Comments on National Bioethics Advisory Commission Report on Research Issues Involving Human Participants
February 17, 2001
Comments of Laura McBride, M.S., Peter Lurie, M.D., M.P.H. and Sidney M. Wolfe, M.D.
Public Citizen's Health Research Group
on the National Bioethics Advisory Commission Report Ethical and Policy Issues in Research Involving Human Participants
This report of the National Bioethics Advisory Commission (the "Commission") addresses weaknesses in and potential solutions for the current system of protections for research participants. The report is a significant review of the system and provides many thoughtful recommendations on improvements. The report describes the challenges facing the current system as "significant" and states that they call for "major change" (Chapter 1, p. 40, lines 3-4). We agree with this assessment, but feel that the changes recommended often fall short of "major change," particularly in the area of conflict of interest.
We first address two general issue areas that the Commission should consider in revising its report. Thereafter, we list a number of more specific comments on the report's recommendations.
1. The primary focus of the system — to protect research participants — seems to get blurred occasionally in the report. Only as a secondary goal should promoting efficient research be considered.
Although research obviously can benefit society greatly, the guiding ethical principles established by the 1978 Belmont Commission should remain sacrosanct. In several instances, this focus seems to get confused in the report. In Chapter 1 (p. 4), the report states, "in the ideal, an adequate and responsive oversight system should enhance the quality of the research while simultaneously protecting participants" (emphasis added). Later on p. 9, the text reads "[the current system] is more cumbersome and less effective than it must be both to support research and protect participants." This equal emphasis on ethics and efficiency seems to be a departure from the Commission's charter, which states that the Commission is to be review the "protection of the rights and welfare of human research subjects." Changing such language to emphasize ethics over efficiency can help reinforce to readers the importance of the primary focus and goal of the protection system. Moreover, it addresses the temptation to justify weakening ethical protections by invoking claims of efficiency.
2. The report contains only scant discussion of conflict of interest concerns.
The report strikingly lacks a thorough discussion of one of the biggest and growing concerns that led to the formation of the Commission: the commercialization of research. If, as the Commission points out, the research community must retain the trust of the public, acknowledging the existence of conflicts is key. The Commission cannot be seen to be brushing this critical issue under the rug.
In the over 20 years since the Belmont Commission released its report, clinical research has become increasingly commercialized. More biomedical research is now funded by private industry than by the government, and much of it occurs in doctor's offices, out of the purview of academic Institutional Review Boards (IRBs). Research sponsors stand to reap enormous profits from the marketing of their products. Research investigators and institutions can receive thousands of dollars in revenue for their services. The size of the money pot has placed enormous pressure on investigators and research institutions to enroll patients and work faster, and this puts patients at risk through both inappropriate enrollment and inadequate protections during the trial. The report fails to acknowledge these important pressures up front and gives them only minor attention in later chapters. In fact, some report language seems to acquiesce to the need for speed. In Chapter 2, the report reads, "the system should allow such research to proceed with minimal delay ... An oversight system succeeds to the extent it encourages the protection of participants in all aspects of research, but does not create any unnecessary barriers to the conduct of research." (p. 3, lines 15-18)
The cursory review of the significance of conflicts of interests is of particular concern since the Commission intends to continue the system's reliance on a "moderately protectionist approach." This approach is defined as "the investigator remains responsible for providing protection, including obtaining informed consent from participants, but the research study undergoes external review by a group not involved in conducting the research." (Chapter 2, p. 3, lines 24-26) Reliance on investigators for the brunt of human subjects protections is clearly not adequate when the investigators themselves face conflicts from financial, institutional, and professional rewards and is hardly the kind of "major change" the report has stated elsewhere is necessary. To be effective, this report should create a list of situations for which it believes conflict of interest can be relieved only by taking measures that eliminate the conflict (e.g., stock ownership by researchers in the drugs they are investing, "finder's fees" for doctors) and others where disclosure would suffice (e.g., disclosure of funding sources to patients and journal readers).
