Money and Mission? Addressing the Barriers to Evidence-Based Medicine

Although the central concept of evidence-based medicine (EBM), that medical practice should be based on critical review of the best available evidence from clinical research, may seem unassailable, the formalization of EBM is a recent phenomenon. Virtually unheard of before 1990, EBM is the topic of a growing number of publications. Practicing physicians still struggle with the concept and its implementation, however, as shown by McAlister and colleagues.¹ Although the majority of physicians McAlister et al. surveyed felt EBM can “play a positive role in medical practice,” and said they employed EBM “always” or “often,” the physicians actually relied mainly on clinical experience, review articles, and the opinions of colleagues when making clinical decisions. Many possible barriers to EBM have been catalogued.² The study by McAlister and colleagues suggests that EBM advocates still have to contest with three in particular: Physicians remain uncomfortable with the quantitative approach to medical practice; even those who accept this approach in principle have few opportunities to learn about it in detail; and physicians are often frustrated when the EBM process does not discover strong enough evidence to let them make decisions with confidence.

Understanding the historical context of EBM may illuminate these barriers. As Kerr White recounted in Healing the Schism,³ public health and medicine diverged in the early 1900s. The “schism” developed between the philosophy and priorities of older medical schools and those of new schools of public health, which had flourished with the aid of funding from the Rockefeller Foundation. Medical schools, in response to the Flexner Report, emphasized biomedical sciences and the investigation of the mechanisms of disease, while public health schools emphasized environmental and social influences on health and disease. Robust growth of the biomedical sciences lead to a dazzling array of new tests and treatments. However, physicians lagged in their understanding of the benefits and harms of these technological advances and their abilities to determine which patients would benefit most from which intervention in what circumstances. By the mid-twentieth century, the bad effects of the discrepancy between the power of the clinical armamentarium and physicians’ limited abilities to evaluate and choose among the tools in it became increasingly apparent. White and others responded with a movement to help physicians “see and believe that epidemiological concepts and methods would help them to practice better medicine.”⁴ Several foundations, including the Rockefeller Foundation, which may have felt some responsibility for generating the schism, began supporting academic clinical epidemiology units in medical schools, starting with the University of Pennsylvania (1978) and McMaster University (1980). Results of this movement included the formation of the International Clinical Epidemiology Network (INCLEN), first a Rockefeller Foundation project, now independent. This network helped to support many early advocates of EBM, and teaches physicians around the world about clinical epidemiology and related fields. Thus, the schism may be starting to heal.

Nonetheless, traditional medical education has produced “successive generations of physicians … who are largely innumerate,”³and thus may be uncomfortable with the quantitative cognition that EBM requires. Although short courses in EBM, like the ones sponsored by McMaster, are attracting more North American physicians, and EBM is beginning to infiltrate the medical school curriculum and graduate medical education,^5–7 only a few North American medical schools and graduate training programs seem to have given it significant emphasis. Despite grass-roots efforts by many dedicated clinicians, in most institutions, EBM remains the “missing curriculum.”⁸ To change this situation, EBM advocates first may have to convince the traditionally trained physicians who control medical education to change their own thinking about clinical problems before they are likely to restructure teaching programs in favor of EBM. This may be a tall order.

It would help if there were more money to support education about EBM. Few federal funds are targeted to EBM education. Managed care organizations (MCOs) have called for physicians trained in EBM: “what MCOs really want from training programs is a competent physician skilled in the application of knowledge, and with an ability to discern what is relevant, not necessarily a physician with a unique fund of knowledge.”⁹ Therefore, it seems logical that MCOs should be willing to pay to improve EBM-related teaching. So far, there is little evidence that they have.

Physicians’ discomfort with EBM may be manifested by concerns about its challenge to the “art of medicine.” If the “art of medicine” means addressing the values, concerns, and preferences of individual patients when making decisions, however, then EBM can facilitate this art. Evidence-based medicine encompasses decision analytic techniques, which can take into account human values when making decisions.¹⁰ Because techniques for assessing such values, (“utilities” in decision analytic parlance) are still rather crude, EBM advocates need to work diligently to improve them. It may still be better to measure values crudely but explicitly than to artfully guess at them.

