PRIORITIZING OF CLINICAL RESEARCH QUESTIONS

Fewer than half of all the medical treatments delivered today are supported by evidence (IOM, 2007), yet the United States lacks a clear prioritization of the gaps in medical evidence and an allocation of clinical research resources to efficiently and effectively fill these evidence gaps. The federal government, industry, academic institutions, patient advocacy organizations, voluntary health organizations, and payers each have incentives to develop research questions that suit their unique interests. The value of a particular research effort is judged by stakeholders according to their own cost–benefit calculation. Reflecting the diversity of stakeholder value judgments, and in the absence of a broad national agenda, clinical trials are conducted in a “one-off,” narrowly focused fashion.

Because clinical trials are necessary to obtain regulatory approval in the United States, they are a high priority to companies. It was noted by a number of workshop participants that the prioritization of clinical research questions by companies seeking regulatory approval is distinctly different from the priorities of society in general, which may prioritize the comparison of two commonly used therapies. This divergence between the priorities of society and industry is notable as the nation discusses how to address the current gaps in clinical research and medical decision making.

As an example, in investigator-initiated research, academic investigators seek federal funding (primarily from the National Institutes of Health [NIH]) to conduct research they deem important to advancing science and/or medical practice. But James McNulty, Vice President of Peer Support for the Depression and Bipolar Support Alliance (DBSA), believes the NIH peer review process for research grants is inherently conservative and fails to reward innovative research into areas about which little is known. McNulty believes this conservative approach has contributed to serious gaps in knowledge in the area of mental health, specifically in schizophrenia, depression, and bipolar disorder. In terms of formulating relevant research hypotheses, the U.S. Department of Veterans Affairs (VA) was cited as one example of a health system that successfully engages practicing physicians in noting potential research questions that arise in the day-to-day care of patients. The VA Cooperative Studies Program works to ultimately take physicians’ questions into the clinical trial setting.

Industry-sponsored trials are conducted largely to gain U.S. Food and Drug Administration (FDA) approval to market a new drug or a previously approved drug for a new indication. Preapproval trials include a simple protocol (i.e., ask a limited number of questions) and test a drug in a highly selected patient group designed to provide the most robust evidence on the drug’s benefits and risks. Conversely, the federal government conducts large clinical trials to answer medical questions unrelated to gaining regulatory approval for a new drug or therapy. These studies can involve a wide range of patients and seek to answer a number of relevant clinical questions at once. Several presenters in the diabetes session of the workshop suggested that government-funded clinical trials for diabetes would not be conducted by industry or other sectors. New therapies for type 1 diabetes are often of limited interest to pharmaceutical companies because of the small patient population, whereas drugs for the exponentially larger type 2 diabetes population are avidly pursued.

The beginnings of a coordinated prioritization of research needs can be seen in the recent increased interest in comparative effectiveness research (CER). To enhance the ability of clinical research to generate knowledge that can better inform clinical practice, Congress included in the American Recovery and Reinvestment Act (ARRA) of 2009 an allocation of $1.1 billion for federal agencies (the Agency for Healthcare Research and Quality [AHRQ], NIH, and the Department of Health and Human Services [HHS]) to jumpstart the national CER effort. CER seeks to identify what works for which patients under what circumstances, providing evidence about the costs and benefits of different medical options. One-third of ARRA funds ($400 million) were designated as discretionary spending by the Secretary of HHS to accelerate CER efforts. The Institute of Medicine (IOM) was tasked with recommending national CER priorities to be supported with these discretionary funds and to guide the nation’s creation of a long-term, sustainable national CER enterprise.² Recently enacted health care reform legislation (Patient Protection and Affordable Care Act passed in March 2010) created the Patient-Centered Outcomes Research Institute (PCORI)—a nonprofit institution positioned outside the federal government to define and execute comparative effectiveness research methods.

Several speakers and workshop participants raised questions about the ability of the current clinical trials system, which is already showing signs of strain, to absorb a substantial amount of the anticipated CER studies. Many voiced concern regarding the overall organization of clinical research in the United States: how it is prioritized, where it is conducted, who oversees it, how it is funded, who participates, and who staffs it. Presenters and participants also described the diminished capacity of the current clinical trials system. These observations, and proposed solutions, informed the discussion over the course of the 2-day workshop.

THE DIVIDE BETWEEN CLINICAL RESEARCH AND CLINICAL PRACTICE

Janet Woodcock, Director of the FDA’s Center for Drug Evaluation and Research (CDER), identified bridging the divide between research and the clinical practice of medicine as one of the most critical needs facing the clinical research enterprise today. The limited involvement of community physicians in clinical research reduces physician referrals of patients to clinical research studies, as well as the total number of investigators available to conduct the research (see the discussion of narrow incentives for physician participation in clinical trials below). Furthermore, the findings of research conducted in academic medical centers rather than in community settings are less likely to be adopted by physicians in their daily practice. The poor rate of adoption of effective clinical practices is reflected in one study that examined adherence to 439 indicators of health care quality for 30 acute and chronic conditions and preventive care. Results indicated that American adults receive on average only 54.9 percent of recommended care (McGlynn et al., 2003).

