|
Curbing The Cardiovascular Disease Epidemic: Aligning Industry, Government, Payers, And Academics
Despite decades of progress in the diagnosis, treatment, and prevention of cardiovascular disease, its prevalence continues to grow in both developed and developing countries. We have constructed a model, the "cycle of quality," which connects the innovation of initial scientific discovery with validated methods of translating research into effective delivery. This model can serve as a basis for evaluating proposed efforts to improve interactions among private and public aspects of health care to accelerate development and appropriate adoption of new treatments, and to achieve greater penetration of effective behavioral therapies and established technologies, resulting in major improvements in cardiovascular health.
AMERICAN SOCIETY HAS WITNESSED a remarkable convergence of effective diagnostic, preventive, and therapeutic approaches to cardiovascular disease (CVD). At the same time, the overall prevalence of CVD and its sequelae are rising as a result of the aging of the population in developed countries, greater longevity in developing countries, and the global epidemics of diabetes and obesity. A global view is especially important, because the greatest increase in prevalence and mortality will occur in countries with developing economies, and much effective CVD prevention and treatment does not require expensive technology, although technological advances have the potential to improve the health of patients with CVD on a global scale.1 We have developed a model, the "cycle of quality," for connecting innovation in the discovery phase of science with measurement of effectiveness and the effective delivery of proven technologies.2 Our thesis is that improvements in rules by which major sectors of the CVD enterprise interact could accelerate the cycle of innovation and adoption, leading to unprecedented improvements in cardiovascular health in both economically developed and developing countries. The U.S. cardiovascular community can be divided into five broad categories: (1) thousands of medical product companies that develop and sell products to treat CVD; (2) government agencies, including the Food and Drug Administration (FDA), the Centers for Medicare and Medicaid Services (CMS), the National Institutes of Health (NIH), the Centers for Disease Control and Prevention (CDC), and the Agency for Healthcare Research and Quality (AHRQ), concerned with improving cardiovascular care as a public mission (although these agencies have different roles—research, public health, health care quality, regulation, delivery—their joint effect is intended to continuously improve the health of the U.S. population); (3) the clinical practice community and its academic centers, which provide basic training and scientific infrastructure and have the responsibility of delivering effective technologies to individuals and populations; (4) payers and employers, concerned with purchasing care for their beneficiaries at a reasonable cost; and (5) patients and their families, who have most at stake in health care policy debates. In our analysis we focus on systemwide impediments to the successful prevention or treatment of CVD, with an attempt to identify areas in which formal and informal arrangements or rules governing interactions among these sectors could accelerate progress in reducing death and disability from CVD, particularly through more-effective public-private partnerships (PPPs).
The global community of CVD experts has increasingly converged on a set of fundamental concepts about efforts needed to create systematic approaches to translating knowledge across the continuum from discovery science to public health intervention. The cycle (Exhibit 1  ) begins with the discovery of fundamental biological, physical, and social constructs. Once a discovery is made, it undergoes a development cycle including extensive preclinical applied research before it can be developed as a treatment with plausible human benefit. Evidence is then gathered in human experiments, and assessments are made about the intervention’s value; these evaluations continue after the treatment is clinically available.
Clinical practice guidelines have evolved as the prevalent mode for defining professional standards for disease prevention and management. When findings from these guidelines reach a level of definitive evidence in a quantifiable context, the guidelines can be translated into performance measures that are then used to quantify the degree to which best practices are followed by practitioners, practices, hospitals, health systems, communities, or countries.
Massive funding of the NIH by U.S. taxpayers has been a major stimulus for discovery, although the medical products industry has recently increased its investment in fundamental research and development (R&D) to similar levels.3 Studies of successful discovery science tend to conclude that most breakthrough findings are serendipitous, in that an individual, stimulated by curiosity rather than directed research, develops an insight into a previously unknown biological or physical mechanism.4 Fundamentally, then, discovery research is not amenable to plan-based, specifically anticipated results. Accordingly, there is no substitute for continued fueling of enterprises in which scientists are free to pursue knowledge based on their curiosity rather than external direction, but remain accountable for productivity and academic excellence based on scientific peer review. However, every phase of the cycle beyond discovery research is amenable to improvement in efficiency and translation into a reduced burden of CVD.
