Coordinated Ad Hominem Attacks on Physician-Industry Relationships in Guideline Development: The Next Frontier?

Recently, we saw concerted attacks on clinical guidelines committees, but interestingly, not on the science coming out of them.  Instead, the attacks were focused on whether the writers of those guidelines have a financial interest in the area they are working in. 

Two articles looked at conflict of interest (COI) disclosures as they relate to guidelines creating panels.  The first article, published in the British Journal of Medicine (BMJ), did a cross sectional study to examine the extent of financial COIs among members of guideline panels who participated in the development of guidelines on hyperlipidaemia and diabetes between 2000 and 2010. 

The second article, published in Forbes blog, attacks the Journal of Clinical Lipidology and its new Recommendations from an expert panel assembled by the National Lipid Association (NLA), based on an analysis, which found that some members of the panel have collaborated with industry on other work.

Also this week medical malpractice law firms began pining for guideline whistleblowers.

What is problematic about both articles is that they do not account for the tremendous breakthroughs Americans have benefited from over the past several decades in areas such as diabetes, heart disease, hypertension, and related diseases and symptoms.  For example, 

• Since 1970, the death rate from heart disease has dropped nearly 60%,
• Deaths from stroke are down 70% since 1970
• The average life span of Americans increased from 69.7 years in 1960 to approximately 80 years in 2007. 

If it were not for the declines in death rates from heart disease and stroke, we would lose 1 million more Americans every year.  Moreover, for every $1 spent on: 

• Statin therapy for heart attack survivors produced as much as $9.44 in health gains.
• Routine use of beta-blockers for acute heart attack sufferers produced as high as $38.44 in health gains
• Intensive glycemic control in newly diagnosed type 2 diabetes patients produced $3.77 in health gains 

In addition, a recent article from CNN noted that, “the prevalence of heart disease in the United States is declining.”  Specifically, the Centers for Disease Control and Prevention (CDC) found that the prevalence of coronary heart disease decreased from 6.7% to 6.0% a remarkable 10.9% drop from 2006 to 2010. The results, based on a national telephone survey, were published Thursday in the CDC’s Morbidity and Mortality Weekly Report.

“That’s a very significant decline, from 6.7% to 6% in five years,” said Dr. Jerome Cohen, a board member of the National Lipid Association and professor emeritus in preventive cardiology at St. Louis University.  Consequently, the decrease in heart disease prevalence reflects the work of industry on developing statins and the medical profession in developing guidelines, and medical education in ensuring the data and guidelines reach prescribers.

Additionally, a recent study from JAMA found that because of “specialized pacemakers and better use of medications such as ACE inhibitors that relax blood vessels, diuretics that prevent fluid buildup, digoxin that boosts heartbeat strength and beta blockers that ease strain on the heart, hospital stays for heart failure fell a remarkable 30 percent in Medicare patients over a decade.  All of these tools are products created by physician-industry collaboration.

Physician Industry Collaboration

Despite the clear benefits Americans have experienced from these breakthroughs, a small minority of individuals are concerned about physician-industry collaboration.  These critics call the issue a matter of “conflicts of interest,” which has recently surged in the media and government investigations over the past several years. 

However, the relationships that physicians, academia, and government have with industry have been subject to a great amount of misunderstanding, exaggeration, and outright stigmatized for the past few years, in part because of what some believe are widespread unethical practices. These include “frequent and large ‘consultancy’ payments to physicians, the practice of ‘ghost writing’ by drug company employees, and the prevalence of industry funded ‘key opinion leaders’ in medicine,” which have raised concerns that physicians’ financial relations with industry may undermine the practice and promotion of high quality evidence based care.  The reality however, is that such “conflicts” are extremely rare, and have never been shown to harm patients directly.

Consequently, the authors of the study noted that, “one area in which the presence of COI may be particularly concerning is the development of clinical practice guidelines.”

Guidelines serve to standardize care, to inform evidence based practice, and ultimately to protect patients, so their freedom from bias is particularly important. Over the past decade, most organizations that produce guidelines have adopted COI disclosure policies for members of guideline panels. Some organizations, such as the UK National Institute for Health and Clinical Excellence (NICE), have gone further, excluding authors with COI from relevant decision making. US and Canadian guidelines are issued by medical specialty societies, non-profit organizations, government agencies, and professional associations, each with their own guideline development processes and COI disclosure policies.

The authors explicitly recognized that “most organizations mandate some form of disclosure.”  However, they assert that, “complete transparency is often not achieved, and simple disclosure of interests may not be enough to prevent panel members’ bias from influencing recommendations.”  Yet they cite no evidence to support this claim.

Instead, the authors rely on a non substantiated report from the Institute of Medicine (IOM), which called for “the exclusion of panel members with financial COI, the appointment of a chair without any relationships with industry, and an end to direct funding of guidelines by industry. IOM also recommends full disclosure of the COI policy of each guideline panel, along with the potential interest of all panel members. Lastly, based on personal opinion they recommend that if appointing panelists with interests is unavoidable, their presence should be limited to a minority and they should be prohibited from voting.”

