Improving Human Research Participant Protection Program Performance and Clarifying Roles
Throughout this report, the committee has emphasized that protection of human research participants is most effective when delivered in the context of a “system.” This system is complex and multifaceted and sometimes operates through elements or modules that cross organizational boundaries—and includes a number of distinct and definable processes.
This chapter continues the committee’s theme that program responsibilities include much more than the ethical review of protocols. Previous chapters have stressed the need to promote the incorporation of ethical principles in the design and conduct of studies, ensure that independent scientific review occurs—as well as considerations of financial conflicts of interest—and that mechanisms are in place for continuing review and monitoring of protocols, particularly those that pose more than minimal risk.
This chapter further argues for the need to “purify” the role of the Research Ethics Review Board1 (Research ERB) and the informed consent process by differentiating participation protection from other institutional matters. Also included are descriptions of various conflict of interest issues
and recommendations regarding the need to compensate participants for research-related injury, a topic that has been discussed at the national level for decades but never adequately addressed in practice. This chapter also highlights the need for continuous quality improvement (CQI), a critical means for ensuring that the various Human Research Participant Protection Program (HRPPP) functions are performing at optimal levels, and the potential of accreditation programs.
CONTINUOUS QUALITY IMPROVEMENT
Quality improvement (QI) in the context of health care involves individuals working together to improve systems and processes with the intent of securing the best possible outcomes. A catch phrase used in the quality field is, “If you always do what you always did, you will always get what you always got.” The premise, of course, is that standing pat is not a viable strategy when better performance is demanded. “Zero defects” may be a reasonable description of public expectations of protection programs, but it is far from a reality in current practice, in perception or in fact. Formal, systematic QI methods are widely used in the health care system and are at the heart of health care accreditation. One of the promises of accrediting protection programs previously highlighted by this committee is the much greater visibility of and attention to QI in participant protection efforts. Programs seeking accreditation will have to learn and implement this management approach.
However, even in the absence of accreditation preparation, programs can and should work on CQI of their program. Elements of CQI include the following:
identifying standards for the program,
benchmarking performance against that of leading programs,
searching for best practices to accomplish program functions and processes,
adapting identified best practices to the individual institution’s or sponsor’s situation,
performing self-assessments to determine the degree to which these processes are being successfully implemented,
using continuous improvement techniques to further refine the best practices, and
disseminating these refinements to aid other programs in the research community through journal articles and other channels.
Best practice in this context should not connote a belief that no further performance gains are possible—or expected. This term is used in QI to
indicate a proven approach to carrying out a work process efficiently and effectively. The word “proven” is important; most definitions of best practices anticipate that the process improvement has stated objectives, has been evaluated, and that sufficiently robust measurements exist to establish that the organization accomplishes the prescribed objectives. Hence, rather than an ultimate end, a best practice merely becomes the next target for additional process improvement, analyses, and refinement.
An illustration of a nascent QI intervention development process can be found in the area of informed consent, in which advanced protection programs have utilized “consent monitors” in studies involving significant risk and/or participants with impaired decision-making capacity and have assessed the monitors’ impact on the informed consent process (Silber, 2001). Such innovation is desirable in any program and can produce significant breakthroughs in QI processes.
The value of data to support both problem definition at the national level and quality assessment and QI at the program level cannot be too strongly emphasized. Data provide the program with the means to discharge its responsibilities for the participant protection system and enable decision makers to make programmatic decisions and allocate resources. Data on outcomes are especially important in assessing system performance. Yet the committee has been struck by the paucity of even the most basic information.
Information Tracking at the Federal Level
Recommendation 6.1: The Department of Health and Human Services should commission studies to gather baseline data on the current system of protections for participants in the research that it oversees and to assess whether the system is improving over time.
In recognition of the continuing need for information collection on the national human research protection system, the committee repeats this recommendation from its earlier report, Preserving Public Trust: Accreditation and Human Research Participant Protection Programs (IOM, 2001a, p.90). Clearly, this represents a formidable undertaking, yet it is one that as a society we should carry out if we are to understand and make appropriate changes in the current system (see also Recommendation 9 in Appendix B). The committee provides some suggestions about the kinds of data that are needed (Box 6.1), but recognizes that not all of these data can be collected at once and that some may be better suited to special studies than to ongoing reporting and collection mechanisms. In many cases, conducting scientific surveys involving representative samples rather than a full census will serve the development of policy as well and more cost effectively.
An initial step to the collection of data is to establish a prospective plan and to begin thinking about data needs, their sources, and their priority for action. Some conceptual and categorical clarification will be necessary in order to identify these data needs. For example, a study of research injury or harm cannot proceed without the prior development of a taxonomy that encompasses the spectrum of types of injury and harm and the contexts in which they occur. Research injury or harm is not limited to physical injury, and it does not always occur at the level of the individual. Research injury may include harm to dignity, psychological harm, or harm at the social level, such as the stereotyping or labeling of a group or community.
The studies and the data collection that would take place under this recommendation have several uses. They will provide essential data on
which to base policy decisions in the future. They will also point to ways in which the protection system can be improved and may help prioritize strategies for improving the system by pointing to strengths and weaknesses in the current national approach. Finally, the availability of these data could reassure the public and policy makers about those aspects of the current system that are functioning well and more clearly define those that are not.
In addition to sponsoring studies about the system as a whole, federal agencies facilitate performance improvement in other ways. For instance, the National Institutes of Health (NIH) intramural program participated as a test site in the pilot phase of the accreditation efforts. Agencies and other organizations interested in promoting accreditation and/or QI can support and promote such efforts by, for example, disseminating best practices, QI data metrics, and databases and by providing other general tools.
Notably, both the OHRP Division of Assurance and Quality Improvement and the Department of Veterans Administration (VA) Office of Research Compliance and Assurance (ORCA) have developed self-assessment tools2 that should enable programs to establish useful baseline measures against which they can assess their progress (Mather, 2002; OHRP, 2002a,b; ORCA, 2002; Roswell, 2002). These activities move in the right direction, because they highlight how oversight offices with responsibilities in compliance as well as education can facilitate improvement rather than focus solely on punitive measures. Based initially on a quality assurance (QA) self-assessment tool, OHRP is encouraging programs to work with it to identify areas for improvement in what ultimately will be a multiphased approach.
QA is an approach that compares current practice to defined standards of good quality (in the human research area, it would be rooted in compliance with basic regulatory requirements) and includes self-assessment, which provides a structured (and generally nonpunitive) way to determine the degree of compliance and areas of significant shortfall. QI goes beyond QA’s focus on identifying and correcting errors (Box 6.2); QI is a methodology and set of statistical and qualitative analysis tools that programs use to ascertain the most common underlying causes of shortfalls in, for example, Research ERB work processes and procedures, and develop improvements that would eliminate them. OHRP indicates that it will phase in a QI assessment tool and subsequently a CQI process at the institutional level (OHRP, 2002b).
An important aspect of the OHRP program is the intent to protect information submitted for QA/QI purposes from the compliance investigation function of OHRP, in recognition of the reality that institutions would not want to submit an honest self-assessment if it were to result in penalties
The OHRP tool is available online at ohrp.osophs.dhhs.gov/humansubjects/qip/qatooli.htm. The ORCA tool is available online at www.va.gov/orca/docs/Human_Subjects_Checklist.doc.
QA and QI are complementary, yet distinct strategies. QA asks the question, “Did we do the things that should be done?” and redresses problems as they are identified. QI asks the question, “What causes us not to do the things that should be done?” and seeks to modify the cause.
