Flaws in the development of guidelines that are related to conflicts of interest, bias among the members of the guidelines panel, and suppression of competing viewpoints were central to the antitrust investigation of IDSA by Attorney General Blumenthal. Moreover, these flaws mirror the growing list of problems with evidence-based guidelines in general, which has led many to call for reform in how guidelines are developed . In their article in the Journal of the American Medical Association, Sniderman and Furberg highlighted critical deficiencies in the current process:
Conflicts of interest of 'experts' on guidelines panels may substantively drive the content of the guidelines in a manner that does not hold the interest of the patient paramount.
A paucity of high-level evidence and an overreliance on 'expert opinion' may result in the formulation of guidelines that merely replace one 'authority-based' system with another.
Legitimate controversy may be suppressed by artificially 'unanimous' panel recommendations and by the exclusion of divergent viewpoints from the panel.
Specialty medical societies, which use guidelines to expand their competitive sphere of influence, may publish guidelines in their own journals essentially 'as-is,' without submitting them to the type of independent, external peer review that might vet issues of bias or conflicts of interest on the part of guidelines panel members .
All of these aspects undermine the credibility of guidelines and permit personal bias to determine the care of patients -- the very problem that evidence-based guidelines are intended to avoid. Indeed, these deficiencies may undermine the ethical foundation of medicine, which requires 'physicians to put the needs of patients ahead of personal gain, to deal with patients honestly, competently, and compassionately, and to avoid conflicts of interest that could undermine public trust in the altruism of medicine' .
The integrity of medicine, and, ultimately, its value to society depends upon recognizing and safeguarding this ethical foundation. Sniderman and Furberg highlight the most important components of that ethics foundation in the context of guidelines: conflicts of interest, overreliance on expert opinion, and failure to acknowledge legitimate controversy. Distilled down to their essence, these issues are based on two ethical principles: the need to hold the interests of the patient above other commercial interests and the need to preserve the 'treatment choice' that gives life to the ethical principle of autonomy.
The ethical obligation to hold patient's interests paramount can be compromised by conflicts of interest resulting from financial ties of the panel members. The ethical obligation to respect patient autonomy depends upon the preservation of treatment choice for patients and can be jeopardized when guidelines fail to acknowledge legitimate controversy and do not provide treatment options. Choice is also jeopardized when there is a paucity of evidence and evidence gaps are filled by the 'expert opinion' of the panel. The problems that arise from conflicts of interest and failure to preserve treatment choice are furthered when the medical specialty societies that sponsor and publish these guidelines fail to adequately police these risks or have industry ties themselves. The issues regarding guidelines development and conflicts of interest, quality of evidence, and patient choice are discussed in more detail below.
Conflicts of Interest
One of the paramount deficiencies identified by Sniderman and Furberg in guidelines development arises from conflicts of interest. The Institute of Medicine (IOM) defines a conflict of interest as 'a set of circumstances that creates a risk that professional judgment or actions regarding a primary interest will be unduly influenced by a secondary interest' . This determination requires identifying a primary interest and a secondary interest. The primary goal of medicine that should stand at the center of treatment guidelines is to '[improve] health by providing beneficial care to patients' . Secondary interests 'may include not only financial gain but also the desire for professional advancement [and] recognition for personal achievement' . This emphasis on primary versus secondary interests arises from the divergent goals of medicine (improving patient health outcomes) and commerce (ensuring a financial return to shareholders) .
Not all conflicts are of equal severity. The IOM suggests that conflicts should be assessed by considering both the likelihood of adversely affecting the primary interest and the seriousness of the harm caused by the conflict:
Likelihood depends on the value of the secondary interest, the scope of the relationship between the professionals and the commercial interests, and the extent of discretion that the professionals have. Seriousness depends on the value of the primary interest, the scope of the consequences that affect it, and the extent of accountability of the professionals .
