Evidence-Based Medicine and the Law
The Courts Confront Clinical Practice Guidelines
The "Politics" of Clinical Practice Guidelines
As some commentators have recognized, the development and promotion of EBM has important political dimensions. The rapidly evolving science of outcomes assessment and its expanding application to the everyday delivery of health care have the potential to change substantially the dynamics and power relationships in the health care provider community, affecting the prestige and pocketbooks of powerful vested interests. Not surprisingly, there is an inherent cultural (some would call it "political") conflict between those who support EBM and those who support continuance of the traditional regime in which individual physician's subjective and independent judgment holds sway. The ripples from this culture clash extend beyond health care practice and will surely influence how the law deals with medical evidence. The cultural/political dimension merits consideration here; and, in this regard, three additions to the literature in the 1990s offer useful insights. Consideration of these insights, interesting in their own right, will also inform our later exploration of the advisability and feasibility of a guidelines certification program.
Knowing and Acting in Medical Practice
Sandra Tanenbaum's 1994 article, "Knowing and Acting in Medical Practice: The Epistemological Politics of Outcomes Research," describes the results of a small-scale study she conducted in a department of internal medicine to discover how physicians reason about clinical care, learn from their experiences and other information available to them, and pass their knowledge on to their professional colleagues. Her study gives valuable indications as to whether physicians are likely to "buy into" outcomes research and let its results influence their diagnostic and therapeutic decision making. She observes that, at a basic level, physicians have two very different ways of processing information: "At the level of clinical practice, realism and empiricism represent two approaches to medical problem solving: deterministic and probabilistic. Deterministic reasoning searches out mechanisms of illness and therapy, including etiology, pathophysiology, and the mechanisms of action. Probabilistic reasoning draws on what past experience predicts, whatever the cause. The probabilist plays the odds while the determinist imagines the process" (Tanenbaum 1994: 30-31).
Tanenbaum goes on to analyze the various ways physicians make sense of what they experience in their clinical practice. She notes that they tell "interpretive stories" about their cases, using a deductive reasoning process to distill their experiences into values that guide their future treatment of patients and are passed on to their colleagues in ways both informal (e.g., lunchroom conversations) and formal (e.g., grand rounds, conference presentations, journal articles, etc.). Through these processes, "Every doctor accumulates a vast and idiosyncratic knowledge of medicine" (ibid.: 33, emphasis supplied). From these and several other perceptive and intriguing observations, she concludes that physicians' decisions are not likely to be greatly influenced by outcomes research. They are inclined to be distrustful of outcomes research studies and the conclusions drawn from them. Moreover, even when they accept that the research findings are accurate in general, they tend to find reasons to believe that their particular patient is atypical and, therefore, merits treatment different from that which might normally be indicated by the study results. To put it perhaps too simply, given the detail and subtlety of her analysis, doctors are strongly inclined to hold to their idiosyncratic views, even in the face of strong contrary research evidence. "The clinical medicine I observed was essentially interpretive and therefore irreducible to probabilities, no matter how rigorously derived" (ibid.: 31). Tanenbaum's analysis lays an important foundation for Gary Belkin's (1997) and Matthews's (1999) observations regarding doctors' skepticism about the kind of information that forms the core of evidence-based medicine.
The Technocratic Wish
Belkin's article "The Technocratic Wish: Making Sense and Finding Power in the 'Managed' Medical Marketplace" (1997: 509) discusses the "technocratic wish," which Belkin defines as "an appeal to objective measures to resolve contentious issues and/or clothe their resolution as scientifically logical and natural." In the arena of managed care, the technocratic wish takes the form of a search for a (seemingly, at least) objective and verifiable rationale to justify the shift of control from an entrenched medical elite to a new cadre of health services researchers, MCO executives, and government policy makers. This latter group views the country's health care needs, and thus leans toward allocating its health care resources, using a systems approach, looking at issues on a macro rather than micro level, and employing population-based rather than individual-based measures to assess the utility and cost-effectiveness of health care inputs. To put it another way, the "technocrats" tend to measure the success of health care activities by looking at their aggregate effect on populations rather than on individual patients, contrary to the clinician's natural tendency to focus on the individual patient she or he is currently treating.