The discussion of the regulations and guidances on p. 4 is unbalanced. While some appropriate observations about the merits of guidances are made, there is little discussion about the advantages of regulation or of the limitations of guidances, notably the tendency of some parties to simply ignore the latter since they are unenforceable. While researchers may prefer guidances to regulations, guidances are often insufficient to adequately protect study participants.
We are also concerned about the idea of allowing IRBs to delegate some minimal risk research to be reviewed by administrative personnel. In our view, initial reviews, even of minimal risk research, should be reviewed by IRB personnel, particularly because one of the more persistent problem areas in biomedical research ethics has been the categorization of research with more-than-minimal risk as minimal risk.
Recommendation 2.1: Application of Uniform Set of Regulations to All Human Research
On pp. 18-20, the Commission lays out a number of startling examples of unethical research, including a horrible display of ethical ignorance or malfeasance on the part of reconstructive surgeons who experimented with two different surgical techniques, one on each side of their patients' faces, without any evidence of informed consent. Other examples included questionable research conducted in weight-loss centers and dentists' or psychotherapists' offices. The Office for Protection from Research Risks (OPRR; now known as Office of Human Research Protections [OHRP]) was powerless to do anything as the research fell outside of its limited jurisdiction.
We therefore strongly support the Commission's recommendation that the protection system govern all research in the United States. Such a requirement is absolutely fundamental to an ethical system of protections and we join with the Commission in urging Congress to take action to accomplish this immediately. There can be no ethical justification for having one study to be regulated, while allowing an otherwise identical project located in another setting or funded by a different institution to proceed in an unregulated fashion. We cannot continue to allow research currently outside of the scope of the current system to continue without protections. Probably tens of thousands of individuals are participating in research without any protection whatsoever.
Recommendations 2.2-2.3: Creation of National Office of Human Research Oversight (NOHRO)
Moving the current responsibilities of OHRP (other than those only applicable to the Department of Health and Human Services) out of DHHS and into a national office for protections could improve coordination of and facilitate compliance with federal regulations. The proposed structure of NOHRO also reduces conflicts of interest and bureaucratic barriers that have plagued OPRR since its origin within the National Institutes of Health and that continue, somewhat improved, with OHRP.
An office with national prominence such as that recommended by the Commission would be an affirmative statement about the importance of consistent participant protections across federal departments and the private sector. However, much of the structure of the NOHRO remains vague. Through stronger recommendations, the Commission would more likely ensure that the office would actually come into existence and receive the necessary authority and resources to adequately carry out its function.
The Commission, however, provides an alarming description of NOHRO's activity in the area of monitoring and enforcement. In both areas, the Commission is far too willing to defer important functions to accrediting bodies and institutions. This is yet another example of the impact of the report's ignoring of the rampant conflict of interest issues currently plaguing the clinical research system. The Commission is recklessly placing the important charge of monitoring and enforcing compliance in the hands of those involved in the conflicts themselves. Moreover, there is decided lack of clarity over which functions would be carried out by NOHRO and which by groups equivalent to OHRP. Here again, if inherent conflicts of interest are to be avoided, most monitoring and enforcement will remain in NOHRO. It is not sufficient to only involve NOHRO with "serious violations or repeat offenders." (p. 25, line 21)
We strongly urge the Commission to forcefully recommend that the NOHRO have a greater role than OPRR had in monitoring and enforcing human subject protections. Throughout its report, the Commission refers to the startling incidents of noncompliance and other protection violations uncovered during the too-infrequent site visits conducted by OPRR. Federal monitoring is a crucial role and helps assure to the clinical research community and participants themselves that someone is watching over the research process. Monitoring compliance is the ultimate form of ensuring accountability of all players and not a role that should be shirked by the national office.
Recommendation 2.3, continued:
The Need for Accountability — The report acknowledges this important facet of the protection system, but provides only minimal explanation of how this would be carried out. The report should provide stronger recommendations to federal agencies (because we believe they should play the central enforcement role) on the need for and content of site visits. For example, inspectors should observe studies actually being conducted, not content themselves with the review of paperwork. The visits should be conducted both in response to allegations of noncompliance and also as routine, prospective evaluations of site protections. These important responsibilities cannot be left to an accrediting body.