Because they may affect multiple patient care decisions and criteria for insurance coverage, clinical practice guidelines are considered by the IOM to have a serious potential for harm . This is compounded by the fact that guidelines panels have not usually been held to be legally accountable to patients under malpractice laws because the duty of care that creates liability exposure only arises in the context of a direct physician-to-patient relationship . Discretion of those on guidelines panels increases when the scientific evidence base is weak and the panel elects to plug evidence gaps with 'expert opinion' rather than acknowledging the uncertainty and providing treatment options. These factors increase the risk and seriousness of harm to patients and emphasize the importance of managing conflicts of interest. Parenthetically, the growing importance of guidelines in public policy may increase the application of negligence and strict liability law to their development process .
Unfortunately, conflicts of interest on guidelines panels are common. It is not unusual for 'expert' panels to include 'key opinion leaders', usually academic researchers for whom industry ties are vital to ensure the research funding on which their careers depend . Choudhry and colleagues found that 87 percent of participating 'experts' had financial ties to pharmaceutical companies, and 59 percent had ties to companies whose products were considered in the guidelines authored by the 'experts' . The roll call of guidelines formulated by 'conflicted panels' that appear to further the interests of pharmaceutical companies includes those for the treatment of sepsis, anemia among kidney patients, and high cholesterol . A New York Times article commented that a conflicted guidelines panel that adopted an industry-friendly new definition of high blood pressure 'illustrate[d] connections -- among the pharmaceutical industry, academic physicians and societies that formulate opinion -- that can ultimately affect patient treatment,' and called this 'the monetarization of medicine' .
A central value of research is the pursuit of objective, unbiased information, and many 'key opinion leaders' believe that they can rise above commercial conflicts of interest . Common sense and recent investigations indicate the opposite . When money enters the equation, the nature of the debate simply changes: for example, Choudhry and colleagues found that research articles have either been published or discretely shelved based on ties to industry .
Choudhry and colleagues argued that the subtle influence of money on 'key opinion leaders' forms the basis of a substantial portion of pharmaceutical marketing to promote the industry's interests . Nonetheless, they noted that the vast majority of researchers believe that at least their own research is not compromised by monetary ties with industry . The Association of American Medical Colleges reported that, in contrast to deliberate corruption, many ethical problems in the professions are based on unconscious bias that permeates professional conduct; that is, professionals with conflicts of interest may selectively seek out information, may interpret information in a biased fashion, and may easily be derailed from good intentions, but are typically unaware of these effects .
When researchers dominate the panels that write guidelines, they may bring their research biases with them. Ernst and Canter describe the adverse effects that bias may have on research:
Bias is ubiquitous, and medical research is no exception. From the very outset, investigator bias can influence the general attitude towards a research project....The biased researcher, however, has preconceived ideas and is likely to approach a project to prove a point. For example, a researcher who is convinced of a particular treatment or, worse, has a vested interest in it, could misuse science to demonstrate the efficacy of his therapy. Equally, an investigator with a preconceived negative attitude towards a particular intervention can set out to disprove its efficacy .
While some bias is unavoidable, biases introduced by conflicts of interest frequently are and should be avoided to the extent possible.
Academic researchers who sit on panels may have conflicts of interest not only in the form of industry ties (particularly if they are regarded as 'key opinion leaders'), but also other competing financial interests identified by the IOM, like career advancement or recognition of personal achievements . For example, citing their own research in guidelines may promote both future research funding and further their academic careers . A report from the Institute for Clinical Research and Health Policy Studies at Tufts University describes researchers who have staked their careers on a particular theory and concludes that they 'are likely to resist accepting that the whole field in which they have spent their lives is a null field'. These factors may impair the ability of guidelines panel members to critically select and evaluate research findings in an unbiased fashion.