Methods and Findings

The authors used the IOM recommendations as a framework to determine the prevalence of financial COI among guideline panelists from organizations considered likely to reflect best clinical practice and influence behavior. Specifically, the authors did a cross sectional study of members of panels that have produced guidelines over the past decade (2000-2010) through national organizations in the United States and Canada that covered screening for and treatment of diabetes and hyperlipidaemia.

Fourteen (14) guidelines met the search criteria, of which five had no accompanying declaration of interest by panel members. The authors searched each guideline for declaration of COI by the panel members. If no conflict was declared in the guideline or in an accompanying document, they searched Medline to identify financial interest, accounting that any relationship with industry whether applicable to the guidelines or not were by the writers to be considered a conflict of interest.  They also searched all publications of each panel member in the two years before publication and the year of publication of the guideline for data on COI.  If the Medline search identified no COI, the authors used Google to search the internet, combining each panelist’s name with the name of each of the major manufacturers that develop and market drugs for hyperlipidaemia and diabetes.

Among the 14 guideline producing organizations, nine provided a document listing the COI of panelists. Five guidelines did not provide an opportunity for panelists to publicly declare financial COI, and four of these were sponsored by the government.  288 panel members had participated in the guideline development process, of which:

• Overall, 150 (52%) panel members had relationships with industry, of which 138 were declared and 12 were undeclared.
• 138 (48%) reported relationships at the time of the publication of the guideline
• 150 (52%) either stated that they had no such conflicts or did not have an opportunity to declare any.
• Among 73 panelists who formally declared no conflicts, 8 (11%) were found to have one or more.
• Of the 14 guidelines, 12 identified a chair. Of these 12, six had a interest, all of which were declared. Among guidelines that provided public declarations of interests, six of seven chairs reported interests
• Panel members from government sponsored guidelines were less likely to have conflicts of interest compared with guidelines sponsored by non-government sources (15/92 (16%) v 135/196 (69%).
• Of the 211 panelists who had an opportunity to publicly declare interest, 73 stated that they had no COI, among whom 11% (n=8) had undeclared interest identified through the authors search strategy
• Among the 77 panel members who did not have an opportunity to publicly declare interest, the authors found 5% (4) to have interest through their search strategy
• Of the 12 undeclared interests, they found one through Google search and the remainder through Medline search.

The authors defined financial COI as the “direct compensation of a guideline panelist by a manufacturer of a drug used to treat the disease of interest in the guideline, in the form of grants (including research), speakers’ fees, honorariums, consultant/adviser/employee relationships, and stock ownership.”  The authors classified COI into three categories: 1) declared in the guideline; 2) undeclared in the guideline but identified through our search strategy; and 3) no opportunity to declare COI in the guideline but identified COI by our search strategy.

Based on the above findings, the authors claimed that COI “were present for the vast majority of guideline panels reviewed.”  However, only 52% had COI’s, hardly a vast majority.  The authors also exaggerate claims about members who did not report.  Only one in nine (11%) reported no COI but were found to have undeclared COI.  However, the authors did not show any evidence or discuss why the authors had not declared.  This could be reasons that are outside the scope of the individual panel member, such as the policies of the guidelines committees or the dates ranges in which they had served the companies.   

Consequently, the authors noted “that in spite of increasing requirements for disclosure of COI by guideline panelists the rate of COI remains similar, suggesting that mandating transparency may not translate into decreased COI on guideline panels.”

The authors suggested that because government sponsored panels “had significantly fewer panel members with COI,” that expert panels can be convened from members predominantly without COI.  However, this claim is completely misguided.  As recent numbers from Food and Drug Administration (FDA) advisory committee panels have demonstrated, it is extremely difficult to find experts who have not collaborated with industry.  If a federal agency such as the FDA does not have the means or resources to find adequate panel members and experts to serve important advisory committee roles, finding such experts for guidelines committees will be even more difficult.

National Lipid Association

The National Lipid Association’s (NLA) mission is to enhance the practice of lipid management in clinical medicine. In support of this objective, the Association convened an expert panel to author a paper entitled, “Clinical Utility of Inflammatory Markers and Advanced Lipoprotein Testing: Advice from an Expert Panel of Lipid Specialists,” for publication in the September/October issue of the Journal of Clinical Lipidology, which was mailed to the membership and published online this week to the Journal website and ScienceDirect.

The paper discusses the consensus view of the expert panel, which evaluated the use of selected biomarkers [C-reactive protein, lipoprotein-associated phospholipase A2, apolipoprotein B, LDL particle concentration, lipoprotein(a), and LDL and HDL subfractions] to improve risk assessment, or to adjust therapy. The panel recommendations are intended to provide practical advice to clinicians who wrestle with the challenges of identifying the patients who are most likely to benefit from advanced lipid testing, to provide the optimum protection from CV risk.