To implement their health care philosophy, those who share the technocratic wish collect data from entire populations, crunch the numbers, and express their conclusions as to what works best in terms of population-wide statistics. Their methodology is more quantitative and, on the surface, at least, more objective than medicine's traditional way of assessing the effectiveness of health care, that is, by cumulating the qualitative observations of individual physicians based upon their experiences with individual patients. The first approach appears more objective, the latter more subjective, since with the latter every observation must go through an individual physician's personal, subjective filter before becoming a data point in the analysis. Belkin (ibid.: 512) points out, however, that the seemingly objective quantitative approach is vulnerable to a "subjectivity of objectivity." In other words, researchers necessarily make choices among the many different ways of constructing scientific tests and measures and of interpreting and framing the results. When the range of possible approaches is recognized, it can be seen that analysis that appears to be objective and inevitable actually is a mosaic of subjective choices. Thus the difference in objectivity between the new approach and the traditional approach may not be as great as devotees of the technocratic wish would have us believe.
In Belkin's view, managed care has embraced the technocratic wish in its desire to find a rationale and a mechanism for standardizing medical practice and reining in physicians' natural inclination to treat each patient as a special case. MCOs, health insurers, and federal health programs commonly seek to control, or at least guide, how physicians treat patients by pointing to the results of outcomes research and asserting that, by and large, this should determine what care is provided and how. This approach obviously runs counter to the natural tendencies of physicians described by Tanenbaum. Belkin's analysis bears on our present discussion by helping to address the question of how influential customary professional practice should be in setting standards for measuring physician performance when the application of EBM methodologies would point in a different direction. In other words, how much weight should EBM be given when it comes into conflict with the traditional dynamic of subjective judgment and physician autonomy—either the autonomy of individual physicians or the collective autonomy of professional groups?
Practice Guidelines and Tort Reform. Building on Belkin's analysis, Matthews's response, "Practice Guidelines and Tort Reform: The Legal System Confronts the Technocratic Wish" (1999), describes the battle for control of the health care system as one that pits the qualitatively oriented clinician—the "artful practitioner," as Matthews terms her—against the quantitatively oriented scientist. Operationally, the contest is between the subjective judgment of individual professionals derived from and applied to case-by-case experiences and across-the-board rules distilled by health services researchers from "large n" studies. It is also a contest between professional discretion and "cookbook medicine," between the "art" of medicine and cold, hard science. Tanenbaum, Belkin, and Matthews all recognize a broad distinction between hands-on clinicians and the more data-driven types who tend to be drawn into the fields of research medicine, health plan administration, and health policy formulation. The conflict between their worldviews is mirrored in the face-off between those who practice medicine in the trenches and those who run MCOs, governmental health programs, and the federal health policy process.
As a side note, it is interesting to contrast the view of physicians put forward in these three articles by Tanenbaum, Belkin, and Matthews with that expressed by Paul Starr in his classic 1982 book, The Social Transformation of American Medicine. Among Starr's many keen insights is his observation—not all that profound in itself—that physicians as a group acquired their prestige and power in large part by holding themselves out as men of science. Their devotion to the systematic, scientific pursuit of truth set them apart from, and above, lesser members of society. To now look in a more detailed and precise way, using the lens of the three more recent commentators, and see physicians—at least the archetypal physicians, clinicians—cast not as scientists but rather as "artful practitioners" with a bias against the emerging science of outcomes research and EBM is, at first, surprising. Upon reflection, however, this shift reveals something we all know, namely that things look different from up close than they do from farther away or, to put it another familiar way, that "all things are relative." Clinicians may fall at the less rigorously scientific end of the physician spectrum, according to Tanenbaum, Belkin, and Matthews; but don't forget that the spectrum in question is an array of scientifically oriented professionals. Therefore, it is possible, and mostly accurate, to regard physicians on the whole as people of science while, at the same time, noting vast differences among this group in terms of their approach to science. Just as in George Orwell's Animal Farm some pigs were "more equal than others," so in the world of medical science some physicians are "more scientific" than others.