Many experts in technological development have pointed to gross inefficiencies in bench-to-bedside translation as a major problem in adapting scientific advances to the clinic, a problem called the "first translational block."5 Despite dramatic advances in fundamental biology and engineering technology, as well as increasing investment in R&D, the number of new molecules achieving FDA approval has declined, and many discoveries languish in laboratories for years before emerging.6 Scientists are often unfamiliar with regulatory processes for applied research in technology development and the need for team processes for clinical development, and they often lack adequate resources to sustain these efforts. Promising approaches often fail because of "off-target" effects—the unintended effects of drugs on pathways for which they were not designed, leading to unanticipated toxicity. Preclinical systems and animal models do not accurately predict the balance of effectiveness and safety that should be understood when medical interventions are used in practice. Accordingly, an enormous number of molecules and concepts must be tested in humans to find a few that yield a net balance of benefit over harm. In response to these systemic shortcomings, a major area for PPP has been identified in the FDA-sponsored Critical Path Initiative (CPI).7 This initiative has focused on a body of knowledge termed "precompetitive"—that is, concerned with aspects of development where keeping knowledge proprietary confers no advantage to a company, while conversely, failing to share knowledge devalues the entire industry; it includes preclinical toxicology findings and biomarkers. The goal is to develop a common knowledge base about effective approaches to preclinical translational research that transcends any particular corporation, university, or government agency. Thousands of compounds have been evaluated in systems involving cell cultures or whole-animal toxicology. Researchers must then deduce whether toxicologic findings predict safe human applications. Unfortunately, when a compound causes toxicity, these findings are often sequestered in internal files of pharmaceutical and device companies or the FDA, which does not have the independent authority or means to make such findings public. Accordingly, common understanding of the significance of preclinical findings in predicting future outcomes is severely impaired; a complete record of failures as well as successes would greatly improve translational researchers’ ability to understand the relevance of their findings to human health. The CPI thus proposes creating a database of preclinical findings, to develop predictive standards that will be available as a publicly accessible source for predictive toxicology. This includes the creation of a nonprofit center to house databases and provide a mechanism for funding of research projects with an aim to using data in the public interest, with financial support from both private business and philanthropic foundations. To illustrate how the CPI PPP could provide major benefit to CVD treatment, let us consider the electrocardiogram, a common, inexpensive test. Many drugs effective in treating CVD affect sodium and potassium channels in the heart muscle. When the permeability of these channels is altered, a change occurs in the electrocardiogram.8 This change—prolongation of the QT interval—is associated with a finite risk of a lethal cardiac arrhythmia called torsade de pointes. QT interval prolongation is measurable when almost all older and atypical antipsychotic medications are given to patients; many other examples exist.9 This phenomenology has led the FDA to require detailed measurements of QT intervals in electrocardiograms of test subjects for most drugs under development.10 Drugs that cause no QT interval prolongation do not lead to torsade de pointes, but drugs that cause modest QT interval prolongation will probably lead to some number of potentially lethal arrhythmic events, although the exact relationship between QT interval prolongation and risk of torsade de pointes is not well established. The FDA is working with an academic and professional society consortium to develop a data warehouse that will enable stakeholders to access data that are not identified by drug name, to examine relationships between QT interval and biological and clinical phenomena.11 By taking data that before were proprietary and mostly unpublished and making them available by applying a set of rules, companies will benefit from a better understanding of how to use QT measurement in making decisions about drug development, resulting in quicker development of safer drugs. This type of PPP takes data held by a public organization (the FDA) and provides access to academic medicine and industry to improve the common body of knowledge. The NIH has launched a bold effort to change the way that translation occurs from discovery through human studies.12 The recognition that the nature of research had shifted to an increasing reliance on "big science" and team research has led to the NIH Roadmap program.13 A major element of this effort is "transformation" of academic medical centers into structures that facilitate translational medicine across the spectrum of health care. By providing a home for clinical and translational researchers, along with academic support and means of advancement and promotion, it is hoped that a stable base of expertise will evolve. Although this effort represents less than 2 percent of the NIH annual budget, it is a major public investment by any criteria and is specifically intended to improve the speed at which basic discoveries are translated into clinical practice. As part of the effort, academic centers are expected to develop creative approaches to partnerships with industry that make the most of their combined capabilities.