An illustration of QA and QI in participant protection might be found in continuing review of protocols. For example, a QA self-assessment might show that the Research ERB is late in completing continuing reviews on 20 percent of studies that, under regulation, should have ceased enrolling patients once the approval period expired. Having identified these errors, it would be necessary for the Research ERB to notify the investigators of this and to take steps to bring the institution and the studies into compliance. In contrast, a QI approach would be to prevent this from occurring.
QI would collect data about the processes used by the Research ERB and analyze more frequent and less frequent causes of the failure to learn why these deadlines were missed. Hypothetically, the QI study might determine that there is no system to trigger reminders, that the investigators fail to respond to notices, or that the Research ERB sometimes loses its quorum and cannot complete scheduled reviews. The most frequent causes would be subject to a review and refinement of the Research ERB’s work processes to prevent future occurrences—e.g., education of investigators or a longer lead-time in a reminder system. Subsequently, a QI study would remeasure the same variables to see if improvement occurred and to identify remaining causes of failure.
This illustration is a simple one, involving a process measure used in research reviews, but, of course, QI studies should also be undertaken to posit and measure participant protection outcomes.
for noncompliance.3 Unlike OHRP, ORCA does not have regulatory authority, and thus is currently better able to emphasize proactive, culture-building efforts within its oversight activities (as this committee has encour-
aged OHRP to do previously within this report).4 It is also important to note that although it is Research ERB processes that are the main focus of the current OHRP and ORCA self-assessment tools, the responsibility for assessing and improving program quality rests with the research organization, as the Research ERB is only one element of the protection program.
Information Tracking and Quality Improvement Within the HRPPP
Efforts to initiate QI measures in the research community have been stymied by the lack of empirical data regarding the performance of HRPPPs, measurable outcomes or other criteria for the ongoing evaluation of protection programs, and the scant formal knowledge of the approaches and methods by which effectiveness of protection programs has been improved. This is particularly surprising in clinical research programs, because CQI has been a prominent feature of health care QI for two decades (Berwick, 1990; Hughes, 1988; Juran and Godfrey, 1999). There is an extensive literature available in both scientific and professional journals about CQI in health care generally, addressing both the results of CQI efforts as well as methods, approaches, statistical metrics, and other aspects of the CQI process itself.5 Yet, with the exception perhaps of studies involving ways to conduct informed consent, there is a notable lack of published CQI literature on the elements of human research participant protection, much of which paradoxically takes place in the same settings in which CQI evolved and currently flourishes. The CQI field can accelerate QI if programs and their protection functions are made the focus of sound scientific research. In this way, experts in many disciplines—health services researchers, social and behavioral scientists, ethics researchers, and quality measurement experts—can contribute to building a new empirical knowledge base.
Recommendation 6.2: Research sponsors should initiate research programs and funding support for innovative research that would develop
criteria for evaluating program performance and enhancing the practice of quality improvement.
The development and validation of criteria for evaluating effectiveness is one type of research that is needed, as few efforts have been made to define appropriate, quantifiable outcomes of participant protection. For the most part, performance assessment has been based largely on how accurately protection processes have been conducted—adequacy of record keeping, adequacy of disclosures and warnings in consent forms, for example. There is a need to connect this process information in a scientific and measurable way to develop evaluation criteria that accurately reflect program performance. For example, criteria should be developed for assessing participant understanding and for measuring and reassessing it on a continuing basis. One key aspect relevant to these criteria would be measures of understanding of the differences between research and treatment. Once data systems are in place for monitoring safety, it will also be necessary to develop protocols for assessing whether injuries or other negative outcomes were avoidable.
Similarly, research into innovative or more effective ways to conduct HRPPPs is vital to address the need to devise suitable end points for measuring effectiveness. Another important research area is QI methodologies devised and/or adapted for use in the unique health settings of protection programs, as there is little experience to draw upon that is explicit to participant protection efforts.
Dedicated research funding would help move this objective forward. Federal agencies should provide support, given the public policy importance of this effort. The committee lauds the NIH effort to provide short-term interim support for institutional activities to strengthen participant protection efforts at institutions that receive significant NIH support for clinical research and the reopening of the program announcement on ethics (NIH, 2002b,c).
This research area would also be a fertile field for philanthropic organizations with an interest in enhancing the contribution of science to people’s lives. Moreover, industry sponsors have a need, an opportunity, and a responsibility to provide investigators and programs with information about processes central to their research and should support independent researchers in the study of human protection, in particular, in clinical trials. The use of a very small fraction of the resources now committed to clinical trials would vastly improve what is known about human protection.
Quality Assurance Database Needs
Many programs lack routine or automated systems for tracking key information regarding the studies under their purview. Information track-
ing systems are integral to the protection of research participants because they feed into QI and QA efforts, are likely to be required for some accreditation purposes, and are a means for measuring compliance. Mechanisms should be in place at critical junctures in the research process to ensure that the safety and interests of individual participants are maintained throughout the course of a project and that the data generated are valid.
Collecting research review data is a complex task that should be integrated into the practices of the program. Databases could be used to track, for example, protocol activity at each site, research personnel involved in study conduct, and appropriate credentialing and training per institution requirements (Box 6.3). There are often unforeseen uses for the types of information gathered in the conduct of human research, and a central database, with appropriate archiving and security measures to ensure confidentiality, assures that this information is available for self-assessment, policy development, research purposes, and QA support. Although QA is not QI, it provides programs with data about the conduct of human research, and it provides investigators an opportunity to learn through external evaluation. Data collection and analysis are a necessary precondition for both QA and QI.
STATUS OF ACCREDITATION
In its first report, Preserving Public Trust, this committee recommended the careful implementation of pilot projects for nongovernmental accreditation programs for HRPPPs and the research organizations responsible for them (IOM, 2001a). This recommendation was based on the potential for a constructive, performance-based accreditation system to facilitate within protection programs an emphasis on outcome measures as well as to provide a proactive, responsive mechanism that was able to incorporate feedback from accreditation stakeholders in order to meet evolving program needs. Further, participation in accreditation programs is a form of QA, as efforts to prepare to meet accreditation standards should ordinarily have beneficial effects, and at a minimum, will help ensure that programs will conduct self-assessments, presumably noting and addressing deficient areas.
Accreditation has considerable potential to systematize and accelerate QI processes. Site visits by accreditation programs determine if activities meet the standards set by the accreditation process and whether the organization has documented that it meets them. In addition, however, they require the organization to demonstrate that it has undertaken individualized local efforts to improve its activities. For example, the National Committee for Quality Assurance (NCQA) accreditation program, recommended in Preserving Public Trust as a suitable pilot program for the Department of Veterans Affairs (VA) medical centers, identified numerous areas in which it will review program QI activities. The expectation is that programs routinely collect QI data, systematically perform QI studies and analyses, and act to implement them. Examples explicitly identifying QI requirements in the NCQA standards include the following:
databases and information systems that provide QI data;
compliance in drug/device studies, correction of deficiencies;
adequacy and effectiveness of Research ERB processes; and
appropriate investigator conduct of informed consent process (NCQA, 2001).
Accreditation site visits provide a mechanism for identifying performance deficiencies, as well as for finding and commending strengths and excellent program performance. Accreditation organizations, such as the Association for the Accreditation of Human Research Protection Programs, Inc. (AAHRPP) and NCQA, as well as federal agencies participating in and promoting accreditation, can expedite the wider adoption of best practices by identifying them and, with the permission of the particular program, extending their reach through broader dissemination.6
Included within the AAHRPP website is a section devoted to the dissemination of Best Practices identified in the course of accreditation evaluations (www.aahrpp.org/best_practices.htm).