Turning back to the difference of perspectives between medical practitioners and those who run health plans and participate in the policy process, it is apparent that this parallax reflects, and magnifies, the underlying tension between doing the best for individual patients and making the right systemic allocation choices for patients as a whole in what is essentially a zero-sum game. The Benthamesque (social utilitarian) goal of trying to achieve the greatest good for the greatest number is often in conflict with the goal of trying to serve individual patients unreservedly and rendering the best (although it is often the most expensive) care for the individual without regard to the longer-term systemic consequences.
Matthews's article extends Belkin's thesis by examining how the courts' treatment of CPGs can be expected to affect their adoption and use by the medical community. His analysis supports the view, expressed above, that allowing the use of CPGs only for defense purposes, as in Maine and Minnesota, was a political decision meant to stimulate the adoption and use of guidelines by a physician community otherwise reluctant to accept them, in part because of its fear of the liability consequences. Getting physicians to rely upon and conform their practice to guidelines based on quantitative, empirical, outcomes research is challenging. Even more challenging is getting them to acknowledge this reliance in a formal courtroom setting. As Matthews (1999: 287) observes, for "most of the twentieth century, the legal system has validated the medical profession's belief that the authority of clinical judgment derived from personally acquired disciplinary expertise." For that reason, permitting physicians to introduce evidence that they followed practice guidelines is "not merely a technical change in the law; rather, it is using the regulatory function of law to facilitate the health economists' construction of authoritative medical judgment based on the ideal of impersonal and standardized objectivity, that is, the technocratic wish" (ibid.).
It is interesting to note that in speaking of the "regulatory function of law," Matthews posits that law has two broad functions: litigation, the settlement of disputes, which is retrospective (backward-looking) in orientation, and regulation, which is prospective (forward-looking). The law's regulatory function is much more compatible with the notion of EBM, which casts things in probabilistic terms. Regulation, which looks to the future, necessarily deals with what might happen, a probabilistic notion, rather than trying to determine what actually did happen in a past incident, the obvious focus in litigation. Since our focus here is on how courts will deal with CPGs in a litigation situation, the tension between the essentially probabilistic nature of EBM evidence and the deterministic requirements of assessing and assigning responsibility for a past event is at its greatest—or so Matthews might argue.
Matthews's analysis addresses the factors that will dispose physicians to use CPGs in litigation; but what about the factors that will dispose courts to embrace them? Obviously, a court will admit CPGs as evidence and use them in a particular way if directed to do so by the legislature; but in the absence of a legislative mandate, what will courts be disposed to do? Just as the medical profession is accustomed to relying on the subjective opinions of physicians, courts have a long tradition of relying on the testimony of medical experts to establish what is professional custom, which, in turn, sets the legally required standard of care for the case at bar. It is intriguing to ponder possible parallels between the opposing mind-sets of the two groups described in the Belkin and Matthews articles and the mind-set of judges who will be called upon (still assuming the absence of a legislative command) to decide how to treat CPGs when they are offered as evidence. Presumably the reason that the law decided many years ago to use professional custom as the basis for setting the legal standard of care, was that courts had no way to know what was the proper treatment for a patient in a particular instance. The obvious solution was to look to the medical profession and assume that what it customarily did was the right thing to do. Today, a judge might be faced in such an instance with two proffers of conflicting evidence: one the traditional proffer of an expert witness who would testify how the profession commonly treats such a case, the other a practice guideline proffered to show the current EBM-based judgment as to what treatment works best in such a case. How would the judge likely rule on the admissibility of the latter type of evidence? If the evidence were admitted, how would the judge likely instruct the jury to treat it? These questions recall earlier discussion as to the various ways CPGs might be treated by the courts. However, the question here is not just what treatments are possible but, rather, what would te court most likely be disposed to do?
One might answer, simply, that the court would do whatever legal precedents require it to do. But, as Matthews's (1999: 288-294) discussion of a group of cases involving guidelines reveals, courts are not so constrained by precedent that they have no choice in this matter; and courts have gone in different ways when presented with that choice. As Matthews (ibid.: 289) puts it, "Regardless of which party in the dispute introduced the guideline, however, the litigation always casts the same issues into especially bold relief: the conflict between the impersonal objectivity of a guideline and the personalized expertise of the physician, or the conflict between the guideline's focus on general decision procedures and the legal (and medical) profession's focus on the particular facts of the case under consideration."