Woodcock stressed that, to generate relevant research based in clinical practice, community practitioners must be actively involved in the clinical trial process. She suggested it is not surprising that the uptake of evidence-based practices is slow when practitioners are not engaged in the research that supports the changes. In many instances, the characteristics of the study population, their comorbidities and therapeutic regimens, and the setting and conditions under which the trial is conducted bear little resemblance to typical community practice. Indeed, the outcomes are often quite different as well. It is little wonder that community physicians may be hesitant to modify their treatment practices to reflect clinical findings developed in this manner. According to Woodcock, the divergence between physicians conducting research and those in community practice is one of the greatest barriers to successfully translating study results into clinical practice. She argued that, to develop a truly learning health care system capable of self-evaluation and improvement, the currently separate systems of clinical research and practice must converge.

GLOBALIZATION OF CLINICAL TRIALS

The increasing trend toward conducting clinical trials outside the United States is an important consideration in discussing ways to improve the efficiency of trials. The number of patients enrolled in clinical trials is decreasing in the United States and increasing abroad. According to Woodcock, when development programs are conducted entirely outside the United States, the FDA questions the extent to which the results can be translated to U.S. clinical practice. The applicability of foreign trials results depends on the disease being studied and the state of current clinical practice in that area.

Califf suggested that the difficulties inherent in conducting clinical trials in the United States have contributed to the relative decline in U.S. clinical trials described in Chapter 2. Citing a recent paper that he coauthored, he noted that one-third of phase III trials for the 20 largest U.S. pharmaceutical companies are being conducted solely outside the United States (Glickman et al., 2009). For these same firms and studies, a majority of study sites (13,521 of 24,206) are abroad (Glickman et al., 2009). Califf stated that the situation is the same across study sponsors—NIH, industry, and academia all look to conduct trials internationally.

Califf suggested that globalization is a positive trend overall, one in which he and his home organization, the Duke Clinical Research Institute (DCRI), are engaged. However, the current situation in which clinical research is being sent abroad just to get trials completed is unsustainable. One reason for this situation is that clinical trials in a number of other countries cost less than they currently do in the United States (Table 3-1) (see also the discussion of costs below). If a large outcome trial requires enrolling tens of thousands of patients, for example, selecting trial sites in Russia or India instead of the United States can result in hundreds of millions of dollars in savings. The overall cost associated with gathering the necessary resources to conduct a clinical trial is an important factor in the choice of a trial site. For instance, physician salaries in a number of countries are lower than in the United States. In these countries, the charges to clinical trial sponsors for conducting a clinical trial with physician involvement are lower than they would be in the United States. Some also argue that clinical trials conducted outside of the United States are of higher quality because of better adherence to trial protocols and better patient follow-up.

TABLE 3-1

Global Research Costs: Relative Cost Indexes of Payments to Clinical Trial Sites.

THE COST OF CLINICAL TRIALS

Clinical trial costs can vary widely depending on the number of patients being sought, the number and location of research sites, the complexity of the trial protocol, and the reimbursement provided to investigators. The total cost can reach $300–$600 million to implement, conduct, and monitor a large, multicenter trial to completion. Table 3-2 outlines the various costs of an exemplar large, global clinical trial, which in this case add up to about $300 million.

TABLE 3-2

Breakdown of the Costs for a Large, Global Clinical Trial (14,000 patients, 300 sites).

Christopher Cannon, senior investigator in the Thrombolysis in Myocardial Infarction (TIMI) Study Group, stated that two clinical trials on which he is working cost a total of $600 million. To put this cost in perspective, it represents approximately half of the $1.1 billion allocated for comparative effectiveness research in the American Recovery and Reinvestment Act of 2009. According to Cannon, the exorbitant cost of clinical trials today points to the need to move toward simpler large trials that would study a broader population, include less data, and cost less overall.

