CASE FOR EVIDENCE-BASED MEDICINE

J. Michael McGinnis, M.D., M.P.P.

Americans look to cutting-edge biomedical research and innovation to protect and improve their health and health care, with the expectation that they will receive the most appropriate care based on the best scientific evidence. Unprecedented innovations in the diagnosis, prediction, therapy, and long-term management of disease are bringing Americans closer than ever to a vision of personalized health care. Breakthroughs in just the past decade in human genomics, stem cell biology, proteomics, and immunology offer glimpses of new approaches than can now only be imagined. At the same time, this vision of progress is clouded by uncertainties about the health and economic consequences of clinical practices.  What diagnostic and therapeutic value is actually gained from current practices and new approaches, at what cost, and with what consequences?  Using our resources to take full advantage of genuine breakthroughs depends on knowledge that will inform wise choices among available options.

With exciting advances in medical procedures, pharmaceuticals, devices, and biotechnology, comes a sometimes confusing array of choices for patients and their care providers, making decisions increasingly complex about which medical interventions are best—for whom and under what circumstances. The lack of critical information and studies necessary to better inform these decisions places in jeopardy both the effectiveness and the efficiency of the medical care delivered to Americans.

Experience teaches that some clinical interventions in use today will be found to offer little or no clinical advantage over lower cost alternatives, or to present some unanticipated risk, or both.  At the same time, the adoption of promising new technologies can be delayed by a failure to link technology diffusion to timely demonstrations of value. A deeper commitment to evidence-based medicine can improve care on both counts—to accelerate the adoption and dissemination of effective, safe, and worthwhile interventions while avoiding those which are harmful or more costly than equally effective alternatives.  

To confront these emerging issues and increase the quality and value of health care delivered, better and timelier evidence is needed. As a nation, we ought to be better able to provide the guidance needed by capitalizing on the tremendous growth in biomedical science and innovation, the development of evaluative clinical sciences, and advances in health information technology. Capacity exists in both the public and the private sectors, but it is underdeveloped and inadequately applied. To effect needed systemic changes, the way forward will require cross-sector engagement of the key stakeholders in a careful consideration of the priorities, the methods, and the policies involved in developing and applying evidence for sound medical care. 

Health and Health Outcome Perspectives

Ultimately, our health care must be judged by how well it performs in improving health status. On this dimension, there have been many encouraging developments in the past two decades. Infant mortality has reached an all-time low for American babies, in part because of improved prenatal care, but also in part due to striking improvements in the care of high risk, low birthweight babies. Among children, treatment of several childhood cancers—such as soft tissue sarcoma, leukemia, and Hodgkin’s lymphoma—has advanced dramatically. For adults, advances in the treatment of high blood pressure, elevated serum cholesterol levels, coronary artery disease, HIV, and certain types of cancer—joined by lifestyle improvements such as reduced tobacco use, shifts away from saturated fat consumption, and greater focus on physical activity levels—have ushered in unprecedented reductions in adult death rates and in disability rates among older people.

Still, if we take the experience of other countries as an indication of what should be readily attainable for a nation so richly invested in health care, we are falling far short of the possible. Our infant mortality rate ranks 27^th in the world, and at 6.9 deaths per 1000 live births is far higher than those of Singapore (2.3) or Sweden (2.8). At 75 years for men and 80 years for women, we rank 27^th and 30^th, respectively, in life expectancy at birth. In a World Health Organization rating of overall health care system performance, the U.S. ranks 37^th in the world.

A primary challenge in this respect relates to inadequate application of available evidence. Numerous studies have documented a persistent variation in the adoption of proven health care interventions. One widely cited study suggested that U.S. adults currently receive only about half of recommended care for conditions assessed, with quality varying significantly by medical condition. For example, only about 45% of established care was delivered for diabetes, 23% for hip fractures, 65% for hypertension, and 54% for asthma. This problem has been shown to be ubiquitous, affecting even the most prestigious academic medical centers. Such quality deficiencies result in increased mortality and morbidity, as well as decreased quality of life and a less productive workforce.