Review by Bodies Above IRBs — We support the Commission's recommendation that risky research be reviewed by a body above local IRBs. Prior reports have brought attention to the apparent lack of expertise often found at the local level to review today's increasingly complex research. However, because local IRBs may have more knowledge of what is occurring at the local research site, they must have ongoing communication with the central body. The local body must act as the eyes and ears of the central body which, by its nature, has only minimal presence at each research site.
Certification and Accreditation — Mandatory certification of investigators and IRB members and accreditation of IRBs and research institutions will help ensure that these crucial players in the research process have an understanding of the ethical issues involved in clinical research. Such certification and accreditation is long overdue. Much is needed in this area and the current lack of education of IRB members and researchers is a gross scar on the current research enterprise.
Recommendation 2.4: The Definition of Research
We are concerned about the use of "intent" as a factor in determining whether human subjects research is being conducted. It will always be convenient to claim that the purpose of the experiment was merely to benefit a particular patient (and thus classifiable as "practice" under the Commission's schema) rather than to generate knowledge (and thus classifiable as "research"). We would not be surprised if many of the researchers involved in the examples of abuses catalogued by the Commission made the claim that they were not in fact conducting research, simply treating patients. As in criminal cases, intent is always difficult to divine, and thus provides an opportunity for improper defenses of unethical research.
Some forms of data collection that the Commission deems "public health practice" should also not be exempted from regulation as human subjects research. In particular, outbreak investigations can involve potentially significant breaches of confidentiality, including details of personal behavior. We can see no reason for an outright exemption for this form of research.
A further illustration of the deficiencies of the research definition recommended by the Commission appears in Chapter 4, where the report declares that "[Phase] 4 [studies] are not research" (p. 31, lines 12-13). Phase 4 studies can range in design from long-term prospective studies to placebo-controlled, randomized controlled trials to short-term pharmacokinetic studies. Clearly, all of these carry the potential for serious harm for patients; it is illogical to exempt such studies from all of the elements of ethical protection simply because they occur after a drug is approved.
In the end, the Commission has provided general recommendations when specific ones would move the field forward. Why not provide a more inclusive list of examples of studies considered research and those not considered research? This would be far preferable than merely delegating the responsibility of making these distinctions to a NOHRO which does not exist (and may never).
Recommendation 2.5: The Definition of a Human Participant
While the exclusions from this category are generally appropriate, we are concerned that "information revealed about others" is one of the them. We would agree with this exclusion only if the report clarified that this information would not be linked to any identifiable individual.
One responsibility of IRBs is to consider alternatives to the study that has been proposed. As noted by the Commission, the IRB has the responsibility for assessing whether the study can answer the research question (p. 9, line 24). Unmentioned, however, is the responsibility of IRBs to make an assessment of whether an alternative research design that is more protective of participants might answer the research question as well or approximately as well as the research proposed. At a minimum, this is the sort of question that can be posed to researchers should the IRB have any doubts.
Similarly, while the report recommends not including "standard procedures" in the IRB's ethical analysis (p. 19, lines 18-19), there is a regrettable trend to not provide (or even remove) standard procedures as an element of the research. Recent examples include discontinuation of neuroleptic medications, failure to provide hepatitis B vaccination in prospective studies of at-risk individuals and randomization to placebos (or other inadequate interventions) when known effective interventions exist. (The latter may be acceptable when the condition is minor or self-limiting or the placebo effect is substantial.)
Recommendation 3.2: The Definition of Minimal Risk
We favor the current schema of permitting expedited review if the research is both considered "minimal risk" and falls into one of nine categories. As noted, the "minimal risk" designation is particularly prone to abuse and, particularly if some minimal risk research is to be reviewed only by administrative personnel (a Commission proposal we oppose), restricting "minimal risk" to limited categories of research seems necessary.
Recommendations 3.3 - 3.4: Informed Consent
The Commission acknowledges the three elements of informed consent (information (disclosure); comprehension (ascertaining understanding); and voluntarism (ensuring voluntariness)) when it states, "It is time to place the emphasis on the process of informed consent with the goals of ensuring that information is fully disclosed, that participants fully understand the research so that they can make informed choices, and their decisions to participate are always made voluntarily." However, discussion of the latter two elements is weak despite the strong evidence presented showing the failures of the current informed consent process.
