Leadership Commitments to Improve Value in Health Care: Finding Common Ground: Workshop Summary (2009)

Coordinator

Janet Woodcock, Food and Drug Administration

Other Contributors

Mark Benton, State of North Carolina Department of Health and Human Services; Nancy Derr, Food and Drug Administration; Mark Gibson, Oregon Health and Science University; Karen Milgate, Centers for Medicare and Medicaid Services; Jane Thorpe, Centers for Medicare and Medicaid Services

This chapter presents perspectives on two ways that the regulators sector—in particular, state and federal healthcare regulators, the Food and Drug Administration (FDA) and the Centers for Medicaid and Medicare Services (CMS)—can contribute to accelerating progress in the delivery of health care that is evidence driven. Although FDA and CMS regulate different aspects of health care—FDA regulates the marketing and use of medical products, whereas CMS regulates reimbursement for healthcare products and services for two of the largest healthcare programs in the country (Medicare and Medicaid)—both agencies share a critical interest in the safety and effectiveness of pharmaceuticals, medical devices, and healthcare services.

The Institute of Medicine (IOM) Roundtable on Evidence-Based Medicine is focused on furthering the use of the best available evidence in clinical decision making, that is, evidence that will help provide an understanding of which diagnostic or treatment intervention, among an array of possible options, is best for a given patient. For medical products, this can happen only if the system for identifying, quantifying, and qualifying the evidence that is gathered on the product throughout its life cycle is formalized. As the nation moves toward the personalization of treatment, it is critical

that ways to improve the evidence base be found. Making use of existing and emerging information management technologies (i.e., standards and information management systems) will be critical to efforts to formalize an evidence-based system of healthcare practice. Key to this effort, however, is combining the strengths and resources of the many stakeholders involved.

As this chapter explains, FDA’s contribution to this effort will primarily be in its ability to improve the quality and the type of evidence generated during the early phases of a medical product’s life cycle, as well as to improve the development, communication, and use of risk and efficacy information throughout the product’s life cycle. CMS’s key contribution, which will occur at both the federal and state levels, lies in its ability to leverage the broad healthcare system through initiatives and incentives that advance evidence-based medicine.

The following paragraphs describe the roles of FDA and CMS in evidence-based medicine. Although these two agencies are very different types of regulators, they share a keen interest in both the safety and the effectiveness of medical products and services and in informing and minimizing the risks associated with the diagnostic and therapeutic choices made in the delivery of health care. Medical product use can be seen as a continuum that begins with the discovery of a candidate product; moves through product development, testing in clinical studies with humans, and FDA marketing approval for specific indications; and proceeds to postmarketing use for the approved indications, including possible subsequent approval for other indications. Throughout this life cycle, substantial data are collected and analyzed to evaluate whether a product is safe and effective for its indicated use. In the premarketing phase, statutes and regulations require that a product’s safety and efficacy profile be carefully monitored and that the data be carefully quantified and qualified. Once a product goes on the market, however, the generation of further evidence about a product’s safety and effectiveness is not structured. To achieve the IOM Roundtable’s goal of achieving medical care that is more solidly based on the evidence, the careful quantification and qualification of evidence related to medical products needs to be extended throughout the life cycles of products.

The FDA has regulatory responsibility for drug and biological products and medical devices, beginning in the premarketing phase and lasting for the duration of a product’s life cycle. Before a product is marketed, FDA

oversees and advises the sponsor on the development of the data submitted in the marketing application to demonstrate the safety and efficacy of regulated products. FDA reviews the collected data submitted in marketing applications and makes approval decisions on the basis of safety and efficacy data. After a product is marketed, FDA is responsible for monitoring both the safety and the effective use of the product. In fulfilling its regulatory responsibilities, FDA carefully assesses the data that inform the use of regulated products and that can have a significant influence on the design of studies intended to delineate the attributes of a regulated product. In this role, FDA is also active in facilitating the implementation of advances in the biomedical, product development, and regulatory sciences. As discussed later in this chapter, it is in the latter role that FDA has the best opportunity to contribute to improving the evidence on which clinical decision making is based, but its work to improve monitoring of a product’s use is also key.

Recognizing the pipeline problem—that is, the slowdown during the past decade in the development of new and novel medical products—FDA launched (FDA, 2007a) its Critical Path Initiative (CPI) in 2004. This initiative is a long-term effort to modernize the ways in which regulated products are developed, evaluated, manufactured, and used. As part of this initiative, FDA has launched, often in partnership with others, a series of projects¹ to modernize the scientific tools (e.g., in vitro tests, assays, computer models, qualified biomarkers, and innovative study designs) and harness the potential of bioinformatics to improve the ways in which it evaluates regulated products and to help better predict the safety, effectiveness, manufactur-ability, and use of those products throughout their life cycles. Many of these efforts will result in the improvement of the quantity, quality, and utility of primary clinical effectiveness data as well as facilitate the analysis of secondary clinical effectiveness data.

The white paper developed as background for Roundtable deliberations, Learning What Works Best: The Nation’s Need for Evidence on Comparative Effectiveness in Health Care, describes two broad categories of clinical effectiveness research. The first category, primary clinical effectiveness research, which comprises direct comparisons between interventions or between an intervention and no therapy, includes a range of randomized clinical trial designs as well as observational and cohort studies. Primary clinical effectiveness research delivers the principal data that product developers rely on to demonstrate the safety and effectiveness of an intervention. The second category, secondary clinical effectiveness research (or evidence synthesis), is defined as the structured assessment of evidence from multiple primary sources

mostly for the purpose of reaching conclusions about an intervention(s) that cannot be deduced from the individual studies alone (e.g., meta-analyses). FDA more typically uses data from secondary clinical effectiveness research to understand safety issues related to a regulated product.

As explained in more detail below, FDA’s statutory authority may not always permit the agency to require significant primary clinical effectiveness research that both is comparative and provides a sound basis for understanding how a new intervention is best used among the existing treatment options. Typically, the primary focus of the major efficacy studies intended to support the approval of new products is the development of data that establish that an intervention has a beneficial clinical effect (i.e., that it is effective and safe for a given use in a defined population). The findings obtained from types of study designs that are the most effective and efficient at establishing that an intervention has a clinical benefit—prospective, randomized comparisons of the new intervention and a control intervention—are often not broadly generalizable to populations beyond the populations enrolled in the study (e.g., the findings might not apply to younger or older patients or to patients with different stages of a disease). In addition, the efficacy findings might not contribute much to the understanding of how the intervention differs from among an array of options. FDA can, however, implement or facilitate a range of activities that can improve the quantity, quality, and utility of primary clinical effectiveness data generated for regulatory purposes. FDA is doing that under the CPI.

With expenditures of approximately $650 billion in 2006 and with more than 90 million beneficiaries, CMS plays a key role in the overall direction of the healthcare system. It is CMS’s mission to ensure effective, up-to-date healthcare coverage and to promote quality care for its beneficiaries. CMS’s mission is to achieve a transformed and modernized healthcare system in which patients and doctors together can make informed decisions about the most effective medical care, based on timely access to the latest evidence, in a way that delivers the highest-value care. This will help to ensure the right care for the right patient at the right time.