But important barriers to better care also derive from deficiencies in the ways we generate and apply evidence for new interventions. The more public examples of these deficiencies fall in the category of too rapid adoption of unproven interventions. Prominent examples of interventions widely applied in advance of adequate information include hormone replacement therapy, COX-2 inhibitors for pain relief, autologous bone marrow transplant for advanced breast cancer, and fluoride treatment for osteoporosis. Our national capacity to generate and apply higher quality, more timely, and more reliable evidence about the relative effectiveness of candidate interventions is vital to the improvement of health care for Americans.    

Health Cost Perspectives

Development of new medical treatments, rising prices and growing utilization will continue to contribute to the upward trajectory of health care spending. In 2005, nearly $2 trillion—roughly 16% of the GDP—was spent to pay for health care among Americans. These costs will continue to rise and are predicted to reach 20% of the GDP by 2015.  While medical progress has improved health care for many, rising costs are a growing burden for households, businesses, and governments. Increasingly, health care costs are associated with reductions in the depth and breadth of employer-based health insurance coverage for U.S. workers, and are often cited as a factor reducing the ability of companies to remain competitive. For the third consecutive year, chief executive officers of U.S. companies have cited health care costs as their number one economic concern—a concern compounded by questions about the quality of the return on that investment. 

Many studies of the appropriateness of health care services, or the variability of service intensity for a given condition, suggest that far too high a proportion of our investment goes to activities that do not improve patient outcomes, and in some cases, may even be detrimental. Yet our national attempts to improve the efficiency of system performance are not directed at better understanding effectiveness. As the pace of patents filed for new drugs, diagnostics, medical devices, and biological products quickens—presenting both  opportunity and challenge—efforts to address health care spending have often focused on shifting and minimizing costs rather than understanding how scarce resources can be better targeted to effective, high-quality care that really makes a difference. Our current capacity is simply not up to the task of identifying which diagnostics and treatments work best, and only a small fraction—likely less than 0.1%—of each health care dollar is currently devoted to systematic research and assessment of the comparative effectiveness of various diagnostic and therapeutic options.

Consumer Perspectives

The American public has traditionally expressed strong support for medical care, research and technology development, while also expressing a strong interest in both individual patient prerogative and better information to aid patient and provider decision making, individually and in consultation with each other. Confidence in health care delivery may be beginning to wane. In a 2005 Research! America survey, 60% of Americans said they didn’t believe that the U.S. had the best health care system in the world, 41% said they knew of a time when they or a family member had received the wrong care, and 56% said there should be more investment in health services research.

A variety of recent initiatives, some by patient groups and some by medical and scientific groups—including the Institute of Medicine in its 2001 report Crossing the Quality Chasm—have emphasized the need for health care to focus more on the delivery of individualized, patient-centered care. This is the essence of evidence-based medicine. The movement towards a better understanding of what works best in practice, the transparent consideration of known risks and benefits, and a continual assessment of evidence gaps will provide tools for doctors and patients to better interpret evidence and risk.  This shift to the patient in terms of knowledge and responsibility will allow consumers to take a more active role in making informed choices about their health care.

A number of initiatives have been developed to improve the information for those choices through the application of assessment metrics that provide better insights into the quality of the care delivered. Work on quality standards by various organizations—public and private—has focused on producing information that might be used by purchasers, providers, and consumers to compare the performance of health care plans and institutions. Other organizations also work to diffuse these performance measures into everyday clinical practice in the interest of accelerating quality improvement in patient care. Central to the success of these efforts is the availability of reliable evidence on the comparative effectiveness of different interventions.