Amendments to improve and strengthen accreditation standards are also sometimes derived from these site visits. For example, descriptions of standards in accreditation guidance, responses to inquiries, Web site documents, and manuals produced and disseminated by federal and other research sponsors have the potential to raise the bar for future human protection program expectations and accomplishments.
Subsequent to the committee’s initial report, both the NCQA and AAHRPP accreditation programs have continued to work toward implementation. Each has developed its own set of standards and has conducted pilot site visits to begin refining them.
National Committee on Quality Assurance
NCQA is in the second year of a five-year contract with VA to develop and implement an accreditation program, which will apply to more than 120 VA medical centers that conduct research involving human participants. On November 15, 2001, NCQA released its final VA Human Research Protection Accreditation Program Accreditation Standards,7 to remain in effect until July 1, 2004 (NCQA, 2001). However, in response to problems identified during the initial series of site visits, revised standards currently are under development8 (Otto, 2002a). NCQA indicates that it will review and revise the standards annually in the future (Briefer French, 2002). As drafted in November 2001, however, the standards continue to fall short of sufficiently ensuring meaningful participant protection at various levels of program decision making and policy making.
During 2001, NCQA conducted pilot tests and subsequently more extensive field tests at VA hospitals to prepare to conduct active accreditation visits. Accreditation site visits began in September 2001, with visits to additional VA medical centers planned to take place at approximately weekly intervals. One anecdotal impression concerning the initial evaluations offered to the committee was a curious lack of awareness of CQI on the part of research institutions (Briefer French, 2002). Formal QI efforts— which are at the heart of health care delivery and hospital accreditation— seem more or less unknown and little practiced in research programs within the same settings. If this is so, it is a gap that accreditation preparation should close through the provision of training programs, the dissemination of research reports, and the provision of greater specificity in standards, guidelines, and site visit measurement tools.
sThe NCQA standards are available online at http://www.ncqa.org/Programs/QSG/VAHRPAP/vahrpapfinstds.pdf.
As this report went to press, the revised standards were available for comment at http://www.ncqa.org/Programs/QSG/VAHRPAP/vahrpapdraftstds.htm until October 4, 2002.
NCQA announced the results of its first 12 accreditation assessments in early April 2002. Within this group, nine VA medical centers received conditional accreditation and three failed to pass the evaluation9 (Otto, 2002b). The most common deficiencies identified included the lack of local policies and procedures regarding IRB structure and operations, inadequate procedures relating to the informed consent process and consent forms, and problems in the documentation of the initial protocol review evaluations IRBs are required to make (Roswell, 2002).
In response to feedback from VA centers that participated in the first round of the accreditation process, NCQA suspended its accreditation visits to institutions in April 2002. This pause was requested in order to assess and respond to those areas already identified through the accreditation process as requiring refinement or further development (Otto, 2002a). The committee notes that it is this ability to identify problems and take responsive action that makes the nongovernmental accreditation model advantageous.
Association for the Accreditation of Human Research Protection Programs
AAHRPP is a nonprofit organization established in 2001 that seeks to accredit organizations engaged in human research. AAHRPP’s declared intent is to provide accreditation of organizations involved in biomedical as well as social sciences, humanities, and other nonmedical types of research, such as business and engineering. AAHRPP states that its accreditation process “is voluntary, peer-driven and educationally focused, and aims to foster ‘a culture of conscience and responsibility’ within institutions seeking its services” (2002b). The accreditation process involves rigorous self-assessment, followed by a site visit from AAHRPP accreditors who are experts in practicing, teaching, and promoting human research protections.
AAHRPP released interim standards for public comment on October 15, 2001. The group was responsive to concerns expressed in this committee’s initial report10 about the need for broader utility within the standards, inclusion of more specific standards regarding participants and sponsors, and attention to CQI. One of nine “principles” enunciated by AAHRPP (Box 6.4) is that “Standards should promote the development and implementation of outcome measures that can provide a basis for demonstrating quality improvement over time” (AAHRPP, 2002c). The
SOURCE: AAHRPP Accreditation Principles (AAHRPP, 2002b).
group began pilot site visits at the end of 2001 and finalized its standards11 and procedures based on public comment and the results of pilot site visits in the spring of 2002, at which point it also began accepting applications for accreditation (AAHRPP, 2002a; Softcheck, 2002; Speers, 2002a). Accreditation evaluations of applicant institutions are expected to begin in the fall of 2002.12
Under AAHRPP’s program, an institution will receive Full Provisional Status or Qualified Accreditation, or Accreditation Withheld, based on a self-evaluation process and a subsequent site visit. The program will operate on a fee-for-service basis, with fees depending on several variables, including the number of research protocols and Research ERBs at an institution (AAHRPP, 2002d). The accreditation will be valid for three years.
Future Opportunities in Accreditation
The committee is encouraged by the efforts of these two organizations and notes that the fact that difficulties were encountered during the initial roll-out of accreditation programs is not unexpected. As the committee stated in Preserving Public Trust, “accreditation will not be successful until it is widely accepted as a mark of excellence” (IOM, 2001a, p.86), and this will require consistent and iterative feedback between the various parties involved. AAHRPP indicates that its accreditation standards are intended to apply to universities, hospitals, and pharmaceutical companies among others. NCQA, while currently applying its accreditation standards and procedures to VA medical centers under its contract, indicates that it is developing a business plan for accreditation of other sites and research sponsors as well (Briefer French, 2002). These two programs are to be commended for their progress, but the committee stresses that accreditation remains a nascent process that will require substantial time and development before a meaningful assessment of its added value can be made (see Recommendation 6.4).
Recommendation 6.3: Human Research Participant Protection Program accreditation programs should include a standard directed at establishing and identifying accountability for specific protection functions.
AAHRPP’s accreditation standards are available online at www.aahrpp.org/standards.htm.
Personal communication, Marjorie Speers, Executive Director, AAHRPP, August 19, 2002.
Current efforts to establish accreditation systems are just under way, and the proposed standards remain relatively new and untested. The process for the accreditation of programs is still being configured, and the organizations thus far identified to carry it out are taking on an unprecedented task. The committee therefore offers further suggestions for areas that can still benefit from action by NCQA and AAHRPP:
Continue to move toward valid performance measures in lieu of static “documentation” reviews.
Identify strategies and dissemination opportunities to share best practices and measured outcomes with the research community.
Contribute, by making it explicit in the standards, to clarifying and systematizing accountability for all functions of programs within the various settings and systems in which they can and do operate.
Consider accreditation not only of the research organization, but also of organizations established to carry out only one of the functions of a program, such as protocol review (e.g., independent Research ERBs).
As stated in this committee’s first report, independent, nongovernmental accreditation programs, operating under a voluntary mechanism, are likely to be more responsive to the changing demands and needs of protection programs than other existing models (IOM, 2001a). Emerging accreditation programs are, however, still best viewed as pilot projects that should be evaluated in light of field experience. Any accreditation system should be constructed as an evolving tool, and it cannot be expected to immediately correct deficiencies in the collective protection system. As a component of a long-term strategy to improve the quality of research oversight, however, these nongovernmental accreditation processes show promise. It remains unclear, however, how the research universe will be sorted between the two organizations and what ramifications any distinctions between programs might bring. It is encouraging to note that efforts are underway to develop a mechanism that will allow Research ERBs that serve VA facilities and are elements of an academic protection program that has been accredited by AAHRPP to be exempted from NCQA inspection (Otto, 2002a; Speers, 2002b).