Might not judges in general be more disposed to rely on medical expert testimony than on the more technocratic input, the guidelines? Given its emphasis on precedent (past decisions), the legal system is inherently conservative and backward-looking. Therefore, since reliance on expert testimony is the long-established norm, change can be expected to come slowly. Moreover, law is much more a qualitative (verbal) than a quantitative discipline; judges might be less comfortable relying, or instructing a jury to rely, on a guideline derived from quantitative, statistical processes rather than on expert witnesses' "interpretive stories." On the other hand, in the post-Daubert era, as Shuman articulates, federal judges have a duty to ensure that juries are presented with scientific evidence that is enlightening and not confounding. Even state court judges, who are not directly touched by the Daubert ruling—except in those states that have opted to follow Daubert—may feel more inclined than previously to direct juries toward standard-of-care evidence that is more definitive, authoritative, and clear.
Does this mean that judges will favor the use of guidelines over the testimony of expert witnesses? Not necessarily. This is currently an area for speculation because, despite the academic work done to ferret it out, there is scant evidence to date of how judges will react to this new form of medical evidence. The only thing safe to say is that we cannot assume courts, left to their own choice, will move toward greater use of CPGs. If we want CPGs, and EBM generally, to have greater prominence in the courts, legislative action is needed. Maine, Minnesota, Kentucky, Florida, and a handful of other states have spoken, albeit not that fully and clearly, on the subject of guidelines. Most states, though, have not yet taken a position. This commentator, for one, believes that they should, but with federal leadership. EBM approaches, including the development and use of CPGs, are a better way of guiding the practice and progress of medicine. The nation's courts should be opened to the evidentiary use of guidelines. If it takes legislation to assure that this will come to pass in a timely fashion, a focused effort to promote such legislation seems the right prescription.
The Implications of the Political Terrain
The following section proposes a certification program for CPGs. How does the "political" terrain described above bear on the feasibility of such a program? Some general observations can be made, and, frankly, they offer little encouragement. If physicians are not disposed to accept the notion of CPGs in general, they are not likely to welcome a proposal to give guidelines a federal imprimatur and increased importance in the courts, moves intended to stimulate greater development and adoption of guidelines. In short, the terrain is not—at first glance, at least—all that favorable. Let's examine that terrain, though, as an architect might do in considering what kind of house, if any, can be built upon a difficult piece of land. Sometimes, as in the case of Frank Lloyd Wright's Fallingwater, a very serviceable and impressive edifice can be constructed on seemingly inhospitable real estate.
If, as Tanenbaum observes, clinicians are inclined to distrust the science underlying outcomes research, EBM, and CPGs, or to discount their relevance to individual practice situations, and if, as Belkin and Matthews contend, guidelines are largely seen as a mechanism for nonclinicians to use in controlling clinicians, a guidelines certification proposal can expect a tough, uphill climb to passage. Three questions will help to measure the steepness of the ascent. First, can physicians be brought to understand that they have more to gain than to dread from CPG certification? Second, how much political muscle can the clinician population exert against the program's passage? And third, where are the other political players in this contest likely to come down on guidelines certification?
In trying to identify the various political forces at work here and to predict how they will react to a guidelines certification program, one of Paul Starr's early articles comes to mind. Long before his defining work, The Social Transformation of American Medicine (1982), Starr recognized the interplay of professional, business, and public interests that sets the stage for legislative action. In "Malpractice: The Doctor's Discomfort," published in the New Republic in 1975, he explained how, and why, physicians and other health care providers, liability insurers, the legal profession, and the general public would position themselves around the countless legislative programs being put forward to correct the so-called medical malpractice crisis of the 1970s. His words, although addressed to a somewhat different issue, seem just as apt in the present context:
A crisis of any kind is a moment of potential discontinuity in social affairs, simultaneously a danger and an opportunity: a danger to the orderly operation of a social system but also an opportunity for those who wish to change it in some way. For this reason, crises are not always resisted; on the contrary, they are often welcomed. A crisis can be a truly marvelous mechanism for the withdrawal or suspension of established rights, and the acquisition and legitimization of new privileges. The power to create crises (and the opportunities they bring) and, just as important, the power to define them are not distributed equally in society. Some groups are better positioned than others to promote their conception of events and to enlarge their discomforts into social concerns. This is the case with the malpractice crisis. It is the malpractice problems of physicians and insurers, rather than those of patients, that now constitute a crisis. This is a reflection not so much of the relative seriousness of their difficulties as of their relative power to turn private into public troubles.