Complete versus incomplete data. Once a potential treatment has passed initial preclinical screening and enters the regulatory process, it must be developed so that it can be deployed in the health care system and evaluated to determine the balance of risk and benefit deriving from its use. Some have argued that rules in place for regulatory agencies are out of step with the state of scientific knowledge regarding evaluation of technology.14 In particular, we now know that the risk/benefit balance of a technology usually cannot be depicted accurately using biomarkers or surrogate endpoints and that limited populations entered into clinical trials performed for the purpose of regulatory approval for marketing seldom provide information adequate for estimating risks and benefits for "real-world" populations.15 However, requiring complete data for such purposes prior to marketing would create financial hurdles that could drastically reduce investment in new therapies. A more holistic approach might allow early marketing of technologies with labeling that reflects the actual state of uncertainty at that point in development, while definitive broad and longer-term studies are done.16 This could be particularly important in creating incentives for development of new treatments for chronic diseases, such as asymptomatic atherosclerosis, an area where it may take decades to truly understand the risk/benefit balance. Improving the trial system. The systems used to conduct these necessary human experiments likewise need major revamping. The current inefficient clinical trials enterprise lacks common standards and has minimal infrastructure in academe for developing and maintaining an intellectual discipline for the enterprise. In most cases, each study is designed, investigators and research sites are recruited, a case record is constructed, and after extensive training of research staff in trial specifics, patients are recruited. Enormous funds are then spent monitoring study conduct. These high costs derive from a failure to create a systems approach to quality based on sampling methodology; instead, large amounts of money are spent auditing almost all study data.17 When the trial is completed, the research structure is disassembled, and the next study starts the labor-intensive process again. These inefficiencies are exacerbated by a complex series of well-intentioned regulations that add major costs and delays to clinical trials, particularly in the United States.18 We believe that assessment of therapeutics could be dramatically accelerated through relatively modest public investment in several key areas. First, key federal agencies could lead a systematic approach to data standards and nomenclature for cardiovascular medicine. By tapping the expertise of the professional community, the FDA, NIH, and CMS could develop approaches that would be interoperable for both publicly and industry-funded research. These same standards could be used for billing and quality measurement in medical practice, thus allowing use of the same data collection for clinical practice and clinical research. Standardization would be driven by requirements that cardiovascular practitioners conform to CMS standards; medical products companies conform to FDA standards; and academic researchers conform to NIH, CDC, and AHRQ standards. This publicly driven approach alone could reduce clinical trial costs to both public and private sectors by up to 50 percent, not only by cutting data collection costs but also by reducing the need for idiosyncratic training and monitoring for each trial.19 Second, these same agencies could develop infrastructure that links clinical practice with human research in a manner that would drastically reduce costs while also improving the relevance of research results to the problem of reducing mortality and morbidity from CVD. An increasing portion of U.S. cardiovascular specialty practice is captured in registries developed to measure quality that are maintained by the Society for Thoracic Surgeons, the American College of Cardiology (ACC), and the American Heart Association (AHA). Using these registries as a foundation for clinical trials would ensure that generalizability of populations enrolled in trials could be assessed and that uptake of treatment strategies could be measured in what approaches "real time." The CMS recently reinforced this concept by requiring that all implantable cardioverter-defibrillators (ICDs) placed in the United States and paid for by the CMS have a core data set collected by the ACC’s National Cardiovascular Data Registry (NCDR).20 Appropriate support for the embedding of trials in disease registries by the National Heart, Lung, and Blood Institute and other federal agencies needs to be further developed. As electronic health records (EHRs) are developed, they should be integrated with CVD-specific registries as mentioned above. By developing interoperable standards, EHRs across health care systems will be able to populate disease registries and clinical trial data fields without extra effort by practitioners. This will require conversion of EHRs from instruments designed purely for health care delivery into instruments that are capable of serving research and quality improvement purposes. Progress toward this goal is being made by groups such as the HMO Research Network and AHRQ’s ACTION Network, which link integrated health systems interested in improving quality of care.21 The creation of a national infrastructure for research has been envisioned in the NIH Roadmap, and its successful implementation is expected to lead to NECTAR, a broad national "network of networks" that could dramatically reduce the cost of evaluating therapeutic approaches to CVD while also providing context for rapid implementation. The public portion of this envisioned effort consists of creating standards and providing research infrastructure funding to encourage high-throughput use by the private sector. Such interoperable networks could finally provide the ability to assess cross-disciplinary issues such as cardiovascular toxicity of arthritis drugs.22 Honing the regulations. Given better infrastructure, serious effort must be expended