Furthermore, the advent of these programs should not prevent the development of other strategies and options for the accreditation of participant protection programs. It may be efficient, for example, to incorporate protection program standards into other existing accreditation systems. For instance, most research organizations involved in health research are already involved in other accreditation reviews, such as medical school or university accreditation. Relevant accreditation bodies can usefully look at their overall accreditation program to ascertain if HRPPP functions might
reasonably be added to the multiple domains already covered in the institution’s self-assessment process and accreditation site visits.
Recommendation 6.4: Voluntary accreditation should continue to be pilot tested as an approach to strengthening human research participant protections. The Department of Health and Human Services should arrange for a substantive review and evaluation of the accreditation process after five years, to be conducted under the purview of an independent entity.
Recommendation 11 in the committee’s earlier report called for Congress and the Department of Health and Human Services (DHHS) to initiate studies evaluating accreditation (IOM, 2001a). The committee suggests that accreditation is a major system change and that it may take as long as five years to establish the value of this significant investment on the part of research organizations, accrediting bodies, and others involved in the national protection system. Moreover, an evaluation of such significance would benefit from being conducted in a scrupulously independent fashion by a credible party. DHHS should make arrangements to secure this independent review, bearing in mind that identifying the appropriate measures and assuring the availability of baseline information should be accomplished well in advance of the actual evaluation. Accreditation of HRPPPs may indeed be a powerful tool for accelerating and maintaining improvement in the provision of research protections to participants; however, the research community, accreditation programs, and government regulators should proceed prudently with the implementation and analysis of this strategy’s utility.
ROLE DIFFERENTIATION WITHIN THE HRPPP
As described in Chapter 3, Research ERBs should be reshaped to perform the role that they were originally intended to serve—ensuring participant protection through the careful ethical review of protocols. Although they are the cornerstone of a system in which other entities also have participant protection obligations, Research ERBs should not be expected to assume all of the responsibilities of a protection program, and they should be properly constituted to carry out their duties (Recommendation 3.5). The traditional IRB has too often become the “fall guy” for the institution and the review function, and consequently, it has become a catchall for various responsibilities of the research organization. In the committee’s refocused paradigm, the Research ERB should not be responsible for institutional risk management, for ensuring that the informed consent process protects the institution from harm, or for ensuring institu-
tional compliance with all research rules and regulations. These responsibilities should be clearly assigned to other units within the program.
Recent, widely publicized problems involving research injury have led to legal complaints and lawsuits,13 causing a number of sponsors and organizations conducting research to consider the extent of their liability (Blumenstyk, 2002; Dembner, 2002; Washburn, 2001). However, there is no research that is devoid of risk, and science could not advance without volunteers’ understanding and acceptance of the risks a study is expected to present. A wide range of risks is covered under the Common Rule’s14 current threshold of “more than minimal risk.” Hence, parent organizations of research programs may prudently bring risk management activities into their agreement to conduct research. As laudatory as this may be, however, the focus of risk management is to protect the organization from harm, and it would be unfortunate, from the broader societal perspective, if this became a barrier to capturing the public good of research findings or if it led to the implementation of formal efforts to curtail legal risks in lieu of genuine efforts to protect participants or to carry out ethical research.
Moreover, because it is distinct from participant protection, risk management should be a separate and discrete function from those that reside within the protection program. It is inappropriate to expect members of review boards to conduct their primary duties while also attempting to represent the institution’s need to identify and manage risk. It may be possible to link risk management activities to the deliberations in the scientific review process that precede full Research ERB review—recognizing that a careful review of methodologies and associated human exposures may help identify the true level of institutional risk involved. However, risk management operates through different mechanisms with different objectives than a protection program and cannot be assigned within a protection program.
When encountering risk issues, Research ERBs in particular and protection programs more generally should also consider the parent organiza-
tion’s corporate compliance office and program as a useful adjunct to an effective protection program. Health care institutions, for example, increasingly rely on a corporate compliance office. As contrasted with those responsible for risk management, compliance officers are more likely to focus on institutional conformity with legal requirements than on institutional protection per se. In that sense, compliance is similar to the QA functions described earlier.
Hospitals and other health care providers are probably most familiar with corporate compliance programs as they relate to the enforcement of the Medicare and Medicaid fraud and abuse laws (Box 6.5). Some of the functions of an HRPPP may be similar to those already being addressed by compliance programs. For example, assuring conformity with federal and institutional conflict of interest requirements might be one function performed by the compliance office. Generally assuring compliance with the Common Rule or FDA requirements might be yet another. Compliance offices would also be appropriate venues for reviewing compliance with the Health Insurance and Portability Accountability Act of 1996 (HIPAA) privacy requirements. In particular, they could review whether data had been properly de-identified or whether use and disclosure of data had been properly authorized in conformity with regulatory criteria.
There are several arguments for involving the compliance office in human participant protection:
It has direct access to the leadership of the institution and thus can call the attention of leadership to research compliance problems;
it carries on a standard-setting and education function within the institution that could incorporate research concerns;
it could establish a hotline for participants, as well as provide protection for whistleblowers;
in many institutions, the compliance office will be relatively well resourced, and can thus take pressure off Research ERBs and protection programs; and
it should already be identified within the institution with the authority of the law and thus has the advantage of having its concerns treated as priorities.
In fact, even in institutions in which the compliance office is solely concerned with fraud and abuse enforcement, the office should attend to research issues because of potential liability under the federal False Claims Act.15 Cases claiming research fraud have already been brought against
The DHHS Office of Inspector General (OIG) has published a series of guidances covering hospitals, physician’s practices, and a variety of other Medicare providers (see www.oig.hhs.gov/fraud/complianceguidance.html for more information). Compliance with the guidance is not mandatory, except for Medicare+Choice organizations. The OIG recommends implementation of compliance plans, however, and many health care institutions have followed this advice.
Although the OIG guidances are directed at compliance with the fraud and abuse laws, they are based on the Federal Sentencing Guidelines that apply to sentencing of corporate criminals. The basic idea behind corporate compliance is that an organization that effectively attempts to comply with the law should receive a mitigated sentence if it is later found guilty of in fact violating the law. Under the Federal Sentencing Guidelines, criminal fines can be reduced up to 94 percent under certain circumstances where an effective compliance program was in place.
The Sentencing Guidelines apply to all federal criminal laws, and thus compliance programs should address compliance with all federal laws, not simply the fraud and abuse laws.
An effective corporate compliance program has seven elements, defined in the Sentencing Guidelines and reaffirmed in the OIG’s compliance guidances:
In fact, even in institutions in which the compliance office is solely concerned with fraud and abuse enforcement, the office should attend to research issues. Several cases have already been brought against health care institutions by qui tam relators claiming various types of research fraud. See United States v. Christ, 2000 WL 432781 (S.D. Ohio 2000); United States v. Hektoen Institute for Medical Research, 35 F.Supp.2d 1078 (N.D. Ill. 1999); Moor-Jankowski v. Board of Trustees, 1998 WL 474084 (S.D.N.Y. 1999); Milam v. Regents, 912 F.Supp. 868 (D. Md. 1995). Although these cases have generally ruled against the qui tam relator, when considered together with other cases supporting false claim act liability for noncompliance with regulatory requirements, they support the notion that health care institutions face potential liability for knowing noncompliance with federal research requirements.
health care institutions. Although unsuccessful to date, these suits, when considered together with other cases supporting False Claim Act liability for regulatory noncompliance, suggest the potential liability risk of being aware of noncompliance with federal research requirements (Box 6.5).