CMS is undertaking a number of efforts that support the IOM Roundtable goal. Among those that are particularly relevant to evidence-based medicine are the implementation of SMART health care (science-driven opportunity for management of personal health through affordable, reliable, and targeted care), encouragement of the use of secure electronic health records and personal health records, advancements in the use of electronic prescribing, and the creation of new disease management programs. CMS is also emphasizing prevention, implementing pay-for-performance systems

to promote better-quality and more efficient care, establishing an integrated data repository, and modernizing health information capabilities. Finally, CMS is encouraging health plan and drug plan sponsors to improve the coordination of care and to develop innovative approaches to improving the quality of care for its beneficiaries. These activities directly or indirectly support the development, collection, quantification, and qualification of evidence and encourage the application of evidence to guide the delivery of care.

Evidence is therefore an anchor in CMS operations. CMS uses evidence when making national coverage determinations, when determining whether CMS paid correctly for an item or service, and when determining whether an item or service was provided or performed in a quality manner. CMS is keenly interested in encouraging the development of better evidence and in ensuring that evidence-based medicine is used through a variety of regulatory and other incentives. In addition to the efforts noted above, CMS’s Coverage with Evidence Development protocol, the clinical trials policy, and the use of registries support this work.

CMS recognizes that collaborative partnerships with a variety of organizations, individuals, and institutions are necessary to support quality measurement efforts. For example, CMS works closely with the Agency for Healthcare Research and Quality (AHRQ) as well as the National Committee for Quality Assurance and other public–private quality alliances, such as the AQA alliance (formerly known as the Ambulatory Quality Care alliance), the Hospital Quality Alliance, and the National Quality Forum to develop, adopt, and implement quality measures. Through such collaborations, CMS is working to help create and sustain a better environment for the provision of high-quality, personalized care to every person every time.

Within broad national guidelines set by federal statutes, the states have flexibility in determining the final form of the program in their own jurisdictions. This flexibility includes determining what eligibility criteria are used to control access to coverage under the Medicaid program and to determine which services will be provided for those who are covered. Although there are certain federally mandated benefits (established by CMS), such benefits can be waived for demonstration and research purposes. For example, under such waivers, Utah has developed an approach that emphasizes preventive and primary care and has a very limited hospital benefit. Oregon has developed a priority list, which is linked to its budget that determines which services that it covers. In the formation of each of these programs, the states used clinical evidence and outcomes-based research to inform the decisions that shaped the policies.

Similar to the federal government’s role in Medicare, the state’s role in Medicaid is that of a large purchaser. In fact, Medicaid programs sometimes cover more lives than any other single insurer in a given state. Because of limits in the Medicare program, many low-income adults over age 65 years are often covered by both Medicare and Medicaid. Until the passage of Medicare Part D (the prescription drug benefit under Medicare), these “dual eligibles” depended on Medicaid to help them purchase prescription drugs, and because Medicare has a limited long-term care benefit, many adults over age 65 years depend on Medicaid to provide community and institutionally based long-term care. Across this spectrum of healthcare services, the states are moving steadily toward the greater use of clinical evidence to guide their policies.

A primary difference between Medicare and Medicaid has to do with their funding sources. Medicare is funded solely by the federal government. Medicaid is funded jointly by the federal government and states through a system that uses matching funds. States establish overall spending levels and then state treasuries are required to fund a predetermined share of the cost on the basis of the state’s economic well-being. The federal government then covers its share of the costs incurred. Because states determine the spending level, cost is an important consideration for policy makers in the state system than in the federal system. A delicate balance must be struck between the needs of residents who cannot afford to finance their own care, the demands of healthcare institutions and providers, and the need for policy makers to be good stewards of public resources in light of the limited amount of taxation that the public will bear. Once again, the use of research evidence is helpful in making certain that maximum value is obtained for the limited resources available.

The states directly regulate the practice of medicine and the healthcare workforce. This regulatory authority has its foundation in the 10th Amendment to the U.S. Constitution. Because these duties are not assigned to the federal government by the Constitution, this amendment provides the states the right to enact laws and regulations to protect the health and general welfare of their residents. Because the inappropriate practice of medicine could result in significant harm to the public, each of the 50 states, the District of Columbia, and each of the U.S. territories have enacted medical practices acts that define the practice of medicine and delegates the authority to enforce the act to a board composed mainly but not entirely of physicians. Although the structures of the state regulatory frameworks vary, the common elements that state policies oversee are the initial and ongoing licensing of physicians, including education requirements; examinations; continuing

education; and physician conduct, including the appropriate treatment of patients and the regulation of physician conduct.

As the gatekeepers to all that the healthcare system has to offer, physicians play the pivotal role in whether or not the public will receive the benefit of the knowledge generated through initiatives that increase the amount of good-quality evidence. If physicians are unable to access the information or do not choose to incorporate existing knowledge into their practices, the enhanced knowledge will not translate into better health outcomes for patients. In a perfect world, the rapid adoption and use of the best available clinical evidence would be the community norm; however, there is reasonable evidence that this does not always happen. State regulators are increasingly turning their attention to ways in which they can help physicians keep up with the changes that are sweeping through their profession.

The states are the primary regulators of health insurance in the United States, just as they are the primary regulators of the practice of medicine. However, states regulate only commercial (including nonprofit) health insurance that is purchased from health insurance companies or commercial managed care health plans. This distinction is important, because the Employee Retirement Income Security Act (ERISA) preempts state regulation of employee benefit plans that self-insure, which includes approximately 60 percent of employment-based health insurance coverage.

In addition to ERISA, the federal government has passed several other pieces of legislation that affect narrow but important elements of health insurance for both commercial offerings and self-insured employers. These include the Consolidated Omnibus Budget Reconciliation Act (1986), which requires that certain individuals who have lost their employment-based insurance coverage be given the opportunity to purchase a continuation of that coverage for a limited time; the Health Insurance Portability and Accountability Act, which limits how long insurers can exclude coverage of preexisting conditions for new enrollees; and two other laws that require coverage for minimum hospital stays for newborn children and their mothers and reconstructive breast surgery for women who have had breast cancer. These laws are enforced by the U.S. Department of Labor. The states govern virtually all other features of health insurance. This authority encompasses everything from requirements relating to financial solvency to access to health insurance coverage policies and the oversight of certain clinical issues. Because healthcare organizations and providers depend on insurers for much of their revenue, the regulatory actions of the states play a major role in not only how medical services are financed for millions of Americans but also how health care is delivered in any given location.

Health insurance companies doing business in a given location must meet state requirements. The most fundamental requirement is financial solvency. Regulations require companies to have sufficient fiscal resources and reserves to fund care for conditions found while they are covering a given patient. In addition, state regulations may control everything from what questions can be asked of a person applying for coverage to which services must be included in a policy and which are optional and to whether or not the premiums charged are reasonable. Clinical evidence influences a subset of these activities, primarily the definition of mandatory services for inclusion in policies and the determination of when covered services must be provided.