We reiterate our concerns initially voiced in our November 13, 2000, comments on the Commission's report "Ethical and Policy Issues in International Research" regarding the need for measuring participants' true understanding during the consent process. Consent forms have become ridiculously lengthy (up to 16 or more pages) and are legalistic documents intended to protect the researchers and their institutions. As the Commission states, "the consent form ... is intended only to document the interaction between the participant and the investigator; it is only one part of the informed consent process." (p. 32, lines 8-10) Both IRBs and investigators spend significant time reviewing and altering the informed consent document, but do not, in general know how well participants are actually informed of the risks and benefits of the research. We strongly urge the Commission again to recommend that a random sample of participants should be surveyed to see if their understanding of the trial satisfies a reasonable standard of informed consent. This needs to be done before the research is actually commenced and, if their understanding is deficient, the process must be changed so that the study only begins (or continues) after adequate understanding of the study procedures has been documented. In order to accomplish this, the Commission should recommend that there be research conducted into efficient and effective methods for gauging participants' true understanding so that IRBs and investigators are better educated about the most effective means of informing participants.
Different techniques for consenting participants, such as videos, are potentially very valuable, but the system still needs some way to ensure accountability. A simple signature is important in this regard. Requiring a participant's signature could be as easy as signing a standard statement that they participated in the informed consent process and are satisfied with the information presented.
Monitoring the consent process is also a good way of ensuring the latter two elements of informed consent. As mentioned by the Commission, monitoring could be random spot-checks or could be more thorough for very risky research. Consent monitoring also serves as a valuable tool for IRB education and gets IRB members out of the board room and into the field.
Recommendations 3.7 - 3.8: Privacy and Confidentiality
NOHRO should be charged with ensuring that adequate regulations are promulgated so that privacy and confidentiality of participants and their personal information will be maintained. The increasing use of databases and electronic medical records make it easy for physicians and their staff to scroll through medical records in order to identify potential eligible participants. Investigators advertise the ease with which they can identify participants and the number to whom they have access when they are promoting their services to research sponsors.
Regulations should ensure that no one besides patients' own physicians have access to such private information. Research coordinators, administrative staff, or sponsor representatives should not be allowed access. Particular concerns arise when, as the Commission points out, there are mandatory reporting requirements. In such instances, third parties should not be allowed to gain access to these data warehouses in order to identify possible participants for their research, especially since patient information may be reported without their knowledge.
Recommendation 3.9: Vulnerability
Vulnerable participants do deserve more stringent protections in the current system. We support the Commission's recommendations in this regard. However, the Commission does not give enough attention to the largest vulnerable participant population — one that makes up an enormous percentage of the research subject population: investigators' own patients.
In the discussion of 'deferential vulnerability' the report states, "it should be noted that not all deferential behavior is subordinating" (p. 53, line 17) But, substantial concern exists within the research community and outside that the deferential relationship between patient and physician is indeed subordinating and the trust placed in one's physician can be exploited. Investigators approaching their own patients for participation in research constitutes a good example of when IRB monitoring of the consent process or the use of third-party advocates is crucial.
Recommendations 4.1 - 4.3: Education
We support the Commission's recommendations for education on research ethics in medical school and mandatory, rather than voluntary, certification of investigators. Certification is particularly important as more and more research is conducted outside of tertiary academic medical centers and increasingly private practice doctors are playing a role in clinical research without the checks and balances of the traditional research infrastructure.
Recommendation 4.4: Accreditation
IRB accreditation, too, especially if mandatory rather than voluntary (OHRP has proposed only a weak voluntary accreditation process), can help raise the bar in terms of IRB performance, but we again urge the Commission to be stronger in advocating a closer link between federal agencies charged with monitoring/enforcement and the accrediting bodies. It is essential that if, for example, accrediting bodies were to be delegated some regulatory role in overseeing IRBs and institutions, the information they gather must be shared with the federal bodies and NOHRO. Effective federal oversight requires that information on findings and potential problems be shared with the bodies with the ultimate responsibility for protections.