Some institutions have found both the need and the resources to add compliance functions within the protection program mechanisms. Kaiser Permanente, for example, recently approved the establishment of two new headquarter positions: a research compliance training leader to support investigator and Research ERB training and a research compliance QA leader to focus on FDA-regulated clinical trials and human participant protection. Both positions will support and work with Kaiser Permanente’s eight region-based research programs and Research ERBs, with each region contributing a portion of the cost.16
Separating the Consent Process from Institutional Legal Matters
One expression of how sponsor and institutional interests have become entangled in the effort to conduct the Research ERB’s participant protection mission is the current informed consent document.17 In clinical trials, these documents can run to 8, 12, or even 20 single-spaced pages. It strains credulity to suppose that this amount of text supports the ethical purpose of such documents, which is to appropriately inform a potential participant’s decision to enroll in a study. Rather, these documents increasingly are driven by legal concerns, administrative needs, and many other interests beyond the three principles elucidated in the Belmont Report—respect for persons, beneficence, and justice (National Commission, 1979). The committee strongly believes that in the informed consent process and in its documentation, efforts should be made to separate and highlight the participant protection needs from the legal and liability requirements of the institution (Recommendation 3.4). This can be accomplished by substituting an informed consent process for the current document-driven approach (see Chapter 4). It also requires innovation and simplification in the delivery and communication of relevant information to participants. Non-protectionist issues increasingly have crept into the consent form and the trend seems to be worsening.
For example, revisions to the HIPAA final rule allow an authorization
for use and disclosure of protected health information to be included in a research consent form18 (Davidson, 2002). The committee would instead encourage programs to separate this disclosure documentation from the research consent process.
MANAGING CONFLICTS OF INTEREST
In 2001, the U.S. General Accounting Office (GAO) conducted a survey of five institutions undertaking human research and found that institutional policies regarding financial conflicts of interest had variable thresholds for disclosure, timetables for disclosure, requirements for IRB involvement, and procedures for disclosure (GAO, 2001). Guidelines about what constitutes acceptable and unacceptable levels of conflict and policies for managing them are in various stages of development among public and private organizations (Boxes 6.6. and 6.7).
Recommendation 6.5: Research organizations are responsible for the in-depth review of potential individual conflicts of interest for investigators, primary research staff, and Research Ethics Review Board (Research ERB) members. Such reviews should be carried out by a conflict of interest committee or designated oversight body that is shielded from institutional pressures or influence. Relevant findings should be transmitted to the Research ERB to inform the review process of proposed studies.
Research organizations have the ultimate responsibility for assuring that conflicts of interest are assessed and managed, and the organization where the research is conducted or under whose aegis the research is conducted (in the case of private practice investigators) should establish an independent, chartered, auditable conflict of interest body charged with determining the degree and extent of financial conflict of interest in specific research proposals. Many organizations already have a conflict of interest process and/or committee in place as a result of federal requirements (NHRPAC, 2001), but as noted in Chapter 3, it is imperative that explicit mechanisms be in place for this body to feed information about financial conflicts of interest into the Research ERB’s comprehensive ethical review of protocols. An organization might wish to have a liaison from the conflict of interest committee to the Research ERB. The qualifications of members of the conflict of interest committee should be articulated in an organization’s policies and procedures, which should also ensure that the operations of the conflict of interest body are not subject to organizational pressures.
Public and private research organizations should continue to build on the concordant principles emerging from federal and professional organiza-
One of the first attempts by professional organizations to address conflicts of interest in research was the 1990 report from the Association of American Medical Colleges (AAMC), in which conflicts of interest were defined as “situations in which financial or other personal considerations may compromise, or have the appearance of compromising, an investigator’s professional judgment in conducting or reporting research” (1990).
More recently, AAMC has offered detailed policy guidelines on individual financial conflicts of interest in research involving human participants (2001). AAMC declares that institutions should create and implement policies regarding financial conflicts of interest that should follow federal regulations and should contain a number of specific elements, including definitions; a description of the processes to report, assess, and manage conflicts of interest; the criteria for assessing conflicts; the sanctions for violations; and the processes for appeal.
The Association of American Universities (AAU) also released a report on conflicts of interest in 2001. That report, which addresses institutional as well as individual conflicts, classifies institutional conflicts into two categories: “potential conflicts involving university equity holdings or royalty arrangements and research programs; and potential conflicts involving university officials who make decisions with institution-wide implications, which can include department heads and leaders of laboratories” (2001, p.10). (AAMC released its Task Force on Financial Conflicts of Interest Report dealing with institutional conflicts of interest as this report went to press.) AAU emphasizes the need for effective policies to deal with conflicts of interest in research involving human participants and asserts that management of such conflicts is often more important than the conflicts themselves. AAU also offers a checklist of questions for institutional leaders regarding the management of individual conflicts of interest.
In 2000, the American Society of Gene Therapy (ASGT) adopted a policy on financial conflict of interest that states that “all investigators and team members directly responsible for patient selection, the informed consent process and/or clinical management in a trial must not have equity, stock options or comparable arrangements in companies sponsoring the trial. The ASGT requests its members to abstain from or to discontinue any arrangement that is not consonant with this policy” (ASGT, 2000).
The Association of Clinical Research Professionals’ Code of Ethics exhorts members to “avoid conflicts of interest in [their] own affairs and make full disclosure in advance of undertaking any matter that may be perceived as a conflict of interest” (ACRP, 2001).
Other professional societies also have policies on their members’ conflicts of interest and how to properly deal with them.
In 1995, the Public Health Service (PHS) and the National Science Foundation (NSF) adopted federal regulations for financial conflicts of interest, setting $10,000 or more than 5 percent ownership in any single entity as the threshold for disclosure of financial arrangements [42 CFR 50 subpart F; 60 Fed. Reg. 132, 35809 (July 11, 1995)]. FDA released its own financial disclosure regulations in 1998 that required investigators to report, among other things, payments of $25,000 beyond the cost of research and equity interests valued at more than $50,000 in sponsor companies (21 CFR 54, 312, 314, 320, 330, 601, 807, 812, 814, 860). FDA review of such conflicts is retrospective, whereas the PHS/NSF regulations require disclosure by the time a grant application is submitted. An important point, however, is that the federal regulations were intended to encourage objectivity in research practice—not to protect human research participants per se.
DHHS has begun to address conflicts of interest in research involving human participants, releasing a draft interim guidance on financial relationships in clinical research in 2001 (2001a). Based on a public conference held in August 2000, the guidance includes considerations for institutions, investigators, and IRB members and staff. It also offers suggestions for informed consent considerations. The guidance has not been finalized as of publication of this report.
There is broad agreement among research organizations and policy makers that clear policies and procedures about financial conflicts of interest are needed and that these should apply to all research with human participants regardless of the funding source. Although organizations whose research is funded by PHS or NSF or regulated by FDA are mandated to have such policies and procedures, GAO found inconsistency among policies at five research organizations and notes that the policies are not accessible to staff (GAO, 2001).
The National Human Research Protections Advisory Committee (NHRPAC) and the National Bioethics Advisory Commission (NBAC) also have addressed the topic in their recommendations to federal entities (NBAC, 2001b; NHRPAC, 2001). NHRPAC, commenting on the draft guidance of DHHS, advises DHHS to use the PHS threshold for the disclosure of conflicts and suggests procedures for assessing and managing conflicts of interest. It also addresses the need for disclosure, education, and compliance.
