that occurred starting in the first decade of this century (look here).  Dr Stossel has been known to deploy his own logical fallacies to defend physician-industry interactions (look here), as has Professor Epstein (look here).  Dr Stossel has been known not to disclose his own relationships with industry (look here).

Furthermore, while Rosenbaum attributed a stance in favor of disclosing, regulating or restricting COIs  to former NEJM editor Dr Arnold Relman, it was in the context of doubting his approach, rather than supporting his authority.(3)  Most of the views she cited as opposing hers were not attributed.


Ad Hominem Fallacy: People who Advocate Increased Disclosure, Regulation, or Restriction of COIs are "Pharmascolds"

The ad hominem fallacy is that a posited defect in the character, abilities, competence etc of a person making an argument means the argument is false, see Nizkor.  Rosenbaum wrote,


Physicians know that 'pharmascolds,' as physician-scientist David Saywitz and Tom Stossel have dubbed them, will 'vilify the medical products industry and portray academics working with them as traitors and sellouts.'(3)

The wording thus gives credence to the idea that anyone who advocates for disclosure, regulation or restriction of conflicts of interest is such a "pharmascold."  The articles by Rosenbaum never seeks to balance that assertion with any epithets that might be applied to people who advocate for unrestricted physician-industry interaction.  The implication is that "pharmascolds" are at best excessively sensitive, or worse, engaged in witch hunts. Thus this appears to be at least a back-handed use of the ad hominem fallacy.

Appeal to Pity Fallacy: People Who Advocate a Lenient Approach to Conflicts of Interest are Besieged by a Monolithic Force of "Pharmascolds"

The appeal to pity is an attempt to make an argument more convincing by making the person making it worthy of pity, see Nizkor.  

Rosenbaum started her second article(3) thus,

In 1980, the Journal’s editor Arnold Relman wrote an editorial entitled, 'The New Medical-Industrial Complex' Although it’s hard to pinpoint the moment when a culture forever changed, the editorial represented a seminal event.

She further stated, "In the ensuing decades, endless attention has been paid."  Her examples of this endless attention were two books, the report by the Institute of Medicine, "new rules," and the "recent passage of the Physician Payment Sunshine Act."

Rosenbaum opened her third article(4) thus,

Although I probably couldn’t have explained its rationale, I never questioned the anti-pharma animus that pervaded my medical education. The message I received from certain outspoken classmates and fellow trainees was that interacting with pharmaceutical reps was simply wrong.

She noted that

I suspect my experience was not unique. Indeed, the American Medical School Student Association (AMSA) now grades medical schools on their creation of a 'pharma-free' environment, issuing annual report cards on conflict-of-interest policies and curricula.

As mentioned above, she cited with dismay her interpretation of a single medical student's opinion that a biased lecture caused "violation."  She later cited a Wall Street Journal article and a British Medical Journal article which she thought were too critical of industry.

Near the end of the article was this personal anecdote,

Recently, for the first time, I was asked to consult for a medical products company. My first thought was, 'This would be fascinating.' My second was, 'There’s no way.' I would have to disclose the relationship, my credibility would suffer, and I would be defenseless. That I immediately succumbed to this fear reflects our failure to manage industry relationships effectively.

So the evidence for a huge, powerful, monolithic movement of "pharmascolds" presented was minimal.  Rosenbaum cited a 1980 article and asserted it changed the world, without any real documentation of that.  Otherwise, she cited a few books, a society of medical students, and some personal anecdotes about medical students.  The most telling anecdote was about the author's person perception that her credibility would suffer - presumably unfairly in her eyes - were she to consult on a "fascinating" project, never mind what she would have been paid to do that.  So at the very end, this ostensibly scholarly article concludes with an apparent appeal to pity its poor author for having to give up this wonderful opportunity.  That seems like the essence of an appeal to pity fallacy.

Furthermore, while the evidence of a powerful army of pharmascolds was lacking, the author did not address the evidence that the majority of academic physicians have conflicts of interest, as do the majority of department chairs(5,6).  While she speculated how a medical school dean might be oppressed by the pharmascolds, she did not address how many medical school deans, leaders of academic medical centers, and other top leaders of academic medicine have conflicts (look here).  Finally, she neglected to mention that conflicts of interest mainly come out of corporate marketing and public relations budgets that total billions in US dollars yearly nationally.

So the image of the poor pitiful defenders of the laissez faire approach to industry relationships seems a bit overdrawn.  


Straw Man Fallacies Industry Critics Claim to be Free of Bias, Equate COIs with Rape and Child Abuse, Use Flawed Reasoning, Believe All Physician-Industry Interactions Constitute Fraud

Per Nizkor, "the Straw Man fallacy is committed when a person simply ignores a person's actual position and substitutes a distorted, exaggerated or misrepresented version of that position."  Rosenbaum attributes to all or most supporters of disclosing, regulating, or restricting conflicts of interest all sorts of statements or beliefs without evidence that anyone, or more than a few people actually hold such beliefs, viz...