The federal government funds a large portion of clinical research in the United States, primarily through NIH. In some cases, it has been estimated that NIH institutes pay research sites 20–40 percent less than the actual cost of conducting trials. Michael Lauer, Director of the Division of Cardiovascular Diseases, National Heart, Lung, and Blood Institute (NHLBI), National Institutes of Health, responded that to remedy the issue of appropriate NIH payments to research sites, the solution will likely involve a combination of increasing the amount of money paid by NIH to sites and decreasing the charges associated with conducting the research. Lauer further explained that because NIH’s funding is relatively flat, if research site payments are increased, an equivalent decrease in funding in other areas will be necessary. Given this zero-sum calculation, it will be politically difficult to increase payments for research sites. Lauer believes that simplifying trials could be most effective in reducing their cost. He suggested that good science comes from high-quality observations that are followed by focused experiments to test these observations. The trials that have had the greatest impact on clinical decision making and patient care have been simple (e.g., uncomplicated study protocols, short case report forms). Thus, if the research community could keep trials simple and large enough to answer the study question(s), costs could decrease, while the impact and relevance of the results would increase.

Workshop participants also discussed the inequality of NIH payments to research sites across the various NIH institutes. This variation has created a scenario in which some institutes that pay research sites more are seen as the “haves,” while those that pay less are seen as the “have-nots.” Judith Fradkin, Director of the Division of Diabetes, Endocrinology, and Metabolic Diseases in the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) at the National Institutes of Health, noted that inconsistency across Clinical and Translational Science Awards (CTSA) institutions in the level of clinical trial support they provide makes it difficult for NIH to determine how its payments to research sites should be adjusted to take CTSA support into account.

NARROW INCENTIVES FOR PHYSICIAN PARTICIPATION IN CLINICAL RESEARCH

As discussed earlier, the current clinical research enterprise in the United States is largely separate from traditional clinical practice. In part because the United States does not have a nationalized health care system in which services are provided to all citizens through government-funded providers, clinical research takes place in various types of sites, frequently outside of the community-based, primary practice setting where most patients receive care. Moreover, as noted above, private practice physicians have disincentives to refer their patients to clinical trials. The fewer physicians are involved in developing and implementing clinical trials, the less scientific the practice of medicine will be. A number of workshop attendees suggested that a mechanism to adequately compensate physicians for referring patients to clinical trials could improve recruitment rates of U.S. patients. Making it easier for community-based physicians to participate actively in clinical trials could also have a positive effect on patient recruitment; enhance the engagement of the community in important research; and increase the chances that physicians will change their practice behavior based on research results they were involved in generating, thereby strengthening the trend toward evidence-based medicine in the United States. Workshop participants also suggested that, to encourage physician participation in clinical trials, the study questions and protocol should be designed in the context of clinical practice—that is, the procedures required by a trial protocol should be easily incorporated into practice.

It is also important to consider that the research questions clinical trials seek to answer reflect the incentives and interests of those developing the questions. In this respect, the capability of the health care system to act on trial results is part of the clinical research decision making process. For instance, Amir Kalali, Vice President, Medical and Scientific Services, and Global Therapeutic Team Leader CNS (central nervous system), at Quintiles Inc., explained that his company ran the two largest clinical trials testing the combination of psychotherapy and medication to treat depression. Despite scientific evidence for the benefits of psychotherapy, it has seen limited uptake. According to Kalali, this is because patients have limited access to psychotherapy as a medical treatment in the United States. Thus, the capability of the health care system to implement or act on research findings can be an important consideration in conducting clinical trials to test alternative treatments for a condition.

SHRINKING CLINICAL RESEARCH WORKFORCE

Research involving human subjects has become an increasingly complex environment in which to work and be successful. Thus, it is not surprising that, as noted in Chapter 1, the clinical investigator workforce is plagued by high turnover. Clifford Lane, Clinical Director, National Institute of Allergy and Infectious Diseases (NIAID), shared data from NIAID revealing the trend that fewer professionals are entering the research field than in the past. As Figure 3-1 indicates, the number of tenure-track principal investigators conducting research within the NIAID/NIH intramural program decreased from 74 in 2003 to 42 in 2008. While there was a slight increase to 53 tenure-track investigators in 2009, this number is still well below that in 2003.

FIGURE 3-1

While the number of senior NIAID/NIH tenured investigators is relatively stable, the number of NIAID/NIH tenure-track principal investigators is decreasing. SOURCE: Lane, 2009. Courtesy of John Gallin, 2009 (unpublished).

The majority of phase III clinical trials are conducted by extramural researchers. However, trends in intramural NIH programs add to our general understanding of the issues and challenges facing investigators today. Lane commented that the overall decrease in intramural investigators is due in part to the fact that more researchers are turning to laboratory work because publishing results from this work is easier, and the difficulties of getting a clinical trial protocol approved can be avoided. Accordingly, NIAID has created a protocol development program to decrease the burden of regulatory and administrative requirements and optimize the use of existing clinical research tools. The program’s goal is to allow investigators to focus on their work as clinical scientists rather than having to serve as operations managers of a complex regulatory process. When investigators provide a robust scientific idea with a strong hypothesis, appropriate endpoints, and a sound study design, NIAID’s Office of the Clinical Director helps them navigate such regulatory issues as ethics review, technology transfer, safety concerns, and interactions with Institutional Review Boards (IRBs) and the FDA. This support can help investigators implement a trial successfully. In addition to principal investigators, multidisciplinary support staffs are necessary to complete a clinical trial successfully. Biostatisticians, epidemiologists, laboratory technicians, and administrative support personnel are just a few of the types of staff needed.