To some, however, the “art of medicine” may mean freedom to make intuitive, creative, and idiosyncratic decisions—EBM certainly challenges this sort of art. When EBM can identify the best possible decision options for a clinical situation, using it would necessarily restrict the originality and thus the artistic creativity of the decision maker. Perhaps one consolation is that elegant solutions are still possible even when options are constrained by scientific evidence. Good engineering practice does not prevent the building of beautiful structures; nor do modern acoustics prevent the making of beautiful music. To do the best job for their patients, doctors may have to agree to relinquish some unconstrained originality.

A real frustration for physicians is that EBM often fails to yield clear guidance about how to handle particular clinical problems. Unfortunately, good evidence to answer common clinical questions (e.g., how to treat localized prostate cancer)¹¹ is often lacking. Evidence-based medicine requires the use of the best available evidence, even when the best available evidence does not come from randomized controlled trials,¹² and even if the evidence is conflicting or of poor quality. However, when the best evidence is poor or contradictory, the answers yielded by EBM may be fuzzy and tentative. The EBM process may thus surprise physicians by showing that the evidence supporting current practice is weak. Therefore, EBM may often be the bearer of bad news, not a popular role to play.¹³

The crisis in the quantity and quality of clinical research is too big for EBM advocates to solve alone, but we at least need to raise our voices in support of a solution. In particular, I believe it is vital for EBM supporters to address the disproportionately meager funding available for clinical research on common medical problems. For example, the Global Forum for Health Research estimated that research on the common medical problems that affect 90% of humanity world-wide receives only 10% of the available research funds (the “10/90” problem).¹⁴

United States funding to support clinical research on common health problems is also disproportionately meager. Although domestic health care spending now exceeds one trillion dollars,¹⁵ the National Institutes of Health (NIH), the foremost federal source of research funding, spends only 10%–18% of its approximately $14 billion budget on clinical research,¹⁶ while the number of physicians applying to the NIH for grants to do clinical research is shrinking.¹⁷ Although efforts are underway to bolster clinical research at the NIH, their results are yet to come.¹⁸ The Health Care Financing Administration pays for no significant amount of clinical research, while it continues to spend massive amounts of money for health care of unknown effectiveness. The Agency for Health Care Policy and Research (AHCPR) has done yeoman work in support of research on outcomes and effectiveness, and on methodology to aid EBM.¹⁹ Yet this agency is minuscule compared to the NIH. Further, AHCPR’s political backing remains insecure. It barely survived a political challenge which in part arose out of some physicians’ displeasure with its EBM-based guidelines.²⁰ In contrast, in the United Kingdom, the National Health Service spends about 1.5% of its total budget on clinical research, systematic reviews, and dissemination of these results (Sackett D. unpublished communication).

In the United States, private funds have been insufficient to make up for missing federal funds. One might think that private organizations that pay for health care should be willing to support research that might improve the quality or decrease the costs of that care, at least in their own self-interest. Large employers, however, who pay for about one-quarter of the health care in the United States¹⁵ have played no significant role in funding such clinical research. Although for-profit MCOs may be jumping on the disease-management bandwagon,²¹ they have made no concerted effort to support the clinical research that could really improve disease management. The development of an Academic Medicine and Managed Care Forum may be a step in the right direction, but it is too early to evaluate its results.

The role of pharmaceutical corporations, who are obliged to fund research to satisfy Food and Drug Administration regulatory requirements, must be gratefully acknowledged. Further, a number of foundations have dedicated significant parts of their budgets to fund clinical research and to fund programs to train physicians to do research, notably the Robert Wood Johnson Foundation.²² Yet they cannot take up the slack left by the large payers and the federal government.

In short, McAlister and colleagues have identified significant barriers to EBM. It is now up to all of us who think practicing medicine based on the best available evidence is important to help take these barriers down.