Large quantities of primary clinical data are generated pursuant to FDA requirements to demonstrate the effectiveness and safety of drug and biological products and medical devices before marketing. Typically, multiple clinical studies of various sizes, designs, and purposes support a marketing application. These may include, for example,

• studies intended to determine the pharmacokinetic properties of a drug (absorption, metabolism, and excretion);

• studies intended to determine the pharmacological effects of a drug (its mechanism of action, structure-activity relationships, drug interactions);

• studies intended to determine or optimize dosing;

• small randomized trials to obtain preliminary evidence of safety and effectiveness in individuals with the disease of interest; and

• large randomized trials comparing the new intervention with a control(s) intervention; these are intended to establish that the drug meets the statutory standards for safety and effectiveness necessary to obtain marketing approval.

Clearly, FDA and its regulations affect the quality, quantity, and utility of primary clinical effectiveness data generated for regulatory purposes. However, as the authors of the IOM’s white paper Learning What Works Best: The Nation’s Need for Evidence on Comparative Effectiveness in Health Care, point out, there is often somewhat of a disconnect between the evidence that will most efficiently demonstrate the safety and effectiveness for the purpose of obtaining marketing approval and the evidence that would be best suited to clinical decision making about how to make the

best use of a new intervention in clinical practice, that is, understanding which intervention among an array of options is best for a given patient. The primary use for many of the data generated to support a marketing application is to elucidate the various product-specific characteristics of the intervention that must be known before its marketing, and those data have somewhat limited utility for determining how the intervention fits into the medical armamentarium.

Various organizations have suggested that studies conducted to better understand the optimal use of multiple related interventions might best be coordinated by an independent entity capable of identifying evidentiary needs, prioritizing those needs, and funding research to address the priorities on behalf of a broad coalition of affected parties.

During the postmarketing phase of a product’s life cycle, FDA oversees a mostly passive safety surveillance system—that is, it largely depends on spontaneous voluntary reports from healthcare professionals and patients of suspected adverse events noted during standard patient care—but it may also require the development of additional clinical data after a drug is marketed (so-called Phase IV studies). Manufacturers are required to report to FDA adverse events reports it receives.

FDA’s postmarketing drug safety surveillance program receives more than 400,000 adverse event reports per year. FDA’s initiative to increase the submission of these reports electronically is progressing successfully. Currently, FDA receives approximately 54 percent of total adverse event reports from drug manufacturers electronically; 79 percent of manufacturer reports concerning events with a serious outcome not included among the possible adverse events listed on the product label are reported electronically. The internationally standardized electronic format has resulted in the timely electronic receipt of safety information, enhancing FDA’s ability to make informed decisions more quickly. Furthermore, agency costs have substantially decreased with the electronic receipt of reports.

To enhance the ability to better capture safety signals within the Adverse Event Reporting System (AERS), FDA uses data-mining techniques in its review of AERS data. Data mining is a tool that helps detect safety signals that might otherwise not be recognized or identified in a more timely fashion than by the review of spontaneous case reports. Data mining identifies adverse events in the AERS database occurring with the use of a particular drug more frequently than would be expected. Such signals can then be

pursued with more intensive case evaluation. It is important to realize that data mining looks at associations in the database between drugs and adverse events. It makes no inference regarding a causal role for the drug in the development of the adverse event.

FDA often requests, as a condition for approval, that additional studies (called Phase IV studies) be conducted after approval for the purposes of

• comparing two drugs (e.g., comparison of the new drug with a standard therapy);

• comparing a new drug plus standard therapy with standard therapy alone;

• evaluating long-term efficacy (e.g., when premarketing studies evaluated only short-term efficacy);

• evaluating drug use or dosing in pediatric populations;

• evaluating the use of a drug in geriatric populations;

• evaluating the use of a drug in patients with renal or hepatic failure;

• optimizing dosing regimens;

• evaluating the interactions of a drug with drugs that are likely to be administered concomitantly; and

• evaluating a specific safety concern that arose during drug development.

Data from Phase IV studies make an important contribution to the knowledge base about a drug and often contribute to understanding how a drug is being used. However, their utility is still somewhat limited, in that they ordinarily do not provide the kind of comprehensive evidence that would facilitate optimal therapeutic decision making in situations with many therapeutic options. Discussed here are activities that will enhance FDA’s evidence base for clinical decision making.

FDA oversees both the generation of evidence about the uses of medical products and the communication of that information to healthcare professionals and, to some extent, patients. Prescription drug labeling or prescribing information is the primary information tool for communicating drug information to healthcare professionals. Prescribing information focuses on a drug’s indications (or uses), dosing, and safety concerns. It also describes

the data that were the primary basis for FDA approval of the drug. In September 2005, FDA implemented a final rule that completely revised the content and format of the prescribing information for prescription drugs and biologics and was the most complete change that had been made in more than 50 years. The changes to the prescribing information, including a highlights or summary section, a table of contents, numbering of sections and subsections, and reordering of the content on the basis of the frequency with which healthcare practitioners refer to the different types of information in labeling, are intended to make the information therein more accessible, clinically informative, and easy to use. In addition, the new label format provides clinicians with a list of sections that have been modified or updated in the previous 12 months.

FDA also uses a variety of communication mechanisms to disseminate new information about medical products, in particular, emerging safety issues. In recent years, FDA has made efforts to disseminate emerging information about safety concerns before a complete regulatory review and action has been taken. On September 22, 2006, the IOM released a report entitled The Future of Drug Safety—Promoting and Protecting the Health of the Public (Institute of Medicine, 2006). The IOM report made a series of recommendations, including recommendations about how FDA could improve its drug safety program. In March 2007, FDA issued a response (FDA, 2007b) to that report, in which FDA outlined in detail what efforts were already under way and what efforts were planned to make FDA’s drug safety and risk communications program more robust. Although these efforts are too extensive to be discussed here in detail, a few key life cycle efforts are briefly described.

During the past decade, FDA has invested in a number of areas that are improving its ability to systematically monitor the use of a product over its entire life cycle. The safety of the medical products that FDA regulates has always been a key focus of FDA’s commitment to its mission to protect and promote the public health, and monitoring of safety postmarketing is an important piece of any effort to strengthen the evidence base. A number of efforts are under way at FDA that, once they are implemented, will strengthen the nation’s ability to identify the risks related to medical products and to collect, quantify, and qualify the risk information to help inform the growing evidence base. These efforts are being enhanced through the increasing use of emerging information technologies.

Extensive information on the safety and effectiveness of medical products is available from many different sources, including medical product developers, healthcare practitioners, and healthcare payers, not to mention from FDA, CMS, the Veterans Health Administration (VHA), and others. The information is useful, however, only if it gets to those who need it when they need it. FDA is exploring ways to take advantage of the large amount of data that are available on product risk and effectiveness to enable the sharing of important safety and effectiveness data. To build a solid evidence base for medical products, FDA needs to develop not only a way to collect, quantify, and qualify relevant information but also a way to communicate key findings to point of care in the clinic. FDA is working to modernize the processes by which it collects, analyzes, and communicates to the public important risk information, including the creation of user-friendly information sheets for healthcare professionals on emerging safety issues and a periodic, World Wide Web–based newsletter for healthcare professionals on important drug safety information. FDA’s MedWatch website and listserv disseminate critical health-related information on all FDA-regulated medical products and encourage participants to report possible adverse events to the agency through the website.