Research Perspectives

The nation’s capacity for clinical research is currently inadequate to provide the information needed. Randomized controlled trials (RCTs) represent the current gold standard for answering questions of efficacy—whether or not a diagnostic or therapeutic intervention works—and also provide important knowledge about the safety and costs of medical treatments.  But they are expensive, often difficult to complete, and provide limited insight on effectiveness—how well it works in more typical practice settings and patient populations, and the value it does or does not add to available alternatives.   Several tools, adequately supported and thoughtfully applied, may offer practical approaches to improving the evidence for decision making.  They include:

• “Practical” trials—studies designed specifically to answer questions faced by decision makers, and which therefore compare clinically relevant alternatives, focus on broadly representative patient populations recruited from heterogeneous practice settings, and collect data on a broad range of health outcomes. These include large simple trials, cluster-randomized studies, and time-series analyses of planned changes in care.
• Clinical registries—structured inventories of data on patients receiving particular interventions, with the potential to identify risks and benefits outside the context of formal RCTs and provide insights on both patient characteristics and provider profiles most conducive to positive results.
• Observational studies—using available data sources, including claims and medical records to answer certain questions about the outcomes of a particular intervention, and provide insights on both patient characteristics and provider profiles most conducive to positive results for an intervention.
• Model development—an approach pursued by some researchers as an alternative to clinical trials in those circumstances in which adequate knowledge and information exists about causal chains, physiologic pathways, and mechanisms of action to allow predictions based on information about patient population profiles on key dimensions.  

In addition to the design of dedicated studies, improvement in the application of existing and developing health and medical information sources holds the prospects for the retrieval of important lessons.  National data sets—e.g. the Census, Medical Expenditure Panel Survey, National Health and Nutrition Examination Survey, National Health Interview Survey—represent  rich, but substantially under-analyzed, sources of information useful to evaluation of the effectiveness of various interventions.  And interoperable electronic health records and patient information systems, an accomplished fact in some settings, will change fundamentally the ability to generate real-time information on medical effectiveness as an organic outgrowth of medical record-keeping.   

Because evidence unapplied is evidence wasted, bolstering the scientific foundation for clinical decision making must also include the development of improved methods of translation and implementation of knowledge into clinical practice. 

Recent Initiatives

Beyond the studies supported by manufacturers as part of product development, a number of agencies and organizations are working to improve the evidence for medical care. A number of large employers, private payers and provider organizations are similarly committed sponsors of related work. The efforts of these organizations are challenged both by the need for scale, and, in the case of payers, the perception of bias toward lower cost interventions.

At the federal level, some important, but modest, efforts have begun to engage these issues more systematically.  The Centers for Education and Research on Therapeutics (CERTs) program has operated since 1999 and currently receives $6 million per year from the Agency for Healthcare Research and Quality (AHRQ), supplemented by various public and private sources.  Although not focused solely on assessment of effectiveness, CERTs perform research on the safe and effective use of pharmaceuticals, biologics and devices.  Since 2004, Congress has also appropriated $15 million annually to AHRQ to support syntheses and related work focused on the effectiveness and comparative effectiveness of pharmaceuticals, devices and health care services of particular relevance to the Medicare, Medicaid and State Child Health Insurance Programs. With this support AHRQ has initiated its Effective Health Care Program. 

The Centers for Medicare and Medicaid Services (CMS) has a number of policies in place that help promote the development of better evidence for decisions, including: payment of routine costs of care for beneficiaries enrolled in clinical trials of the National Institutes of Health (NIH), such as those for lung volume reduction surgery and implantable cardioverter defibrillators; payment for experimental devices, under its Investigational Device Exemption policy, that are designated by the Food and Drug Administration (FDA) as similar to already proven devices and in trial expected to lead to approval; and issuance, as part of the Coverage with Evidence Development initiative, of coverage decisions linking expanded coverage to the conduct of practical trials and registries.  The NIH has funded several landmark comparative trials (such as ALLHAT, on hypertension and lipid control and NETT, on emphysema management) that have had substantial influence on medical decision making and underscored the potential impact of systematic work of this sort.