In addition, serving on treatment guidelines panels is part and parcel of being a 'key opinion leader'. In his pharmaceutical marketing textbook, Ronald Evens, a professor at University of Florida, College of Pharmacy, states:
[I]t is paramount to identify national and regional influencers, that can help guide registrational trial program development, conduct solid and timely clinical research, generate leading disease state publications, help draft treatment guidelines, and begin cultivating potential speakers to help leverage clinical trial product use immediately post drug approval .
Conflict of interest policies typically screen for conflicts within the last year or two years, and 'key opinion leader' relationships with industry may span a career . For example, one researcher on the IDSA Lyme guidelines panel, Allen Steere, has relationships spanning more than a decade with Lyme vaccine manufacturers [10, 36, 37]. Although researchers often develop a known expertise and significant influence in a research area, these 'key opinion leaders' may change their affiliations with pharmaceutical companies over a given time frame. Thus it may be unrealistic to assume that conflicts of interest among 'key opinion leaders' are adequately reflected by a time-delimited inquiry.
In addition to the conflicts of interest of individual panel members, the medical society that sponsors and publishes the guidelines may have conflicts of interest . Industry funding of medical societies is extensive . A recent IOM report reviewed the budget of the American Association of Family Physicians for the fiscal year 2006-2007. Less than 38% of its $80 million budget came from membership dues and services, while 42% came from the pharmaceutical industry, (60% from advertising in the academy's journal and 13% from exhibit fees) .
Although medical societies profess to operate for the public good, Noble and colleagues observed that '[g]uidelines promulgated by a particular medical society generally reflect the specific concerns of its members' . The behavior of societies may not only further their own economic interests, but may also take advantage of opportunities to 'sit in judgment of their competitors' . Treatment guidelines expand the society's sphere of influence, and, according to the IOM, medical specialty societies regard guidelines as one of the most valued services they provide . These factors suggest that the interests of medical societies (and those of the individuals serving on a guidelines panel) may not align with the best interests of patients, their professional competitors, or the interests of the broader public .
Moreover, as Sniderman and Furberg noted, 'Few associations submit the final products of the guideline process for external review before they are accepted and, therefore, in a limited but real sense, the committee, which is a creation of the organization, becomes the final arbiter of its process,' bypassing 'one of the core processes of science,' the type of review that might hold these self-interests in check .
How should conflicts of interest issues be addressed in the guideline development process? In testimony before the Institute of Medicine, Merrill Goozner, Director of the Integrity in Science Project at the Center for Science in the Public Interest, emphasized the need to maintain strong boundaries between those who determine how medicine gets practiced and researchers who conduct industry-funded research . The IOM agreed:
Given the important role that clinical practice guidelines play in many aspects of health care, it is important that these guidelines be free of industry influence and be viewed by clinicians, policy makers, patients, and others as objective and trustworthy .
But what if conflicts of interest are the price of expertise, as is often argued? The IOM provided further guidance on this point.
If groups conclude that participants with conflicts of interest are essential to provide the necessary expertise, they should demonstrate to the public that they have made a good faith but unsuccessful effort to find individuals with the required expertise and without conflicts of interests. They should preclude individuals with conflicts of interest from chairing guideline development panels, restrict the number of individuals with conflicts of interest on panels to a distinct minority (e.g., to 25-30 percent of the membership), and prohibit members with conflicts of interest from drafting and deciding specific recommendations .
Hence, 'key opinion leaders' with industry-related conflicts of interests ideally should be excluded from guidelines panels. Where this is not possible, the influence of these individuals should be regulated by limiting participation to no more than 25-30% of the panel and otherwise restricting the leadership role of those with conflicts of interest.
Quality of Evidence
The implicit goal of evidence-based medicine is the provision of medical care that is grounded in science. However, the medical research base lags far behind in providing the clinical evidence that physicians need to treat patients, and the majority of recommendations in most treatment guidelines are not based on high-quality medical research studies .