Electronic prescribing, electronic health records, and electronic adverse event reporting are all important areas whose development and use FDA is encouraging to improve evidence-based practice. At FDA, bioinformatics involves the design, development, and use of modern computer systems to efficiently and effectively manage the regulatory product information that FDA receives and communicates. FDA relies on the management of this information to assess a medical product’s safety and effectiveness throughout its life cycle. The current bioinformatics infrastructure that supports the exchange of information about products is undergoing modernization.

Other FDA initiatives to move the agency into an electronic environment include standardizing the electronic submission of premarketing data, updating and improving the ability to receive and disseminate adverse event reports, and enabling manufacturers to update the registration and listing of their facilities and products electronically using the Internet. In 2007, FDA began requiring that labeling information be submitted to FDA electronically by use of an approved standard, Structured Product Labeling, which is making it possible to efficiently develop and maintain a large database of labeling information on prescription drug products on the Internet. Health-

care professionals and the public are now able to access the most current prescribing information, the professional label, cost free on the Internet at a National Library of Medicine-sponsored website called DailyMed (http://dailymed.nlm.nih.gov/dailymed/about.cfm).

Bioinformatics modernization requires improvements in three important information management domains: access, standards, and interface. Better access to information, more standardized information, and better interfaces to information (i.e., better tools to convert information into knowledge) are needed. As these improvements are made to the system, FDA is consulting extensively in the United States and internationally² with the regulated industry, relevant healthcare-related organizations, and the public to ensure that the vast information at its disposal can be shared across systems in a secure fashion when it is needed. FDA is also working closely with many other federal agencies, (e.g., CMS, VHA, and the National Institutes of Health [NIH], among others) individually and through the office of the National Coordinator for Health Information Technology to standardize and harmonize standards and systems. The potential advantages of this long-term effort to fulfilling the reality of a national electronic health environment are huge, with the potential for research being equally as large. For example, once product information is standardized and can be exchanged across systems, the search and analysis of datasets to try to learn something about a disease’s pathogenesis or differential response because of, for example, genetic variation will become a routine undertaking. Modernizing the nation’s bioinformatics infrastructure is costly, complex, and time-consuming. Nonetheless, it is a necessary step to creating a modern, efficient American health information system.

A necessary key component to modernizing the information management framework (and formalizing the creation of a national evidence base) is to develop, agree to, and implement nationwide shared standards. For example, today, medical product safety information is largely nonstandardized, making even the most basic analyses and data-mining efforts difficult and time-consuming. The lack of data standards also impedes the development of newer analytic tools. For example, although the definitions in the MedDRA (Medical Dictionary for Regulatory Activities) have emerged as

the worldwide regulatory terminology for the reporting of adverse events in the postmarketing setting, other important data elements in a typical safety report (e.g., the medication, ingredient, or device name) are not associated with standard terms or codes that can facilitate data mining and analysis. Furthermore, the lack of a standard format for the exchange of safety information means that information exchange among key stakeholders is often slow, cumbersome, and manually labor-intensive. These areas are the focus of continuing work.

Several elements of CMS policy are key for activities related to the evidence base for medical care. At the most basic level, Medicare’s payment systems affect the development and use of evidence for decisions relating to coverage, coding, and payment. For every new item or service, CMS must decide what it can and should pay for, how payment should be made through coding and coverage decisions, and how much it should pay. CMS also plays a role in encouraging the development of better evidence and in ensuring that evidence-based medicine is used through a variety of regulatory and other incentives.

CMS relies heavily on clinical evidence in its coverage determination process. In determining whether to cover an item or service, CMS must make a determination of whether that item or service is reasonable and necessary for the treatment of Medicare beneficiaries. CMS may make coverage decisions in several ways. The first mechanism is the National Coverage Decision (NCD) process. Through the NCD process, CMS may grant, limit, or exclude Medicare coverage for an item or service. When an application for an NCD is approved, experts review the available scientific and clinical evidence to determine the effectiveness of the item or service in question. A judgment about the adequacy of the evidence for making coverage decisions depends on the methodological quality of the available research and the magnitude of the effect of an item or service on specific clinical outcomes. By using the principles of evidence-based medicine, the aggregate evidence is used to draw conclusions about whether the item or service under review is “reasonable and necessary for the diagnosis or treatment of illness or injury or to improve the functioning of a malformed body member.” Pursuant to the Medicare Prescription Drug Modernization and Improvement Act of 2003 (MMA), CMS has also expedited its process for the consideration of new or expanded coverage.

NCDs actually account for a small portion of coverage decisions. Many coverage decisions are made by local contractors and are referred to as Local Coverage Decisions (LCDs). These decisions are processed faster and are less formal than the NCDs. Because of differences in the practice of medicine in different regions, communities, and institutions, it is often preferable not to have a single decision that uniformly applies to all providers. This may be due to insufficient information for determination of whether coverage should be provided on a national basis or to legitimate regional differences in the delivery of health care. LCDs often provide guidance to providers, suppliers, and beneficiaries in the absence of an NCD as well.

As authorized by the MMA, CMS has also developed the Coverage with Evidence Development (CED) protocol, which is intended to enable Medicare to provide payment for items and services under conditions that help ensure significant net benefits of the treatment for beneficiaries and that give rise to additional information. This evidence also assists doctors and patients in better understanding the risks, benefits, and costs of alternative diagnostic and treatment options. Consequently, the linkage of coverage to data collection will also help to ensure that individual patients are receiving care that is reasonable and necessary given their specific clinical situation.

CMS believes that systematic, protocol-driven data collection has the potential to increase the likelihood of improved health outcomes. The care provided under these protocols generally involves greater attention to appropriate patient evaluation and selection, as well as the appropriate application of the technology. These additional data may alter the course of patient treatment on the basis of the best available evidence and may lead a physician to reconsider the use of the item or service or otherwise alter a patient’s management plan, potentially improving health outcomes. Two current applications of the CMS CED protocol include the implantable cardioveter defibrillator and positron emission tomography registries, which collect information about patients receiving these treatments.

On the basis of the recommendations from the Medicare Evidence Development and Coverage Advisory Committee, in July 2007, the clinical trial policy was expanded to include paying for the investigational clinical service if covered by Medicare outside a trial or required through an NCD (i.e., CED).

This change exemplifies the agency’s commitment to providing access to services for beneficiaries by encouraging the conduct of research studies that add to the knowledge base about the efficient, appropriate, and effective use of products and technologies in the Medicare population, thus improving the quality of care that Medicare beneficiaries receive.

Coding of medical items and services is crucial to the functioning of the healthcare system. Medicare alone processes about 1 billion claims for items and services each year. On each claim, codes indicate what items or services were provided and why. The code sets most commonly used and recognized by the secretary of U.S. Department of Health and Human Services (HHS) as standards pursuant to the Health Insurance Portability and Accountability Act are the International Classification of Diseases, 9th edition (ICD-9), and the Healthcare Common Procedure Coding System (HCPCS). ICD-9 is used for diagnoses in all settings and for procedures in the inpatient setting. HCPCS is used for services and items in ambulatory-care settings.