NBAC notes that “IRB review alone is not sufficient to manage financial conflicts” and suggests that institutions should increase their regulation of investigators’ financial conflicts (2001b, p.59). NBAC also proposes that noninstitutional IRB members can help to mitigate conflicts of interest and that conflicts should be disclosed to participants.
tions for dealing with individual and organizational financial conflicts of interest, regardless of the funding source and in addition to federal guidelines or guidance. These include but are not limited to the need for conflict of interest oversight bodies that are separate from Research ERBs, increased attention to institutional financial conflicts, and meaningful disclosure of conflicts to participants.
Institutional Conflicts of Interest
Recommendation 6.6: Research organizations should establish an external mechanism for the review of potential institutional conflicts of interest regarding research protocols. Findings from this body should be communicated to the Research Ethics Review Board for its consideration in the review of individual protocols.
Although research organizations often implement mechanisms to identify and manage financial conflict of interest at the individual level, they frequently neglect the same issues at the institutional level. The possibility that institutional conflicts of interest may undermine the validity of a research study, cause harm to individual research participants, and ultimately erode public trust in the research enterprise has not been explored sufficiently.
As academic institutions have increasingly entered into financial and collaborative research arrangements with private industry, as encouraged by the Bayh-Dole Act of 1980,19 institutional conflicts of interest have become a topic of growing concern and increasing public scrutiny (AAU, 2001; DHHS, 2001a; Emanuel and Steiner, 1995; Gillis, 2002; Moses and Martin, 2001).
Currently, no federal regulations or final guidance address institutional financial or nonfinancial conflicts of interest, although the DHHS draft interim guidance does address institutional conflicts of interest, as does proposed federal legislation.20 Federal agencies and appropriate interest groups should continue to develop guidelines for evaluating, and if appropriate, managing institutional conflicts of interest with the same rigor that is on-going to the pursuit of professional norms and standards regarding individual financial conflicts of interest.
DHHS draft interim guidance states the following:
When institutions consider entering into such business arrangements, they should consider establishing an independent advisory and oversight committee (institutional conflict of interest committee), if one does not already exist, to determine when their financial arrangements pose a conflict of interest, and if so, how those conflicts should be managed (DHHS, 2001a, p.3).
The committee believes that the conflict of interest committee charged to review institutional interests should be an external oversight or advisory committee if it is to ensure appropriately independent evaluation of conflicts and make recommendations to the Research ERB. Institutions should disclose relevant conflicts of interest to this body, which would then determine how to manage them or whether the institution should be prohibited from carrying out the research in which the conflict exists.
If the external conflict of interest committee or the Research ERB, in consultation with the conflict of interest body, determined that the research could proceed, it could require that the institution divest questionable holdings; conduct the research in question only as part of a multicenter trial; or identify an independent entity to monitor participant recruitment, the informed consent process, participant enrollment, data monitoring, and other aspects of the trial that could be adversely affected (or appear to be adversely affected) by the conflict. Outside experts also might be contracted to perform data analysis and interpretation. If the conflict involves an institutional official, the Research ERB could require that the official be excluded from decisions about the research or that the institution or official sell equity holdings or royalty interests if the research is to go forward at that institution.
Nonfinancial Conflicts of Interest
Although nonfinancial conflicts of interest have existed for a long time, they have attracted little formal attention or analysis. These nonfinancial conflicts of interest are, by their nature, more common in academic settings and largely intrinsic to the research profession itself. Examples include a researcher’s desire to attain academic advancement or tenure or to receive professional prestige or win scientific prizes, the need to obtain grants, etc. (Levinsky, 2002). Every successful investigator has some degree of self-interest in the research, and without the desire to increase knowledge or find new ways to prevent or treat disease, advances in health and medicine would not be possible. Although the focus in the press and in public and private organizations has been on financial conflicts, which are quantifiable, nonfinancial interests are more common and potentially more damaging to participants and to the integrity of the research itself. Self-interest
becomes unethical in human research when it clashes with the protection of research participants. Awareness should be raised at every level within research organizations that conduct human participant research regarding the nature of excessive self-interest and its harmful effects and to promote institutional cultures that do not tolerate runaway ambition (Levinsky, 2002). The evolving accreditation process should incorporate attention to nonfinancial conflicts in its assessments of human research protection programs, and the groups pursuing conflict of interest policies should work to develop guidelines that are as rigorous as those directed at individual financial conflicts.
At the institutional level, structural relationships could threaten the independent activities of individuals or committee operations. For example, junior faculty members serving on Research ERBs may be reluctant to raise concerns about protocols submitted by senior colleagues; this hesitancy may be particularly strong when reviewing a department chair’s project. Similarly, Research ERBs may feel pressure to support institutional perspectives or organizational interests in specific situations (see Chapter 3). Research organizations should take deliberate steps to avoid the potential impact of such scenarios on decisions that affect the protection of research participants.
Disclosure of Conflicts of Interest to Research Participants
Although many private and public organizations agree that information about potential conflicts of interest should be disclosed to participants, agreement on the level of detail and how the conflict should be effectively communicated has not been reached. Concern exists that detailed disclosures in a consent form could be overwhelming and not understandable to most participants. Nevertheless, in order to make informed decisions, participants have a right to know if the investigator, staff, or institution has a potential conflict of interest in the experiment and what that conflict is. Research ERBs should make the final determination about how information about these conflicts is presented to participants.
The consent form might include detailed information about a financial interest and its management (whether it is determined to be a conflict of interest or not). The participant should be advised that more information about the conflict and its management is available upon request. However, simple disclosure is not a substitute for in-depth conflict of interest review and subsequent Research ERB review or for the obligation to adhere to other aspects of ethical research.
COMPENSATION FOR RESEARCH-RELATED INJURY
Research cannot be entirely free of risk. Some research participants may incur a research-related injury even if the study is carried out without negligence and in full conformity with the protocol. Research participants injured as a result of a product defect or malfeasance or incompetence in the design and execution of the study can resort to the tort system for compensation, but those injured through no fault of their own or of anyone else have no legal recourse.
It is the committee’s impression that many research organizations conducting clinical trials agree to provide short-term medical care (during the course of the study) for research-related injuries (IOM, 1994a). The same is true of the medical centers of VA,21 NIH, and the Department of Defense (DoD, 2002; NIH CC, 2000). However, it is also the committee’s impression that few research organizations cover other relevant costs or compensation for lost earnings (IOM, 1994a; Levine, 1986). (The University of Washington, which covers long-term medical costs, is one of the exceptions; see below.) The committee is unaware of any organization that agrees to provide compensation for pain and suffering.
Compensation for research injury is not required by the Declaration of Helsinki or by the European Convention on Human Rights and Biomedicine (the Ovieto Convention) (Council of Europe Publishing, 1997; World Medical Association, 2000). However, the Council for International Organizations of Medical Sciences (CIOMS) does require that participants are equitably compensated for any physical, research-related injury, and the issue is much discussed, both in the United States and elsewhere (CIOMS, 1993). Indeed, a number of countries have already made such provisions (Box 6.8).
Because the contributions of science benefit society as a whole, it seems indisputable that society is obligated to assure that the few who are harmed in government-sponsored scientific research are appropriately compensated for study-related injuries. As the Department of Health, Education, and Welfare (DHEW) Taskforce, which focused solely on federally funded research, noted in 1977, “Because society is both the beneficiary and the sponsor of research, compensatory justice may come into play for the redress of injuries suffered by persons in connection with biomedical or behavioral research conducted, supported, or regulated by the Federal Government” (1977, p.VI-4). The costs of the loss should not fall on the research participant.