Given our reliance on evidence-based guidelines, this shortfall in evidence is disturbing . The IOM reported that '[a] review of guidelines in the National Guidelines Clearinghouse reveals recommendation after recommendation that is supported by weak, mixed, or no evidence' . McAllister and colleagues found that only 68% of the recommendations in guidelines cited randomized controlled trials, and less than 30% were high-quality studies that could be applied to the patient population .
At issue is not only the level of the evidence that is cited (for example, controlled studies versus non-controlled studies), but also the internal and external validity of the research, including the size of samples, the endpoints that are evaluated, and the generalizability of the study to the patient population in the medical community at large. Ioannides and colleagues' analysis of research studies concluded that 'for most study designs and settings, it is more likely for a research claim to be false than true' .
The problem when the 'evidence' in evidence-based medicine does not deliver is the misperception, as Sniderman and Furberg noted, that 'guidelines are evidence based, not opinion based, and therefore their conclusions flow directly from the conclusions of studies' . Bias created by conflicts of interest on guidelines panels is exacerbated when evidence is weak. Under these circumstances, the real question is not what is 'right', but who decides.
Although evidence based medicine regards the randomized controlled trial as the gold standard, many forms of medical evidence exist. Scott Sehon of the Department of Philosophy at Bowdoin College describes the available broader evidence base:
[C]linical experience, observational studies, and RCTs have much in common. All are attempting to ascertain the safety and efficacy of interventions, and all do so by trying the intervention and noting the results....In each case, we will observe the treatment received by a patient, and then we will observe the outcome or endpoint for each patient .
With each of these three forms of evidence, knowledge is based on probability. Hence, to the physician 'knowing a diagnosis means. . .the diagnosis is highly probable [but] not absolute'. The same holds true for observational studies and RCTs, which are based on statistical averages that may or may not apply to a heterogeneous patient population. In any event, regardless of the pretreatment probabilities and the type of evidence employed, the 'absolute' can only be determined by the response of an individual patient to a specific treatment.
The issue of which evidence base should carry more weight in a particular recommendation is important. Put another way, the question is whether centralized medical decision making or individualized care should be given more weight. When the base of evidence is weak, non-existent, or conflicting, the treating physician may become the proverbial one-eyed king in the land of the blind. The clinician alone holds a key to the puzzle that is unavailable to members of a remote guidelines panel: patient histories, physical examination findings, and the patient's responsiveness to treatment. The American Academy of Pediatrics directly tackled this issue in its guidelines on creating guidelines: 'When the evidence is of low quality and the benefit-harm equilibrium is balanced, guideline developers generally should not constrain the clinician's discretion by making a recommendation but instead should designate acceptable alternatives as options'.
Evidence shortfalls that are filled with 'expert opinion' can take treatment choices out of the hands of treating physicians when clinical discretion is restricted and out of the hands of patients when treatment options are excluded from guidelines. Preservation of treatment options is a necessary underpinning to patient autonomy. Guideline panels that fill in evidentiary gaps with 'expert opinion' should be mindful of not interfering with medical decisions that properly belong to the patient. Drawing the line between paternalism and patient autonomy is a subtlety that pervades medical care.
In a clinical setting, the patient can respond to undesired paternalism by exercising 'voice or exit strategies' . Typically, the physician describes the choices and recommends that the patient follow a particular treatment approach. The patient can discuss concerns (voice) and the physician may then modify the treatment recommendation to reflect the patient's individual values . Alternatively, if the physician does not offer treatment options or if the patient does not improve under his care, the patient can exercise his right to leave the relationship (exit) and seek help elsewhere . Either way, the physician is accountable and bears the consequences of his actions, both legally and economically. In guidelines development, there are no voice or exit approaches available to the patient. Hence, guidelines panels need to be particularly reticent about interfering with treatment choices or clinical discretion.