Many areas of the ICD-9 code set are full, making it difficult to add new codes. CMS is considering whether to move to its successor, ICD-10. HCPCS consists of two subsystems: Level I and Level II. Level I is the Current Procedural Terminology (CPT) system and is maintained and updated by the American Medical Association. CPT codes identify the medical services that physicians and other healthcare professionals perform. CMS maintains and updates Level II codes and identifies products, services, durable medical equipment, prosthetics, orthotics, and supplies (DMEPOS). CMS is in the process of phasing in updates to the coding cycle and improving the process for the development of new codes.

Ultimately, CMS is responsible for reimbursing providers and suppliers for covered items and services under the Medicare program. In general, Medicare pays on the basis of fixed payment rates that are revised periodically (annually for most systems). However, as the cost of medical care continues to escalate, CMS is developing incentive programs that can be built into the payment system to encourage the delivery of high-quality, efficient care. Measuring and incentivizing high-quality care requires sound evidence to support those decisions. CMS is working with stakeholders to develop measures that can be used to measure quality and modify current payment systems to enable the use of the results of these measures to affect payment rates.

Just as CMS must determine whether a service or device is reasonable and necessary for the treatment of Medicare beneficiaries, so must the states

make those determinations for Medicaid patients. Although the NCDs and LCDs that CMS makes for the Medicare program can influence whether or not a service will be funded by Medicaid, the states are not required per se to duplicate the coverage. Typically, the NCDs have a greater impact because of the more inclusive and thorough process that accompanies the making of such a decision. However, these federal decisions cover only a fraction of the items that states must review and rule on, and often, the time frame over which the states must make these decisions is shorter than that at the federal level.

Because most Medicaid beneficiaries do not have significant disposable income, Medicaid programs are using evidence to determine which optional benefits will be included in the benefit design. Although it seems illogical, prescription drugs are still an optional benefit under Medicaid. Regardless, all Medicaid programs currently provide prescription drug coverage, but questions of which drugs should be covered and what purchasing strategies are the most appropriate for the needs of patients must still be answered. Although a small minority of states still uses coverage policies that limit the number of brand-name prescriptions or the overall number of prescriptions available to Medicaid patients, more than 40 states use some form of evidence-informed preferred drug list to make these choices. Similarly, decisions to cover certain preventive and restorative dental services are often made on the basis of research that demonstrates the cost-effectiveness of these interventions, especially for children and adolescents. Medicaid programs also annually determine which of the new HCPCS Level II codes for DMEPOS will be covered by the program.

Once the benefit design is complete, Medicaid program medical directors and other clinical staff use a wide variety of sources to obtain clinical evidence to inform their decisions on when it is appropriate to pay for a given intervention. Sources include secondary research (including literature reviews and policy analyses) conducted by state staff, consultant organizations, federal agencies, and international organizations such as the Cochrane Collaboration. States do not generally perform primary research through their Medicaid programs, although other public institutions such as medical schools and universities may. A number of states are involved in pay-for-performance and disease management initiatives, both of which draw on clinical research and system reorganization to improve the quality of care and, it is hoped, health outcomes.

Determining which services must be included in insurance coverage can be a controversial and highly political process in which clinical evidence plays an important but not conclusive role. Traditionally, a policy covers broad categories of services, such as physician services, laboratory and

imaging services, and hospitalization. However, within these broad categories of coverage, many questions regarding the appropriateness of specific elements of care remain. These questions are usually resolved by attempting to determine whether or not the element of care is medically necessary. Such a determination includes a consideration of the effectiveness of an intervention in treating a given condition, a comparison of an intervention with other interventions that might be available, and the overall condition of the patient in question. The levels of evidence required to demonstrate medical necessity vary from state to state.

At the highest level, legislative action mandates what must be included in the insurance policies sold within a state. Screening mammograms, organ transplants, hospice care, and even access to off-label uses of medications for cancer patients are examples of services that state regulations often require. Occasionally, mandates such as those requiring access to autologous bone marrow transplants for patients with advanced breast cancer are enacted in the absence of good supporting evidence and are later repealed as better evidence becomes available. Mandates can also require insurers to provide access to certain practitioners. Examples of such mandates include chiropractors, physical therapists, and midwives. Evidence is often helpful in determining the questions of the scope of practice and competency that inevitably arise in these decisions. The rationale for insisting on good evidence of effectiveness before enactment of a mandate continues to grow because mandates almost always increase the cost of premiums, and cost is the primary reason that millions of Americans are unable to obtain health insurance coverage.

Once a service is included in health insurance coverage, regardless of whether its inclusion is due to mandates, market, or other considerations, there are still questions of when it is appropriate to use that service. In this case, states generally require a process for patients to appeal any denial of care based on appropriateness. A qualified third party with no direct interest in the outcome of the decision often conducts these processes. In these cases, statutes or regulations often call for using the best available evidence and include a ranking of evidence based on the commonly used hierarchy, starting with randomized controlled trials as the most desirable and moving to case series and expert opinion as the least helpful. In addition, there is a growing appreciation of the value of systematic reviews of the existing evidence on a given procedure. Treatment guidelines produced by medical specialty organizations connected to treatment for a given condition are also commonly invoked as evidence that can guide whether a given service should be provided to a given individual. It is also recognized that all guidelines are not created equal, and there is movement toward giving evidence-based guidelines primacy over guidelines based on primarily on consensus. NIH, AHRQ, the National Academy of Sciences, CMS, and

FDA are also frequently cited as trusted sources of evidence that can be used to guide decisions.

Initial medical licensure All states set out minimum requirements for the postgraduate training required before an individual is qualified to take one of the nationally recognized licensing examinations. Typically, 1 or 2 years is required for graduates of a medical school in the United States, and international medical graduates are required to obtain 1 to 2 years of additional training before qualifying for the examination. Those seeking licensure as an osteopathic physician are required to pass the Comprehensive Osteopathic Medical Licensing Examination, and those seeking the license of an allopathic physician are required to pass the United States Medical Licensing Examination. State policy determines the number of attempts allowed to pass these examinations. The states do not write these examinations per se but, nonetheless, deem the standards that they set as sufficient for bestowing a medical license. To the extent that these examinations require would-be physicians to understand how to assess and appropriately apply clinical evidence, they are a critical instrument for helping ensure that the many efforts to improve and produce clinical evidence actually succeed in affecting medical practice.

Continuing medical education Most states require physicians to complete a minimum amount of continuing education to maintain an active medical license. Requirements run from 20 to 50 hours per year, and the types of activities qualifying as continuing medical education are set out by state policy. Often, the activities are required to meet criteria established by the American Osteopathic Association or the American Medical Association. In addition, certain subject matter has been specified by a number of states, including HIV/AIDS, pain management (a particularly thorny issue, given the issues surrounding opioid medications), and end-of-life palliative care. Less often, medical errors and risk management are specific subjects that must be covered. Clearly, these requirements provide a fertile ground for improving practicing physicians’ knowledge of emerging evidence and encouraging them to improve their practice by applying the evidence appropriately.