to address the efficacy of regulations aimed at ensuring research integrity and protection of human subjects. The concept of "learning health systems," which use learning through aggregation of data in practice, is a key approach advocated by the Institute of Medicine (IOM) and is critically dependent on an effective and financially viable regulatory system.23 As multiple layers of regulation have been added to research systems, little empirical investigation has been done to determine which regulations are truly helpful.24 Streamlining research while improving its integrity should be a common goal, because as ineffective regulations are promulgated, the costs of gathering evidence increase, the industry loses efficiency, research flows to other markets, and patients ultimately suffer. The setting of priorities for clinical research should be reexamined, so that human studies are designed to answer questions of greatest importance to patients and practitioners. For practical purposes, human research done to develop medical products now focuses on efficacy and safety in selected groups that might not reflect broader populations. Previously, it was accepted that physicians would "figure out" how to use therapies based on practical experience; we now know that this approach is suboptimal. Additional efforts should address broader public health needs of pragmatic use of therapies in the intact health system environment. However, the NIH provides little funding for such efforts, AHRQ has a small budget, and the CMS is prevented from focusing on research because its primary mission is to fund medical care. The result is that research on medical outcomes is often not done, and assessment of therapies for payment purposes remains a difficult interaction that often becomes more a legal process than scientific weighing of evidence. Different solutions to this funding dilemma have been suggested. All involve some blend of public and private funding, but none have yet been implemented.25 The most attractive proposal from our perspective is the development of a "national problem list" of issues to be addressed in clinical trials. This effort should be a public process in which all sectors participate with a common pool of money created between government and health systems to ensure that critical issues are addressed. One example of a PPP that provides a model for this activity is the Centers for Education and Research in Therapeutics (CERTs), which brings together the FDA, AHRQ, NIH, academe, and industry to perform research and provide education in areas not otherwise funded.26
Creating effective mechanisms to translate clinical trial results into reduced CVD death and disability represents the second translational block. In parallel with the U.S. Preventive Services Task Force, the cardiovascular professional community has been a major purveyor of clinical practice guidelines. A strong alliance has led to CVD guidelines’ being developed jointly by the key U.S. entities, the ACC and the AHA.27 Their combined efforts include specific systematic approaches to coordinating guidelines with the European Society of Cardiology. As experience is gained in developing these guidelines, it becomes clear that processes for devising them should be transparent and that conflicts of interest should be acknowledged and balanced.28 The U.S. government has created a useful partnership with academe and professional societies through the Evidence-based Practice Centers (EPCs) funded by AHRQ. EPCs provide unbiased, publicly funded group expertise that can be tapped by professional societies for analysis of the base of evidence in preparation for developing clinical practice guidelines. Professional groups can then provide clinical oversight to synthesize guidelines, using objective data as the cornerstone.
Performance measures extend the concept of clinical practice guidelines into quantifiable opportunities for action by individuals, groups, and systems. The process by which performance measures are formalized constitutes a key interaction of public and private sectors. Expertise in clinical measurement is produced by academic and health system experts, but the clinical practice community and payers must ensure that measures are feasible and actually improve outcomes. Developing appropriate performance measures to be used in "pay-for-performance" (P4P) systems takes time, and methods are lacking for rapidly updating these measures when evidence invalidates a previous measure because of a new treatment or suggests that a segment of population does not benefit from the treatment. Given the power of performance measures to change clinical practice behavior, particularly when linked to payment, the degree of vetting required across clinical and payer communities is enormous, resulting in major delays. However, expediting such efforts also carries risk, because great damage can result if clinicians and health systems hone delivery of treatments that, in retrospect, are seen to result in previously unquantified harm exceeding the benefit in the population treated. Recent revelations about hormone replacement therapy (HRT) serve as a cautionary example: Prior to trials demonstrating excess harm compared with benefit, HRT was a high-level guideline recommendation; had it been adopted as a performance measure linked to payment, considerable harm could have resulted. The issue of using process versus outcomes in key performance measures needs further study. When effective therapies such as aspirin or statins for patients with known CVD are identified, measuring whether or not practitioners prescribe the treatment for appropriate patients is relatively uncontroversial. However, since outpatients have the option of not taking a treatment, holding practitioners or systems accountable for achieving particular lipid levels becomes much more controversial, because measures to adjust for differences in patient populations are inadequate. Government needs to increase research funding in this area, and academe needs to engage in research to develop better outcome modeling, but this type of pragmatic research requires constant input from all segments of the health care enterprise.