But couldn't industry critics blind spots leave them unjustifiably confident that despite their industry aversion, they are bias free?(3)

There was no documentation that industry critics claim they are free of all biases.

The application of language associated with rape and child abuse to the circumstances of education about effective drugs reveals a feature of the conflict-of-interest movement that has fed its contagion and rendered it virtually unassailable....(4)

Note that this was based on a single Harvard Medical student saying a single lecture lead him or her to feel "violated."  There was no documentation that anyone actually made a comparison to rape or child abuse, much less that such ideas are widely held.

Such flawed syllogistic reasoning has become the norm.(4)

Note that this refers to a "narrative" that someone who works with industry must have a favorable view of industry and therefore must make decisions based not on "clinical and research expertise but a desire for financial gain."  That in turn was derived from a single article in the news media.

'If post-Hart political journalism has a motto,' writes [journalist Matt] Bai, 'it would be: 'we know you're a fraud somehow.  Our job is to prove it.'  A similar motto could apply to much reporting on physician-industry interactions.'(4)

Furthermore,

the climate is so permeated with assumptions of fraudulence that treatments ... that have revolutionized our ability to prevent and treat disease become pawns in the hunt for wrongdoing.(4)

The few examples Rosenbaum supplied of supposedly faulty journalism did not seem to discuss fraud at all.

Summary

The series of articles about conflicts of interest that just appeared in the New England Journal, while ostensibly scholarly, published by the journal's "national correspondent" in the Medicine and Society section, appear to be polemical.  They deployed a substantial number of logical fallacies to make the point that medicine and society have gotten too tough on conflicts of interest.  They are notably short on logical, dispassionate discussion of the evidence.  Thus, they seem more like posts on a very opinionated blog site rather than commentaries in a scholarly medical journal.

By publishing this series of high visibility articles, the New England Journal of Medicine seems to have deliberately muddied the waters of discussion about conflicts of interest.  This is sad, because the journal was once considered the foremost English language scholarly medical journal, but it now seems to be publishing polemics.

This latest publishing phenomenon, or debacle, should be a reminder why conflicts of interest, if unhindered, become so prevalent.  They are relationships that benefit both parties involved.  For example, a pharmaceutical company marketing department presumably benefits from the increased revenue generated by increased sales generated by prominent key opinion leaders touting its products in the guise of professional and/or academic experts.  The KOLs, on the other hand, benefit from their generous payment.  Who loses?  - physicians who are increasingly regarded as pharma shills; physicians, whose decision making on behalf of patients may be hindered by constant exposure to marketing and public relations drowning out logical, evidence based discussion;  patients, who need to worry whether the tests and treatments they get were ultimately too influenced by conflict of interest fueled marketing and public relations, and not enough by evidence and logic.

As we said many times before, the web of conflicts of interest that is pervasive in medicine and health care is now threatening to strangle medicine and health care.  For patients and the public to trust health care professionals and health care organizations, they need to know that these individuals and organizations are putting patients' and the public's health ahead of private gain. Health care professionals who care for patients, those who teach about medicine and health care, clinical researchers, and those who make medical and health care policy should do so free from conflicts of interest that might inhibit their abilities to put patients and the public's health first. 

ADDENDUM (21 May, 2015) - See also detailed comments by Larry Husten on Forbes.and by Dr Susan Molchan on the HealthNewsReview blog.  Both delve into the details of some of the cases and data that Dr Rosenbaum does cite, and thus raise questions about the facts she chose to use, and how she chose to interpret them. Also, corrected citation for "pharmascolds."

ADDENDUM (26 May, 2015) - See additional posts here, here, here and here by Micky on the 1BoringOldMan blog.

ADDENDUM (29 May, 2015) - See posts in the Lown Institute blog by Shannon Brownlee, Dr Vinay Prasad, and Dr Vikas Saini

ADDENDUM (8 June, 2015 - See also comments by Dr Steven Reidbord in the KevinMD blog


References
1.Drazen JM.  Revisiting the commercial-academic interface.  N Eng J Med 2015; ; 372:1853-1854. Link here.
2. Rosenbaum L.  Reconnecting the dots - reinterpreting industry-physician relations.  N Eng J Med 2015; 372:1860-1864.  Link here.
3. Rosenbaum L. Understanding bias - the case for careful study.  N Engl J Med 2015;  372:1959-1963.  Link here.
4.  Rosenbaum L.  Beyond moral outrage - weighing the trade-offs of COI regulation. N Engl J Med 2015; 372: 2064-2068.  Link here.
5.  Campbell EG, Gruen RL, Mountford J et al. A national survey of physician–industry relationships. N Engl J Med 2007; 356:1742-1750. Llink here.
6.  Campbell EG, Weissman JS, Ehringhaus S et al.  Institutional academic-industry relationships.  JAMA 2007;298(15):1779-1786. doi:10.1001/jama.298.15.1779.  Link here.

No comments:

Post a Comment