Consumer information Among the states, 47 have established policies that provide the public with easily accessible profiles of the doctors in the states. In some cases, this activity has been mandated by legislation, and in others, the regulating agency has established it on its own motion. Even those states that do not provide profiles make disciplinary actions taken by

the state accessible to the public. Among the states that provide profiles, some include only educational qualifications and license status, whereas others provide a lengthy additional list of items from specialty qualifications to malpractice records, disciplinary proceedings, hospital privileges, and criminal convictions.

Several regulator sector initiatives, some of which are agency specific and some of which are collaborative, may have direct implications for enhancing the evidence that is used to support clinical decision making.

Many of the efforts described here have been launched under FDA’s CPI. They are dependent on the availability of sufficient resources.

FDA is engaged in or is planning a number of initiatives listed together under the headings of reengineering and streamlining the clinical trials process. Collectively, these initiatives are intended to facilitate adoption of new development and regulatory approaches to facilitate a better understanding of product performance (leading to better patient outcomes), hasten the implementation of personalized medicine, increase the quality and quantity of information that can be derived from clinical trials and other data analyses, and ease administrative and other burdens associated with the conduct of complex, multisite studies. For example, FDA and the Duke University Medical Center have launched a collaboration aimed at modernizing the way in which clinical trials are conducted. The collaborative will include broad representation from government, industry, patient advocacy groups, professional societies, and academia. The goal is to work together to develop new standards and identify new methods and technologies that will improve safety, boost the quality of information derived from clinical trials, and make the research process more efficient.³

Personalized medical interventions present significant challenges to a regulatory process that has historically been geared to evaluating interventions with fairly general populations. Increasingly, selecting the correct

intervention for a given patient will require use of a diagnostic to determine whether the patient is likely to respond to a therapy. FDA is very committed to providing guidance for scientists and the regulated industry to facilitate the development of drug and biological products that target specific subsets of a general disease and the concurrent development of a diagnostic needed to identify the subset to be treated. FDA is developing specific recommendations and plans to issue draft recommendations in 2008. Other initiatives discussed below are also relevant to the advancement of more personalized medical interventions. For example, the HHS Secretary’s Personalized Health Care Initiative, in which FDA and CMS are participants, is designed to improve the safety, quality, and effectiveness of health care for every patient in the United States. By using genomics, the identification of genes and how they relate to drug treatment, personalized health care will facilitate the tailoring of medical care to a person’s individual needs. Successfully speeding up insights on individual variation will require the ability to use networked health data, in effect, creating a network of networks to aggregate anonymous healthcare data to help researchers establish patterns and identify genetic definitions of existing diseases.

Biomarkers are measurable characteristics that reflect physiological, pharmacological, or disease processes. Because changes in established biomarkers after a treatment may reflect a clinical response to a product, facilitating the development and qualification of the next generation of biomarkers is crucial to the successful implementation of more personalized medical care. For more personalized interventions, the selection of the patient population to be treated or the dose to be administered may rely on a diagnostic test for a biomarker that predicts whether a patient has responded or the rate of drug metabolism. New biomarkers are also needed to allow the earlier detection of potential toxicity in treated patients. Genomics, proteonomics, and metabolomics hold great promise as source of biomarkers; and FDA is committed to facilitating the application of these sciences to drug development. Several agencies, including FDA and CMS, are involved in the Biomarker Consortium, a broad-based group of interested parties devoted to identifying and qualifying the next generation of biomarkers.

FDA is committed to the implementation of innovative trial designs. The current paradigm—an empirical clinical trial intended to assess patient

improvement, lack of improvement, and the incidence of adverse reactions—often contributes little to understanding of the disease process and the drug’s mechanism of action against that process. Innovative designs, in conjunction with new and better biomarkers, are needed to obtain more and better information. Innovative enrichment designs and adaptive designs, for example, will likely be needed to demonstrate the safety and effectiveness of a targeted drug therapy and the diagnostic that identifies the target population and to establish the clinical utility of the biomarker that the diagnostic tests for. Trials that define and measure variations in individual responses and that seek a correlation with biomarker status are necessary first steps toward providing the evidence base needed for personalized medicine. As new and innovative trial designs are implemented, it will also be important to develop standardized clinical trial designs and outcomes measures tailored to specific diseases or indications. Standardized designs will, among other things, help reduce variation and error and facilitate more informative cross-study analyses, which is important for evidence-based medicine. They will also improve the efficiencies of clinical trials and product development efforts for subsequent products.

Rapid advances in science and technology have increased the complexity of medical products, resulting in increased attention to safety-related issues by the broad healthcare community. In 2004 and 2005, FDA and HHS announced a series of steps intended to address drug safety issues. FDA created its Drug Safety Oversight Board (FDA, 2005) to, among other things, monitor emerging safety information. FDA also asked the IOM to convene a committee to assess the U.S. drug safety system and to make recommendations to improve risk assessment, surveillance, and the safe use of drugs. On September 22, 2006, after extensive interviews with FDA staff and public outreach, the IOM released a report entitled The Future of Drug Safety: Promoting and Protecting the Health of the Public (Institute of Medicine, 2006). The IOM report made substantive recommendations about how FDA could improve its drug safety program and about what actions other parts of government should take to create a more robust and comprehensive system for better ensuring the safe use of medical products. FDA responded to the report, expressing substantial agreement with most of the IOM recommendations and full commitment to strengthening its drug safety program just as rapidly and efficiently as the available resources allow. The response describes a series of initiatives that are among the highest priorities of the FDA commissioner (FDA, 2007b). One FDA goal that is particularly pertinent to the IOM Roundtable on Evidence-Based Medicine is harnessing the power of bioinformatics and other information manage-

ment technologies to enable the sharing, tracking, and analysis of safety and effectiveness information in a secure fashion across systems.

As the U.S. healthcare system moves into the twenty-first century, it is critical to modernize the way that the nation manages medical product safety information. Many activities are already under way to create a rational and systematic nationwide approach to managing the risks of medical product use. It is very important that FDA work with the healthcare community’s public and private sectors—payers, healthcare practitioners, provider organizations, medical product manufacturers, academia, patients, the states, and other agencies in the federal government—to identify potentially serious problems resulting from the use of medical products (e.g., adverse reactions and effects, product use errors, and product quality problems) and get that information to the public as quickly as possible. With that goal in mind, FDA has begun exploring the possibility of working with the broad healthcare community to enhance the ability to identify problems with medical products and provide information to the healthcare community as quickly as possible. In collaboration with other agencies within HHS, the U.S. Department of Defense, and Veterans Affairs, FDA has begun to explore the feasibility of creating a distributed, electronic, national medical product safety system. Such a system ultimately could foster the seamless, timely electronic flow of medical product safety information from electronic databases and surveillance reporting systems through risk identification and analysis processes to healthcare practitioners and patients at the point of care while protecting patient privacy. It would help make information about the safe and effective use of medical products accessible to patients and healthcare professionals in a timely and efficient fashion. Such a system could be assembled through public–private collaborations and consist of existing systems, rather than creating new systems.