The boundary between treatment recommendations that should be made by the panel and treatment options that should be preserved for patients depends upon (1) the quality of the evidence and (2) whether treatment decisions involve trade-offs between risks and benefits that depend on the values of the patient . Because of the lack of voice and exit options for patients in guideline treatment recommendations, a high degree of certainty should be required before interfering with patient choice and clinical discretion. In his treatise on participatory democracy and bioethics, Thomas May at the Medical College of Wisconsin's Center for the Study of Bioethics addressed the question of choice in the face of medical uncertainty. He observed that 'there is no question that amputation is the appropriate action in extreme cases of gangrene' and concluded that the 'role of bioethics is precisely that of calling into question whether any treatment is appropriate without question' . In the absence of such certainty, May concluded that it is important to preserve choice and to allow the individual whose life is most affected by that choice, the patient, to exercise autonomy of decision.
This comports with the American Medical Association's code of ethics, which states: 'The principle of patient autonomy requires that competent patients have the opportunity to choose among medically indicated treatments and to refuse any unwanted treatments' . As May explained:
Determining 'appropriate' treatment, then, is itself a determination that includes, among other things, a value judgment. As we have seen, what is right for one patient may not be right for another. Whether a given treatment is appropriate depends on whether the potential benefits of that treatment outweigh the burdens it imposes on the patient. This judgment requires that we consider the patient's perspective in assessing the benefits and burdens of treatment. To fail to do so is inconsistent with a liberal constitutional society and with the rights of a patient in such a society .
Ezekiel Emanuel, current advisor to President Obama, explains why choice, autonomy, and informed consent are central issues in medical ethics:
Most health policy analysts...see choice as essential because individuals are the best judges of their own interests. While individuals may not always be correct, they have more intimate knowledge of themselves, more reason to get it right, and more motivation to be dogged in pursuit of their interests. When choice is restricted or relegated to someone else, there is a high chance that individuals will be prevented from realizing their interests or their interest will be sacrificed to someone else's interest. This view is commonly cited by courts and others as one of the justifications for informed consent .
Paternalism is generally rooted in the belief that one is 'right' and that intervention is necessary to protect the patient from making a 'bad' decision. But, as David Buchanan at the Amherst School of Public Health and Health Sciences, explains, good intentions do not justify interference with personal liberty:
Dworkin defined paternalism as 'interference with a person's liberty of action justified by reason referring exclusively to the welfare. . .of the person being coerced'. Paternalism is the usurpation of decision-making power, by preventing people from doing what they have decided, interfering in how they arrive at their decisions or attempting to substitute one's judgment for theirs. . .The moral concern is that the presumption that one is right, and therefore justified in seeking to override other people's judgment, constitutes treating them as less than moral equals .
With autonomy '[t]he critical point is being in the position of deciding, not being decided for' . As May pointed out, the fact that an expert may disagree with the patient is beside the point; autonomy only becomes a real issue when an expert disagrees with the patient . Hence, one of the essential tasks of a guidelines panel is to identify and preserve 'choice' for the patient. Choice can be preserved by acknowledging divergent viewpoints on treatment, providing treatment options, or deferring to clinical judgment. Guidelines of highly influential medical societies need to be crafted with care not to interfere with patient choice.
When evidence is clear, well developed and uncontested, the need to preserve choice is often not an issue. However, it is essential that treatment choices be preserved when evidence is weak, mixed or not yet developed. As was noted in the discussion earlier, the lack of a strong evidentiary base supports the decisive role of clinical judgment exercised by the clinician rather than the remote recommendations of a guidelines panel.
Beyond this, however, what should be done to preserve legitimate treatment choice for the patient? Choice arises out of a diversity of viewpoints. As Sniderman and Furberg pointed out, '[b]ecause gaps in the evidence are inevitable, they must be filled in with judgments, and judgments tend to preserve previous positions. Thus, what is to be decided is often already decided with the selection of the deciders' . Hence, guidelines panel constitution may well determine the range of treatment options available to patients.