These two groups, the physicians and insurers, along with the legal profession, dominate the current situation. Within this triad, all the terms of debate are set. Whatever the resolution of the current crisis, it will almost certainly have to satisfy the interests of the three dominant parties. Less certainly will it have to satisfy the interests of the public. The public has no clear conception of what its interests are, whereas the three dominant parties most emphatically do. (Starr 1975: 17)
The players are somewhat different in the current context from that described above, but physicians are still squarely in the middle. Their voice is powerful and is heard not only from their own mouths but from those of their patients, who still, even in this age of diminished physician dominance, derive much of their view of the health care system from talking to their physicians. The current "backlash against managed care," which has political reverberations all the way up from the local to the presidential level, is substantially driven by physicians interpreting to their patients what is happening on the health care scene. If physicians feel threatened by the guidelines movement, they will see to it that their patients perceive it as a threat to them as well. "Cookbook medicine," especially if its recipes are being written by "the folks who brought you managed care" and endorsed by "big government," can easily be made to seem very threatening to the general public. The chief lesson from all this is simple and straightforward: physicians' opinions matter. If they are to throw their weight behind a guidelines program—not just to pass it but to embrace it in the ongoing fashion needed to make it work—they will have to be courted. The courting, however, should not take the constitutionally and ethically questionable form of "one-way street" legislation that allows use of guidelines only by and for the defense. It should, and must, be based substantially on education. Physicians must be shown how guidelines can work in their favor, not just against it, as they are inclined to believe currently.
A Certification Program for Clinical Practice Guidelines?
In a 1995 article, I floated as a "trial balloon" the possibility of establishing a federal program for certifying clinical practice guidelines (Rosoff 1995: 394). Now, five years later, the idea still appeals to me as a good way of spurring the development and acceptance of high-quality guidelines and of aiding the courts in making determinations involving quality-of-care and entitlement-to-benefits issues. As the use of CPGs continues to grow, courts will inevitably encounter them in proffers of proof by one or both sides in health care lawsuits. Except in the few states (e.g., Maine, Minnesota, and Florida, to name the most prominent) that have enacted legislation on the subject, judges will have no guidance on how to deal with them. As in past instances of case-by-case doctrinal development, they will cast about to find the right approach; and, presumably, they will eventually devise workable rules. However, the time, effort, and other costs of having this process be one of uncoordinated activity in the state courts are hard to justify. A CPG certification program, coupled with some direction to courts as to how to handle both certified and uncertified guidelines, could surely reduce the valuable resources that might otherwise be expended in this process—to say nothing of the false starts and mistakes, and possible injustices to litigants, along the way.
What is needed is a governmental certification program to give guidance to the courts in two main dimensions: first, in distinguishing guidelines entitled to full credibility and weight from those not as deserving and, second, in deciding what evidentiary weight to accord guidelines falling into each of these categories. One thing should be made clear at the outset: this proposal does not intend to be prescriptive about medical practice. It contemplates that there can, and should, and will, be multiple guidelines for given conditions. Any body (or anybody; in theory, it could be an individual as well as an entity) that wants to submit a guideline for certification may do so. The result envisioned would be a free-market competition in guidelines, but with a governmental "seal of approval" that makes it easy to identify a valid guideline. It would be possible for multiple, conflicting guidelines to be certified; and uncertified guidelines would have their place too, albeit a much less favored place. The program would not, I repeat, have government decree how patients should be treated or limit the advance of medical knowledge by enshrining the current state of practice in a bureaucratic orthodoxy. Nor would the program increase the exposure of physicians and other healthcare providers to litigation. To the contrary, it would lessen the need for litigation and, when litigation does occur, it would make it simpler, more efficient, and less costly. Notwithstanding that, in my view, the program's benefits would substantially outweigh its costs, I am under no delusion that it will meet with ready acceptance. Numerous issues and objections can be raised and inevitably will be. The more obvious and important of these are identified and addressed below.