Ultimately, evidence amassed through the "cycle of quality" comes to naught if effective strategies are not implemented. CVD provides an interesting paradigm because of the relatively advanced state of knowledge from empirical clinical trials and epidemiological studies. As discussed above, some have estimated that major cardiovascular event rates could be reduced by 80 percent simply by acting on known risk factors with therapies already shown to be effective.29 Although many consider these estimates to be overly optimistic, there is no doubt that we are falling far short of achievable performance in health systems around the world today. Substantial literature has emerged to describe successful approaches to implementing diagnostic and treatment strategies known to be effective, but the lag between discovery of an effective approach and its broad implementation remains unacceptably long.30 The P4P effort is a critical ongoing experiment in improving implementation of treatment strategies.31 It has long been recognized that the behavior of individual practitioners and systems is greatly governed by the payment received.32 Accordingly, as definitive performance measures have been developed, payers have gravitated toward a strategy of adjusting payment for medical services based on measured performance. The interaction of government and private payers will be crucial in determining ultimate rates of adoption of approaches designed to stimulate improved performance through measurement and financial reward. To the extent that core performance measures can be adopted uniformly across the spectrum of payers, individual practitioners and groups will be able to implement delivery schemes more effectively. Adopting fragmented performance measures among different payers risks creating further disarray in delivery systems and worsening inequities in access to basic, definitive prevention and treatment. Much progress has already been made in uniting the ACC and AHA with government and private payers to work on common performance measures.33
As we enter an age of medical informatics in which primary electronic data capture predominates, measurement systems must weave together broad and specific disease registries with EHRs. These EHRs must be designed to facilitate effective individual health care transactions but will also eventually serve as a rich tool for research and system quality (as they already do in some advanced systems).34 The intersection of professional responsibility for maintaining a knowledge base, corporate responsibility for ensuring that products are used appropriately, and payer responsibility for protecting patients by requiring adherence to appropriate interventions demands more effective PPP approaches to continuing education for health care practitioners. Ultimately, data repositories from disease registries and EHRs shared by government agencies, health systems, and academic centers should serve as the substrate for continuing education of health professionals. Patients and families. Patient involvement is critical in every phase of the quality cycle, including determining research priorities, assessing effects of technology, and determining criteria for quality and should be encouraged. Knowledge bases created through the phases of the quality cycle will be increasingly adapted to provide direct access to patients who wish for greater involvement in understanding their treatment options. Global efforts to provide public access to clinical trials registries, such as the NIH’s ClinicalTrials.gov, and to AHRQ’s DEcIDE Network are interesting examples.35 Behavioral interventions. Exercising, limiting caloric intake, refraining from smoking, and adhering to medications and specific advice, such as dietary sodium reduction for hypertension, have major effects on CVD, yet our methods for achieving these behavioral goals still fall short. We believe that every element of the quality cycle pertains to behavioral interventions as much as to drugs and devices. The major difference is that research funding burdens fall on government, except to the extent that regulations can provide advantages to corporations that promulgate healthy lifestyle choices. The debate over whether emphasis on "low-fat" diets encouraged the obesity epidemic by the unfortunate switch to high-carbohydrate diets illustrates the need for evidence in the behavioral arena, just as with medical technology.36 International implications. Although debates rage about use of expensive technologies in the United States and other developed countries, the largest new toll of death and disability from CVD will occur in developing countries. An important focus for the international community should be using the systems outlined above to identify strategies that achieve greatest impact at lowest cost, with the intent of providing these treatments to as many people as possible. The example of wireless technology demonstrates the potential for new technological development to reduce costs and increase health care systems’ ability to deliver common treatments. Another example is the concept of the "polypill," in which multiple generically available drugs would be packaged together in a single pill.37 Successful development of this concept into a form that can be evaluated in clinical trials will require an approach to PPP that goes beyond current measures. DESPITE EXTRAORDINARY REDUCTIONS in age-specific death and disability from CVD, it remains a global problem and demands new approaches to societal solutions. At every phase of the cycle of quality, limitations to progress in developing effective treatments, assessing them through development of evidence, and using them effectively have been identified. In the United States, potential changes in the configuration of relationships among government agencies, professional organizations, payers, and patients and their advocates have been discussed, and modern informatics can unite elements of the quality cycle in a manner that serves all the constituencies needed to accelerate progress. This could establish a model for other countries as they design policies for addressing this global epidemic.
The authors are from the Duke Translational Medicine Institute (DTMI) and the Duke Clinical Research Institute (DCRI) in Durham, North Carolina. Robert Califf (calif001{at}mc.duke.edu) is the DTMI director. Robert Harrington is the DCRI director. Leanne Madre is program director, Centers for Education and Research on Therapeutics (CERTs) Coordinating Center, DCRI. Eric Peterson is director of Cardiovascular Outcomes Research and Quality, DCRI. Deborah Roth is the DCRI chief operating officer. Kevin Schulman is director of the Center for Clinical and Genetic Economics, DCRI. The authors are affiliated with the Duke Translational Medicine Institute and the Duke Clinical Research Institute, which receive research funding from, and/or have contractual relations with, companies that manufacture drugs and medical devices. Califf, Harrington, and Peterson also receive individual research support from various drug and medical device companies; this funding is administered through Duke University programs.
This article has been cited by other articles:
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Top
A Conceptual Model: The...
Discovery Science