The importance of involving knowledgeable staff throughout a study was highlighted in a discussion of cardiovascular and depression clinical trials. Califf referred to the crucial role of a well-trained, intelligent data monitoring committee tasked with evaluating interim trial results. Data fluctuations revealed by interim trial monitoring require analysis but do not always indicate that a trial should be discontinued. For instance, when death is the primary outcome of a trial, data fluctuations may indicate an adverse effect on mortality of the treatment being studied or a regular clinical occurrence unrelated to the study drug. Califf noted that if it were not for a particularly well-informed data monitoring committee, the ISIS-2 (Second International Study of Infarct Survival) trial would have been most recently discontinued at 5,000 patients, with aspirin showing an adverse effect on mortality. William Potter, most recently Vice President of Translational Neuroscience, Neuroscience, Merck Research Labs, Merck & Co., Inc., indicated that in the depression studies in which he has been involved, interim data that indicate a possible adverse effect usually result in a trial’s being discontinued.

Because clinical trials are conducted in an ad hoc fashion, and study personnel of varying professional quality are recruited and trained anew at each site, inconsistencies in trial execution across sites are not unusual. Woodcock explained that the failure to execute a clinical trial successfully is often attributable in part to the fact that ensuring proper execution of a single trial is no one’s full-time job. The core activities of a clinical trial are largely supplemental responsibilities assigned to a variety of staff in addition to their full-time work.

Califf noted that clinical investigators often are unsupported by their academic institutions and are left largely to their own devices to design a trial and gather the necessary resources. The major reason for this lack of support, he suggested, is that clinical research is not widely respected among academics as a truly intellectual endeavor. Califf explained that, while investigators who are leading large, multisite trials predicted to have a major impact on clinical practice enjoy such respect, this is not the case for those conducting less visible work or just starting out in their research careers. A number of workshop participants expressed their support for rewarding academic researchers who conduct clinical trials. Early career development at the graduate and postgraduate levels could create incentives for more experts to enter the field of clinical research.

NEED TO NAVIGATE ADMINISTRATIVE AND REGULATORY REQUIREMENTS

The internal requirements of an academic institution, federal agency, or pharmaceutical company for reviewing multiple aspects of a clinical trial can significantly delay its initiation. In the case of an academic institution conducting a clinical trial for a pharmaceutical company, the internal review processes of both organizations are involved. In addition to such internal requirements, myriad federal and state regulatory requirements affect the conduct of clinical trials. Adhering to these many requirements is a significant challenge for investigators. Moreover, the delays incurred increase the time cost of a trial and decrease its overall efficiency. U.S. academic institutions typically take longer to navigate the approval process (i.e., from budget/contract to IRB approval) compared to private or academic institutions abroad. The protracted timeline to approve a clinical trial through U.S. institutions is one reason industry sponsors look outside the United States to initiate studies.

RECRUITMENT AND RETENTION OF PATIENTS

A core function of a successful clinical trial is finding patients who fit the predetermined eligibility criteria and getting them to participate. Each disease area addressed during the workshop (cardiovascular disease, depression, cancer, and diabetes) has a unique patient base for clinical trials, and the issues that affect patient enrollment in trials can vary according to features of the disease. In addition, workshop participants identified challenges to patient recruitment that transcend disease status.

Footnotes

1

The term “one-off” alludes to the current situation in which the necessary components of a trial (usually a single coordinating center and multiple research sites) are brought together for a discrete period of time and disbanded once the trial is completed.

2

A list of initial national priorities for CER recommended by the IOM in 2009 can be found at http://www​.iom.edu/Reports​/2009/ComparativeEffectivenessResearchPriorities.aspx.

3

An IRB is tasked with reviewing a clinical trial protocol to ensure that the study is conducted ethically and study participants are not likely to be harmed. An IRB can decide whether a clinical trial should continue as planned or changes should be made.

4

Equipoise is the point at which a rational, informed person has no preference between two (or more) available treatments (Lilford and Jackson, 1995). In clinical research, the ethical concept of equipoise is satisfied when genuine uncertainty exists as to the comparative therapeutic benefits of the therapies in each arm of a clinical trial.

5

The documents and conversations involved in the consent process explain the details of the clinical trial, including its purpose, the treatment procedures and schedule, potential risks and benefits, alternatives to participation, and the rights of participants (NCI, 2010).