FDA, NIH, and the Centers for Disease Control and Preventions (CDC) have been working with stakeholders through Health Level 7 on a national electronic standard for a medical product adverse event report, called the individual case safety report (ICSR).⁴ ICSR will facilitate the timeliness and reduce the cost of reporting, aggregating, and analyzing possible medical product-related adverse events. An ongoing effort within Health Level 7

will harmonize ICSR with other adverse event reporting formats, including the adverse event domain of the Clinical Data Interchange Standards Consortium⁵ standard for adverse event data collected during clinical trials. The development and use of a harmonized adverse event reporting standard across the entire life cycle of a medical product can enhance the ability to integrate all adverse event information collected pre- and postmarketing and can facilitate the development of integrated adverse event repositories, such as the Janus data warehouse currently being developed jointly by FDA and the National Cancer Institute. This should enhance the analytic ability to detect and interpret medical product-related adverse events, regardless of their source. Similar efforts should be pursued in other bioinformation settings.

CMS is seeking a balance in resources in which all participants in the healthcare system can continually learn about the available technologies so that the decisions that are most likely to improve health outcomes can be made. Currently, efforts to develop evidence are focused on getting access to the market. Once access is obtained, far less attention is devoted to ensuring that the evidence used to garner access is verified in clinical practice. When data are collected under pressure to gain rapid access to the market, rarely do those studies include long-term safety and effectiveness head-to-head comparators of a population similar to that of Medicare beneficiaries.

One concept being considered to address this issue is life cycle evidence development (LED). The LED process represents a substantial culture change by way of continuous data acquisition, evaluation, and response to findings. As it is currently envisioned, the concept involves eight domains: identification of healthcare need, proof of concept, safety (both short and long term), effectiveness, comparative effectiveness, quality/appropriateness, and efficacy. The information needed for each domain will vary depending on the current state of information about the technology. At the moment, Medicare’s involvement is primarily limited to effectiveness/comparative effectiveness and quality/appropriateness through Medicare’s CED process and making claims data available to researchers. However, realizing

the goal of LED involves the gathering of data on an ongoing basis by all providers and purchasers. Currently, data are gathered in disparate mechanisms. Sources include healthcare claims, information submitted by clinical trial investigators, manufacturers, professional societies, hospitals, physicians, federal agencies, technology assessments, registries, and other monitored data systems. This data gathering needs to be better coordinated so that current knowledge is available in a standard and open manner. In particular, for LED to succeed, hospitals and other providers of information need systems to coordinate the multiple entries of the same data into various internal databases, clinical trial databases, and other registries. The technology needed to develop common platforms is available, but the lack of national standards for data definitions and for defining the national architecture for such a system are barriers to creating such a platform. In determining effectiveness on a continuing basis, the healthcare system as a whole must determine what questions to ask and answer; agree on consistent definitions, data elements, and collection methods; and develop consistent statistical constructs that are applied across all reporting.

The CMS Chronic Condition Data Warehouse (CCW) is a step forward in the coordination of data collection efforts and analysis across settings. CCW provides researchers with Medicare beneficiary claims and assessment data linked by beneficiary across the continuum of care. In the past, researchers analyzing Medicare data files were required to perform extensive analysis related to beneficiary matching, deduplication, and merging of the files in preparation for their study analysis. With the CCW data, this preliminary linkage work is already accomplished and is delivered as part of the data files sent to researchers.

As noted above, CMS is also working to modify its payment systems to encourage the delivery of high-quality, efficient health care. Specifically, CMS is building incentives into its payment systems. These pay-for-performance programs include the Physician Quality Reporting initiative and the Hospital Value-Based Purchasing Plan. As set out in the Tax Relief and Health Care Act of 2006, CMS’s Physician Quality Reporting initiative will award a 1.5 percent bonus (subject to a cap) to physicians who voluntarily report applicable consensus-based quality measures. In addition, since 2004, hospitals that voluntarily report specified quality measures are entitled to receive the full payment update. This information is also made available for beneficiaries on the CMS website. If the U.S. Congress passes

legislation authorizing CMS to do so, CMS will also implement a hospital pay-for-performance program. Through this program, CMS will seek to align payment policy with the delivery of high-quality, efficient care. One of the core tenets of the program will be an ongoing process for developing, selecting, and modifying measures of quality and efficiency. An equally important element of this program will be continued public reporting of this quality information to beneficiaries.

States are working with each other, not-for-profit organizations, federal partners, and others to expand the use and availability of evidence in clinical and administrative decision making. Many states are in the process of designing and purchasing new Medicaid information systems. This presents an opportunity to enhance the ability of the states to capture data that can be helpful in informing policies and providing providers and patients alike with important clinical and policy information. This effort is a collaboration between states and their federal partners because the federal treasury pays for 90 percent of the cost of the information system. Similarly, CMS has recently awarded Medicaid transformation grants to a number of states. These grants are primarily focused on improving information systems to support clinical care.

Notable initiatives that combine the efforts of states and nonprofit organizations to improve the use of research evidence in policy making include the Milbank Memorial Fund’s program that annually provides travel and tuition for policy makers to attend the Rocky Mountain Workshop on Evidence-Based Medicine and the Center for Health Care Improvement’s initiative that seeks to improve the care of Medicaid patients with multiple comorbidities.

Elsewhere, states have joined together to produce evidence heretofore unavailable. A collaboration of 13 states and the Canadian Agency for Drugs and Other Technologies in Health are producing systematic reviews comparing the efficacy/effectiveness, safety, and effects of classes of drugs on subpopulations. This collaboration has recently expanded its work to include other questions that are important to administering preferred drug lists with the highest clinical integrity. The collaboration, dubbed the Drug Effectiveness Review Project, has completed systematic reviews of 28 classes of drugs as well as numerous updates to the reports on the basis of need dictated by new research. These reports are placed in the public domain.

Eleven other states have also joined together to create a similar collaboration to obtain high-quality evidence to inform other coverage deci-

sions made by Medicaid programs. The subjects reviewed by the Medicaid Evidence-Based Decisions project run a continuum from the appropriate use of various medical procedures, such as bariatric and spinal surgery, to the establishment of which elements of disease management programs are most effective in improving care and saving money and comparison of strategies for obtaining the most value for the money spent in DMEPOS. Similarly, AHRQ has established a learning network for state Medicaid medical directors that allows them to share their evidence, experiences, and policies with each other and to nominate subjects for reviews performed through the Effective Health Care program at AHRQ.

State insurance policy makers in both the legislative and executive branches continue to grapple with the challenges associated with determining which services are appropriate for inclusion in insurance coverage and under what conditions it is appropriate to pay for a covered service. Increasingly, their deliberations are shaped by the inclusion of research evidence. A growing understanding of the nature of good-quality clinical evidence and its strengths and limitations seems to be evident. Some states are even considering the implementation of policies that require vendors to demonstrate through the provision of scientific evidence that new interventions provide improved health outcomes compared with the outcomes obtained with existing treatments. This change could result in a considerable increase in the standards required for payment because many vendors can now provide only the findings of placebo-controlled studies as a basis for a treatment, and such studies often focus only on intermediate outcomes, such as changes in laboratory test results rather than true health outcomes.