The same argument applies to privately funded research, perhaps to an even greater extent, as the economic survival of a company depends largely
The CIOMS Guideline 13 states, “Research subjects who suffer physical injury as a result of their participation are entitled to such financial or other assistance as would compensate them equitably for any temporary or permanent impairment or disability. In the case of death, their dependants are entitled to material compensation. The right to compensation may not be waived” (CIOMS, 1993). (The subtext of the Guideline states, however, that those “who suffer expected and foreseen adverse reactions from investigational therapies or other procedures performed to diagnose or prevent disease” are not owed compensation.)
The International Conference on Harmonization (ICH) Guidelines for Good Clinical Practice defers to national law, with respect to compensation for injury, stating:
“5.8.2 The sponsor’s policies and procedures should address the costs of treatment of trial subjects in the event of trial-related injuries in accordance with the applicable regulatory requirement(s).
5.8.3 When trial subjects receive compensation, the method and manner of compensation should comply with applicable regulatory requirement(s)” (ICH, 1996).
The laws of other nations vary, but most make some provision for compensation:
Germany has long required research sponsors to provide insurance to cover injuries to research subjects in pharmaceutical and medicinal product trials. The German scheme provides for no-fault compensation for research subjects from this insurance fund. The insurance covers economic loss, but not pain and suffering. The research subject must, however, show that the research resulted in an injury to the subject’s body or health, or death, and that no other person was liable in tort or contract for the injury. Also, injuries must occur within three years of the conclusion of the research. Finally, liability is limited, e.g., to 30 million DM for trials including more than 3,000 persons, 50 million DM per year.
New Zealand also provides for no-fault compensation for those injured through clinical trials involving mental or physical health or disease under its general accident compensation scheme.
In France, the research sponsor must carry liability insurance. If a research subject suffers injury, the sponsor will be liable on a no-fault basis for nontherapeutic research. If the subject is injured through research involving treatment, fault on the part of the research sponsor is presumed, but the sponsor may prove that neither the sponsor, the research institution, nor the researcher was at fault and thus escape liability.
Spanish law makes the sponsor of a trial, the principal researchers, and the medical director of the hospital in which research is carried out jointly liable for any injury suffered by a research subject not otherwise covered by insurance during a clinical trial. The law also establishes a presumption that injuries to the health of a subject suffered within a year of the trial were caused by the trial.
Finally, although British law does not require compensation for research participants, absent fault, the guidelines of the Association of the British Pharmaceutical Industry do provide that sponsors should enter into contracts with nonpatient research subjects. Sponsors agree to compensate for “any significant deteriora-
tion in health or well-being caused” by participation in a study, “calculated by reference to the amount of damages commonly awarded for similar injuries by an English court if liability is admitted.” Compensation is to be offered without proof of negligence. When patient volunteers are involved, compensation need only be paid when it is established, on the balance of probabilities, that the injury was caused by the research intervention, and then only for serious injury of an enduring and disabling character (including exacerbation of an existing condition) and not for temporary pain or discomfort or less serious or curable complaints. Although sponsors are counseled to pay compensation in close cases where proof of causation might be difficult, they are also excused from payment in cases in which the treated disease was very serious and the disclosed risk of treatment high. The contract is to provide arbitration for cases for which agreement cannot be reached.
The British National Health Service (NHS) itself does not accept responsibility for compensation for injury to research subjects, although NHS staff is liable for negligence in carrying out clinical trials. NHS is also supposed to enter into indemnity agreements with sponsors of clinical trials carried out within NHS institutions to protect NHS from liability. NHS guidelines do, however, provide that in exceptional circumstances NHS can provide ex gratia payments of up to £50,000 for non-negligent injury. The Royal Commission on Civil Liability and Compensation for Personal Injury recommended in 1978 some form of no-fault compensation for research participants, but no law providing generally for this has yet been adopted.
SOURCE: Deutsch and Taupitz, 2000; Kennedy and Grubb, 2000; Taupitz, 2001.
on the availability of participants to test new therapies, drugs, and other products. Because the participants are ultimately contributing to the profits of the company, any costs that result from the research should be the responsibility of the sponsor. Furthermore, whether a study is privately or publicly sponsored, the results are intended to eventually benefit all of society.
In the absence of a compensation system, lawsuits alleging research-related injury are increasing (Blumenstyk, 2002; Dembner, 2002; Washburn, 2001). The claims to date relate to inadequacy of consent, departure from the protocol, or other negligent activities. Inevitably, these cases invite the courts to expand the legal grounds of recovery, and in an uncertain legal environment cases can be inconsistently resolved.22 A no-
fault system could reduce pressure on the judicial system and allow injured parties speedier resolution of claims. The Tuskegee Syphilis Ad Hoc Advisory Panel has noted that “No matter how careful investigators may be, unavoidable injury to a few is the price society must pay for the privilege of engaging in research which ultimately benefits the many. Remitting injured subjects to the uncertainties of the law court is not a solution” (1973, p.23).
However, when the participant alleges that the injury was caused by a possibly defective product or possible negligence in the design and conduct of the study, the tort system is likely to remain the appropriate channel for redress, serving as a back up to the no-fault compensation agreement for cases in which elements of liability can be proved. (Were a no-fault compensation system in effect, risk management needs and concerns might be lessened to some degree, making it easier to achieve simplified consent forms as suggested in Recommendation 3.4.)
It has been argued that a participant relinquishes his or her right to compensation when giving informed consent, but in the committee’s view, a right to compensation for research-related injury should not be subject to waiver. The DHEW Task Force declared in its report that “The fact that a person has volunteered does not eliminate that person’s right to be compensated in the event of injury, whether or not the injury was foreseeable…. Even if a subject perfectly understands a research procedure and agrees to participate in that procedure, the subject’s consent does not, in and of itself, include, explicitly or implicitly, a waiver of compensation” (1977, p.VI-5-6).
Recommendation 6.7: The Department of Health and Human Services should assemble data on the incidence of research injuries and conduct economic analyses of their costs to help establish the potential magnitude of claims that would arise under a no-fault compensation system for such injuries.
The main impediment to the implementation of a compensation program for research-related injuries in the United States is that, despite decades of discussion and studies by a number of federal commissions, there remains little quantitative information regarding the number and severity of potentially compensable injuries and about the costs of implementing compensation programs (ACHRE, 1995; DHEW, 1977; NBAC, 2001a,b; President’s Commission, 1982a). The 1977 DHEW Task Force report estimated that of the 132,615 research subjects participating in the research reported in its survey, 3.7 percent of them suffered injuries. Of those injured, 79.2 percent of them had trivial injuries, 19.6 percent were temporarily disabled, 3 percent were permanently disabled, and 9 percent died
(1977). Although this information is the most comprehensive available, it has a number of serious limitations23 and is severely outdated.
The University of Washington is one of the few organizations, to this committee’s knowledge, that offers long-term compensation for research injuries.24 When the system was first established in the 1970s, few formal claims for compensation were processed, but officials at the university attributed this to a lack of knowledge about the program among participants (President’s Commission, 1982a). In 1998, University of Washington officials estimated that the School of Medicine enrolled about 100,000 people per year in clinical studies with potential for adverse effects and that the compensation program for those injured paid $2,300 to $5,000 total annually (Marwick, 1998). At the University of Washington, the investigator is responsible for reporting whether an adverse effect was the result of research, at which point the plan goes into effect. The school pays for medical care and related expenses, such as travel.