Hotly contested areas of medicine such as the 'Lyme Wars' arise most often, as Atkins and colleagues noted in Health Affairs, 'when evidence is weaker, outcomes are less certain, and parties disagree about the risks of acting in the face of uncertainty' . The IOM addressed the issue of guidelines panel constitution and concluded:
The inclusion of individuals with a range of relevant professional and other backgrounds on guideline development panels can help check financial, professional, and other sources of bias. . .[and] promote the fuller consideration of potential outcomes, relevant evidence, and aspects of implementation .
The concept of including a diverse panel is to ensure diverse viewpoints--that is, to ensure that the interests of different stakeholders are represented to the extent that their interests are legitimate and central to the practice of medicine. According to the IOM, 'clinical practice guidelines lie at the intersection of medical research, education, and practice' . It follows that researchers and community physicians should be included on the panel. Other stakeholders include patients, insurance providers and pharmaceutical companies.
Most evidence-based guidelines panels are filled with academic researchers rather than with community physicians. In part, this represents a hegemonic shift away from commercially independent physicians who have traditionally controlled medicine, toward academic researchers whose commercial funding as 'key opinion leaders' may frame their viewpoints . As Sniderman and Furberg point out, however, this hegemonic shift may merely transfer power from one authority based expert, the physician, to another authority based expert, the researcher . Clearly researchers offer something of value when they sit on guidelines panels, but their perspective is very different from that of the general practitioner who on a daily basis sees patients who do not conform to research populations and has first-hand experience with the heterogeneity of treatment response, the need to preserve clinical judgment in guidelines, and the importance of providing patients with treatment options. Moreover, as was recognized during the AIDS activist movement, there may be a 'clash between the canons of research [which emphasize protection of research subjects] and the canons of care [which emphasize availability of treatment options]' .
As was noted earlier, the standard of care for physicians is driven more and more by treatment guidelines. Before the trend toward centralized medical decision-making and evidence based guidelines, the standard of care was largely determined by the 'consensus' of practicing physicians. This consensus developed slowly over time and reflected the practices of many physicians. Today recommendations of a guidelines panel may serve as a proxy for the consensus of treating physicians in determining and driving the medical standard of care. In fact, '[f]or many clinicians guidelines have become the final arbiters of care' . Thus, it is critical that guidelines not convey a false sense of unanimity when none exists.
Guidelines need to elucidate rather than obscure serious debate regarding diagnosis and treatment since 'the very nature of scientific debate is that no right answer has emerged' [4, 51]. The only way to assure this transparency is through panel diversity reflecting the range of treatment variation and practice. This means that diverse viewpoints should be represented on guidelines panels, that true voting of recommendations should occur to determine the degree of consensus, and that legitimate controversy should be acknowledged and reflected in the guidelines text.
The other viewpoint that must be reflected in treatment guidelines is that of the patient. The patient's viewpoint is essential to preserve the 'choice' among medical treatment options upon which the right to autonomy rests. When legitimate treatment options are not reflected in guidelines, autonomy has no meaning. Patients may be represented by direct participation on panels or by their treating physicians, and the value of direct participation may vary depending on whether the patient has become an 'expert' on the disease (as those with chronic illnesses sometimes are) .
Direct participation on guidelines panels by 'expert' patients is not common, but is growing. For example, the Cochrane Consumer Network currently provides for patient involvement . The role of the consumer is to 'apply ethical principles such as human rights and civil rights', 'comment on choice', identify 'potential conflicts of interest' and to be 'vigilant to many possible sources of bias' . At the end of the day, the goal of patient representation is to preserve the 'choice' and treatment flexibility that makes the exercise of autonomy meaningful. Hence guidelines that recognize clinical judgment, provide treatment options, and safeguard the patient's right to make treatment decisions can achieve this goal without direct patient participation. However, the best way to ensure that these issues are appropriately addressed is to provide a seat at the table to those whose interests are at stake: the patients.