“The goal of this paper was to help clinicians decide which tests may be useful for various types of patients on an initial visit to assess and also which tests are useful for follow-up after the initiation of lifestyle modification and drug therapy.”

However, an article posted on CardioBrief asserted that, “panel members, along with the NLA itself, have a dizzying number of industry relationships.”  Despite this claim, NLA contributors to the paper provided full disclosures and the contributors affirmed that their participation was not influenced by bias or commercial interest.

Rather, NLA firmly asserted that the viewpoints expressed in the paper “represent the clinical experience and opinion of NLA experts and the scientific evidence, which serves as the basis of clinical lipidology.” All members of the organization and others reviewing this document thoroughly reviewed the study and made sure it met the interests of patient care. The NLA also disclosed its educational grants from multiple manufacturers to support the development of this position. All author and industry disclosures are included in the paper, preceding the reference section, beginning on page 359.

Discussion

First ask yourself “so what.” How does this affect patient care?  The answer is simple, if you look at the health statitics, acess to these experts with “interests” and following their guidelines has improved patient care.   It is a sad day when one thinks that guidelines committees will be denied access to experts who are working to develop some of the most important medicines.  While the authors of the BMJ study believe the high prevalence of COI is concerning, they are exaggerating the findings, considering barely over half had any reported COIs.  Moreover, the authors do not point out any evidence of any harm done by such panel members or any harm the guidelines have caused to patient care.  Rather, as noted above, Americans and patients have seen tremendous breakthroughs and improvements in patient care for diabetes and hyperlipidaemia, and related ailments. 

Additionally, the authors’ suggestion that guideline committees should rely on IOM recommendations is equally problematic.  To “reform” the guidelines committee process, as the authors suggest, is unnecessary, and would pose a significant burden to guideline committees, which do not have the resources or staff necessary to accomplish what the authors recommend.  While the authors suggest it’s feasible to have conflict-free panels, they offer no plan or evidence on how this would be accomplished.

As previously noted, FDA is experiencing significant challenges filling positions on their advisory committees.  If Commissioner Hamburg and CDER Director Woodcock both admit the challenges of finding qualified experts without relationships with industry, following IOM recommendations for guidelines committees could lead to significant absence on guidelines committees.  This would lead to less consistency among physicians and practicing clinicians, which could ultimately harm patients.

Moreover, as Canadian Cardiovascular Society president Dr. Blair O’Neill noted, “it’s impossible for all experts not to have such ties because the level of expertise would be diluted.”  He explained that there is a “limited pool of experts in Canada to begin with and the best ones inevitably have their research funded by industry because the government funding just isn’t there.  “The experts in the field . . . have done research funded by industry because government granting agencies’ purse strings have become ever tighter,” O’Neill said.

What these articles and many we have discussed recently are beginning to do is to use an ad hominem attack.  “An ad hominem is a general category of fallacies in which a claim or argument is rejected on the basis of some irrelevant fact about the author of or the person presenting the claim or argument.  Typically, this fallacy involves two steps. 

First, an attack against the character of person making the claim, her circumstances, or her actions is made (or the character, circumstances, or actions of the person reporting the claim).  Second, this attack is taken to be evidence against the claim or argument the person in question is making (or presenting).  This type of “argument” has the following form: 

1. Person A makes claim X.
2. Person B makes an attack on person A.
3. Therefore A’s claim is false.  

The reason why an ad hominem (of any kind) is a fallacy is that the character, circumstances, or actions of a person do not (in most cases) have a bearing on the truth or falsity of the claim being made (or the quality of the argument being made).”  Other than a few extremely rare exceptions of individuals who have worked with industry and have chosen unethical behavior, the overwhelming majorities of physicians who work and collaborate with industry are ethical, and choose this kind of work because it is in the best interest of their patients and it will further progress healthcare and science.

Consequently, the authors in both of the articles used ad hominem attacks.  Rather than arguing the science (if any existed to support their position), the authors of these articles point to what they believe are negative characteristics of physicians who collaborate and work with industry.  They argue that because such physicians work with industry, any work they produce on guidelines committees must be bias or conflicted or not evidence-based.  This form of ad hominem attack is utterly mistaken.

Most likely, authors using ad hominem attacks in these contexts do so because there is no evidence or scientific proof to even suggest that relationships with industry harm patients or produce negative outcomes.  The only “science” that these critics can point to is “social science,” and that “gifts” influence behavior.  But even that study (Wazana) is problematic because it tried to show how interactions with industry hurt clinical care, but of the publications reviewed by the article, none used patient outcome measures. If anything the outcomes including a 10.9% reduction in heart disease in 6 years would point that those guidelines benefited from the use of experts who work with industry.

Ultimately, there is nothing, as of now, to suggest that the guidelines created by any of these organizations are anything but evidence-based.  There is no evidence of undue influence or bias, other than the authors pointing out that panel members have properly disclosed their interests.  Instead, the evidence we do have is that patients and Americans are healthier and living longer lives.  Why should we change that now?