The DHEW report and the more recent University of Washington examples appear to provide the most relevant information about the number and cost of injuries, but, obviously, this information is both sketchy and inconsistent. However, based on the media attention given to adverse research events in recent years,25 the increased attention given by regulatory agencies to institutional noncompliance and financial conflicts of interest, and the growing pressures on the research system, the potential for diminished public trust in the research community is real (Marwick, 2002).
To ensure credibility, it is critical to have data about the number, severity, and costs of research injuries, and it is not acceptable for society to continue to leave unaddressed a fundamental ethical obligation for the simple want of basic information. Some pertinent information could be abstracted from FDA and research institutions’ adverse event reports or experiences at institutions such as the University of Washington that already offer self-funded no-fault compensation. Similarly, such analysis can make use of the international experience (Box 6.8).
In its telephone interviews, the Task Force relied on data provided by investigators using their own judgment; the interviewers suggested nontherapeutic research if the determination of the type of research was vague; the Task Force had to make assumptions about the length of participation by subjects; and the definition of therapeutic injury did not include the word “unanticipated.” Also, in 1977, the Common Rule was not in effect, and the research landscape was very different from that found today.
For more information see depts.washington.edu/hsd/INFO/MANUAL/99-VII.htm#VII-g.
Some recent articles on the topic include Blumenstyk, 2002; DeYoung and Nelson, 2000a,b; Flaherty, et al., 2000; Flaherty and Struck, 2000; LaFraniere, et al., 2000; Lemonick and Goldstein, 2002; Nelson, 2000; Shaywitz and Ausiello, 2001; Stephens, 2000; Stolberg, 2002; Wilson and Heath, 2001a,b,c,d,e,f,g.
One of the key questions that will require more sophisticated research is the portion of participant illness and injury attributable to the research itself versus the underlying condition of the participant, and the ease with which such determinations can be made. These determinations pertain to the basic issue that must be addressed in any no-fault system—that of causation. Was the injury attributable to the research, or was it a manifestation of the participant’s underlying condition? If this cannot be easily resolved when claims are presented, the costs of resolving the dispute may escalate the costs of the compensation system itself. Making such determinations will require supplemental studies, and the committee recommends that these studies be commissioned as soon as possible in order to guide public policy decisions and accreditation standard development in this area.
Recommendation 6.8: Organizations conducting research should compensate any research participant who is injured as a direct result of participating in research, without regard to fault. Compensation should include at least the costs of medical care and rehabilitation, and accrediting bodies should include such compensation as a requirement of accreditation.
In light of the ongoing need to recognize and address the needs of those who have been harmed26 as a result of research, the committee believes that a fair compensation system should be established as soon as possible. Accordingly, the committee endorses the conclusion reached by NBAC that “a comprehensive system of oversight of human research should include a mechanism to compensate participants for medical and rehabilitative costs from research injuries” (2001b, p.123). Furthermore, this committee believes that, in principle, adequate compensation to those harmed as a result of research should be more generous than that recommended by NBAC and should include full recovery for economic loss, including work-related disability, and in appropriate cases, for lost earnings of a deceased participant. However, in light of the existing uncertainties concerning the number and severity of research-related injuries, the committee recognizes that this objective is attainable only in stages and therefore suggests a two-step approach.
The first step, which should be taken as soon as possible, would implement a compensation program along the lines recommended by NBAC as a requirement for HRPPP accreditation. Accredited research organizations would be expected to identify, characterize, and report research-related injuries and to cover costs of medical care and rehabilitation that are attributable to research-related injury. Meanwhile, voluntary efforts would be
simultaneously undertaken by NIH and private sponsors to establish demonstration programs that would also cover lost income due to temporary or permanent disability or death under various plans for valuation and payment. After three to five years of experience and data collection, the entire compensation effort would be evaluated, including an assessment of whether the fee scale(s) used in the demonstration programs are perceived as fair and easy to administer by those who are affected by it. The assessment should provide a basis for informed judgments about the best approach to take, including the best model for measuring work-related disability.
Under the approach envisioned by the committee, the responsibility for compensation would fall initially on the institution or organization actually accountable for conducting the research, and its terms would be specified in the documentation accompanying the participant’s agreement to participate. Presumably, most research organizations will attempt to insure themselves against such losses, and a market for such insurance may eventually emerge, especially after the necessary data have been compiled. In the context of pharmaceutical research, the allocation of responsibility for compensation between the sponsoring company and the research organization will presumably be determined by contract. This strategy embraces the basic approach of the Association of the British Pharmaceutical Industry (see Box 6.8).
Alternatively, the government could establish a federal compensation program, which could be included as a direct cost within grants, a surcharge on medical bills, or money from general revenues (Annas, et al., 1977). The National Childhood Vaccine Compensation Injury Act of 1986 (P.L. 99-660) addressed similar concerns regarding those injured by vaccines by establishing a federal program to compensate those who were injured as a result of their contribution to public health. The act was prompted by concerns that lawsuits would diminish vaccine manufacturing, and thus limit access to vaccines, endangering public health. On moral grounds, both those who receive vaccines and research participants are contributing to society and deserve compensation if they are hurt while doing so.
PERSONAL LIABILITY OF RESEARCH ERB MEMBERS
As noted, litigation over research injury appears to be increasing (Blumenstyk, 2002; Washburn, 2001). Research organizations with effective protection programs, including effective Research ERBs, should be better protected from legal losses than those that employ less robust efforts. It is the hallmark of both QI and risk management to invest in good practice (“right the first time”) rather than later having to make costly responses to problems that could have been prevented.
Some recent legal filings have sought to name as defendants, among others, members of Research ERBs (Gelsinger v. University of Pennsylvania; Robertson, et al. v. McGee, et al.). Although no such lawsuits have yet prevailed, they may exert a chilling effect on both the willingness of individuals to become board members and on the independence of their decision making. Indemnification, which appears to be a common practice among larger research universities,27 remains important in the present litigation climate. Given the desirability of recruiting and appointing external members to review boards, it is especially important to insulate them from personal liability concerns. Organizations should indemnify both internal and external board members to prevent them from being unduly influenced by the personal risks of potential litigation, a protection that should be extended to individuals of other boards, such as DSMB/DMCs, Scientific Review Boards, and ethics and research expert consultants used by Research ERBs and other review bodies.
An effective protection program should have QA and QI measures in place in order to continuously assess its strengths and weaknesses and to redress those weaknesses. DHHS should facilitate the activities of those studying the effectiveness of the collective system by gathering baseline data about the current system in order to assess improvements. Research sponsors should take on a similar responsibility by funding original research that would enhance the practice of QI.
There is also a need for data about the prevalence, nature, and costs of research-related injuries. Organizations conducting research with human participants should compensate any participant who is injured as a direct result of participation in the research; this compensation should include that for direct medical care, rehabilitative costs, and after appropriate study, lost work time. Accrediting bodies should include this requirement within their accreditation standards.
Accreditation by design is intended to encourage organizations to strive for a high standard of performance. Current accreditation efforts in the field of research involving human participants are nascent and merit review after a sufficient period of pilot testing (likely five years or so). Because of the significance of the task, DHHS should arrange for such a review through a credible, independent entity. Accreditation organizations should also fos-
ter accountability within programs by adding a standard to explicitly establish and identify who is accountable for specific protection functions.
Because the Research ERB is responsible solely for the protection of research participants, it should not be held accountable for matters of institutional interest such as risk-management and the resolution of conflicts of interest. Therefore, these issues should be referred to the research organization’s management for resolution or delegation.