State regulators are playing a catalytic role in moving toward the use of a more consistent, relevant, and accessible approach to measuring and communicating information about a physician’s competence throughout his or her career. Conceptualized as the National Alliance for Physician Competence, this approach seeks to integrate more fully the continuum of physician education, training, licensing, and certification. Its purpose is to assist physicians in achieving and maintaining the highest possible level of competence to improve the health and safety of patients and to demonstrate to the public that the medical profession has created processes that ensure that physicians are of high quality and are competent to render the care that they seek. The alliance’s initial work will focus on identifying the appropri-

ate metrics for measuring physician competence and conceptualizing a data management structure that supports the life-long learning of physicians, the need for physicians to fulfill continuing competence requirements, and the need for the public and other elements of the healthcare system to have ready access to information that demonstrates the competence of a physician. Research evidence will play a crucial role in providing an understanding of what makes a doctor a good doctor, how to measure competence, and how best to create the ability within the medical community to use the findings from clinical research to improve care.

A key element of regulating medical practice is creating a concise statement of what constitutes good medical practice that is meaningful to physicians and that resonates with the public. Toward this end, state medical regulators have joined together with colleagues and associates across the medical community to articulate such a standard. Now being circulated for public comment, the document Good Medical Practice—USA sets out six domains of skills, knowledge, and behaviors that define the elements that make up the appropriate and responsible practice of medicine. The domains include medical knowledge, patient care, professionalism, communication, system-based practice, and practice-based learning. The document calls for physicians to have the ability to critically appraise research and to appropriately incorporate research findings into self-assessment and practice improvement in each of these domains. This clear articulation of the objective of medical education and training will help prepare physicians to sort through and use the vast amounts of clinical evidence that will emerge from the efforts of other healthcare institutions.

In addition, a group of leading medical regulators from Canada and the United States are drafting a report on a survey of state and provincial licensing agencies about their current and desired activities for improving the quality of physicians’ practice. The report will include an analysis of feasible new policies that will make the regulatory bodies more effective in improving quality and the cost of implementing such policies. The members of the group, convened by the Milbank Memorial Fund, were chosen by the federation of regulatory bodies in each country.

Any meaningful effort to advance the application of evidenced-based medicine through the development of data that provide better information about therapeutic options and the implementation of those findings in clinical practice requires a coordinated national effort by the range of interested

players. These players include governments and other third-party payers; the medical community at large and the various professional organizations that provide clinical practice guidance; NIH; CDC; VA; FDA; CMS; and the pharmaceutical, biopharmaceutical, and medical device industries. The regulator sector also recognizes that the resources that will be available to be devoted to evidence-based medicine initiatives are likely to be limited and, therefore, need to be used in the most cost-effective way to achieve the stated objectives. A carefully coordinated effort by all players is the only path forward.

Finally, the types of urgent public health questions requiring answers and the types of studies needed to provide those kinds of evidence-based answers (e.g., head-to-head comparisons of a particular form of therapy) are often not clearly within the mandate or financial interest of any one player acting independently. There needs to be, in addition to existing structures like the VA or cancer cooperative groups, additional consortia of investigators who conduct clinical trials who can rapidly be signed up, identify local site managers, and initiate studies. Models for this also exist (e.g., the International Studies of Infarct Survival in Europe and perhaps the Department of Clinical Research Informatics and the Thrombolysis in Myocardial Ischemia study in the United States). It seems at least possible that some health maintenance organizations with record systems that could greatly simplify patient selection and follow-up could become enthusiastic about answering the kinds of questions that might be posed.

The regulator sector believes that there are two initial areas in which immediate collaboration in the regulator sector could be transformative.

The attempt to identify the areas with the greatest need for evidence should be termed the “national problem list.” An entity, for example, a national think tank with broad representation from the healthcare community (including FDA, CMS, VHA, healthcare providers, industry, and others, perhaps under the sponsorship of the IOM) should be created and tasked with coordinating the development, prioritization, and management of the national problem list. This effort can succeed only if it reflects the views of the full range of people with healthcare interests:

• the medical community;

• professional organizations that prepare clinical practice guidelines;

• third-party payers, both private and governmental (CMS);

• government health organizations (CDC, NIH, FDA, VHA, Department of Commerce, AHRQ);

• the pharmaceutical, biopharmaceutical, and medical device industries; and

• patient groups.

Once it has identified the critical evidence needs, this organization would then be responsible for identifying and brokering the conduct of studies or other activities by parties with the capacity to support such studies or activities. In all likelihood, this would require the establishment of a consortium of entities with the ability to conduct the necessary research. For example, these could include entities with particular clinical expertise (e.g., in cardiovascular medicine, oncology, and diabetes treatment), expertise in clinical pharmacology (biomarkers), or technical expertise (e.g., information technology, data mining, and bioinformatics), in addition to the ability to conduct large-scale clinical trials.

The issue of financial support is complex. As acknowledged above, it is unlikely that any private entity will find it in its financial interest to support a particular study (in contrast to the effort as a whole, the success of which is in everyone’s financial interest); and the resources of public entities such as the NIH, VA, and CMS are limited. Thus, there would undoubtedly have to be some financial entity that could collect and distribute funds (e.g., a foundation, consortium, or public–private partnership). The information that could be gleaned from these studies on effectiveness (or the lack thereof) and the best choices in health care could result in significant savings. Savings aside, the gains in the health of the public that could result from the findings of such studies could also be immense.

There are several proposals, in the U.S. Congress and elsewhere, to establish such an entity. An initial first step might be the creation of a coordinating entity to help identify priorities and design strategies for the establishment of an infrastructure to better support controlled studies and gather data on an ongoing basis.

Cooperative work between FDA and the healthcare community—payers, healthcare practitioners, provider organizations, medical product manufacturers, academia, patients, the states, and the federal government—is important to enhance the ability to identify problems with medical products and provide information to the healthcare community as quickly as possible. Agencies within HHS, the U.S. Department of Defense, and VA have just begun to explore the feasibility of creating a distributed, electronic, national medical product safety network. Such a sentinel system would help make information about the safe and effective use of medical products accessible to patients and healthcare professionals in a timely and

efficient fashion. The system could be assembled through public–private collaborations and connect to existing systems. Although this project is in an exploratory phase, it is consistent with Section 905 of the recently enacted Food and Drug Administration Amendments Act, which provides for such a network to enable postmarketing surveillance.

FDA (Food and Drug Administration). 2005. Drug safety oversight board meetings public summaries. http://www.fda.gov/cder/drug/DrugSafety/DSOBmeetings/default.htm (accessed May 12, 2008).

———. 2007a. FDA’s critical path initiative. http://www.fda.gov/oc/initiatives/criticalpath/ (accessed May 12, 2008).

———. 2007b. The future of drug safety—promoting and protecting the health of the public: FDA’s response to the Institute of Medicine’s 2006 report. Washington, DC.

Institute of Medicine. 2006. The future of drug safety: Promoting and protecting the health of the public. Washington, DC: The National Academies Press.

This page intentionally left blank.