As important as they are, these initiatives might be considered rather limited down payments against a much larger current demand for systematic research and assessment on the relative effectiveness of various diagnostic and therapeutic options, on important safety questions, on individual variation, and on off-label uses.  If our nearly $2 trillion annual investment in health care had just 1% devoted to evaluating its effectiveness, support for that work would amount to $20 billion. 

A New Cooperation: Toward a National Trust for Evidence-Based Medicine

Because of the urgency of the need for greater national capacity to generate the information for doctors and patients in today’s healthcare environment, through the Roundtable on Evidence-Based Medicine the Institute of Medicine of the National Academies is providing an independent venue and stewardship for collaborative initiatives to improve the evidence for medical decisions.  The importance of such capacity, and the need for a public-private cooperative approach, has been an active topic of keen interest among providers, payers, employers, and manufacturers, as well as patient advocacy groups.  It has been promoted and underscored by the work of the Agency for Healthcare Research and Quality and given impetus by the Centers for Medicare and Medicaid Services’ initiative on coverage with evidence development.  Efforts by the Food and Drug Administration to strengthen its drug oversight capacity has also prompted discussion of practical, systematic and collaborative approaches to data gathering and assessment on which to base their determinations.  And it has been a recurrent theme at discussions of the IOM’s Clinical Research Roundtable.  The aim of this work is to explore ways to support reliable, credible, and independent studies that will provide the knowledge base needed to determine what diagnostic and treatment options are most effective for Americans.

Several assumptions provide the rationale for this cooperative approach:

• Multiple stakeholders—patients, providers, payers, purchasers, manufacturers—have interests both in the results of registries and in the broader work on clinical effectiveness.
• Much of the current work is now decentralized, isolated, and ad hoc—i.e. the “field” is scattered, unstructured, and inadequately supported.
• Efficiencies are possible through enhanced cooperation—e.g. shared and accelerated learnings, piggy-backing, economies of scale, methods improvement.
• Common interests are most compelling for common issues—i.e. non-rare events, diseases, treatments.
• Common interest suggests the need for common support.
• Appropriated discretionary government funding will not likely be a reliable source of necessary support.

The objectives of the IOM Roundtable include both finding ways to generate better evidence on which to base clinical decisions, and prompting stronger public and provider commitment to putting evidence into practice. Specific issues and agenda topics will span a broad range of issues in research, medical care, and health policy. Examples include:

• Assessment priorities. Which interventions do stakeholders feel are the highest priorities for evaluation?  By what criteria? How can they best be assessed? How can coordination be improved for assessment activities.
• Assessment approaches. What criteria should guide choices among various assessment tools—e.g. registries, other observational data (natural experiments), randomized clinical trials, practical clinical trials, modeling? How can these methods be improved and the lessons shared? What decision criteria should guide determinations on the sufficiency of evidence?
• Data issues. What are the constraints, issues and opportunities to advancing the public good aspect of health data? How should the protection requirements mandated under HIPAA (the Health Insurance Portability and Accountability Act) be engaged to expedite collection of evidence on medical effectiveness? What is needed for electronic health records to generate real-time evaluative information? Can an interoperable network of integrated health system data be developed to expedite assessment of new interventions?
• Policy issues.  How can policies on coverage and regulation better foster evidence generation? How can the ethical and confidentiality issues be best addressed? How can sustained support be developed for studies needed on comparative effectiveness of diagnostics and treatment interventions?
• Translation strategies. What are the most promising opportunities to accelerate the application of proven interventions in health care? What communication strategies might be most effective in translating to the public the nature of evolving science and medicine?
• Cooperative arrangements. What mechanisms might facilitate cooperative pooling of resources and information to generate better evidence on which to base clinical decisions? 

To address these issues, the Roundtable will work over the next three years in formal meetings, expert panels, workshops, and symposia. (March 2006)