Recommendations 4.5 - 4.6: Registration/Assurances of Compliance
Registration of IRBs and research institutions should be mandatory. As for IRB accreditation, the OHRP has opted for an inadequate voluntary approach, and the report is thus an improvement over the OHRP proposal in this regard. It is impossible to have effective oversight without knowing exactly who is conducting research and where. Equally essential is the verification of actual research practices. Both qualitative and quantitative evidence on the questionable application of research protections is startling and unnerving (see p. 16, lines 1-5).
While verification conducted by IRBs and institutions has potential value, this still raises questions about the potential effects of conflicts of interest. Clinical research means big bucks for institutions and investigators. Institutions cannot be left to be the only source of verification and assuring compliance.
Recommendations 4.7 - 4.9: Managing Conflicts of Interest
As stated earlier, the report contains only a very superficial discussion of investigators' conflicts of interest. This fails to capture the true pervasiveness and significance of these conflicts on the conduct of clinical research today. The Commission turns to IRBs to try and mitigate the effects of such conflicts, but the Commission fails to realize that in most cases, IRBs do not have access to the necessary information. As the Inspector General noted, many IRBs would like to review financial information, but they currently lack the ability to do so. It should be noted that OHRP's recently released draft guidance on "Financial Relationships in Clinical Research" is equally weak in prescription, relying solely on disclosure rather than promulgating stronger positions on what financial relationships should be prohibited. The report should present examples of the types of financial conflicts of interest that would be prohibited. Some examples of these are provided on p. 2 of these comments.
The Commission's formal recommendation on disclosure is more diluted than the text supports. On p. 21, the Commission acknowledges that "such [financial] information may be relevant to an individual deciding whether to participate in a research project," but then goes on only to recommend that the information be disclosed to participants only "when appropriate." (p. 26) This recommendation does not fit the evidence put forth by the Commission on earlier pages. Disclosure is very important and serves as an important method of improving participants' informed consent. The Commission points to Food and Drug Administration's current policy on financial disclosure, but it is important to note that this review of financial conflicts occurs after the research has taken place and, thus, after the risks to participants have materialized.
We strongly support the Commission's recommendation that 50% of the IRB include non-institutional members and 50% include nonscientific members. The current requirement of only one noninstitutional and one nonscientific member, which can be met by the placement of only one individual, is clearly inadequate to ensure an independent review of research. Again, this is particularly true as the significance of clinical research monies skyrockets.
Recommendations 4.10 - 4.13: Continuing Review, Amendments, and Adverse Event Reports
A standardized system for reporting adverse events would most certainly improve compliance with federal regulations and, in turn, enforcement of human subjects protections. As the Commission noted in the draft report, compliance with reporting requirements has been spotty. For example, recent attention to gene therapy research, and in particular a study resulting in the untimely death of an 18-year-old participant, has brought light to this problem. The resulting investigation found that significant adverse event information had not been reported to the Recombinant DNA Advisory Committee in the National Institutes of Health. Also, reports to the Food and Drug Administration were filed later than required by law.
Local IRB review of off-site adverse event reports adds little value to the review process. Further, IRB members often lack the expertise and experience necessary to effectively evaluate an event's significance. Despite the potential value of off-site or central review of adverse event reports, the system must also incorporate feedback channels among all the parties, including Data Safety Monitoring Boards (DSMBs), central IRBs, sponsors, investigators and local IRBs.
The fourth portion of Recommendation 4.13 which now requires "reporting of adverse events in a manner that addresses the concerns of sponsors about the confidentiality of proprietary information." Maintaining the confidentiality of proprietary information is not a legitimate concern for an ethics commission; maintaining the confidentiality of patients is. If science is to benefit from data gathered in prior research and if participants are to be best protected by ensuring that the researchers in the study for which they are volunteering are cognizant of adverse events identified by previous researchers, public disclosure of all adverse events is mandatory, independent of the stage of the drug development process. The Food and Drug Administration has recently proposed just this for gene therapy and xenotransplantation experiments, but we can see no reason that this should not be extended to all fields of human research. This would mean that adverse events in any human experimentation would become a matter of public record, while still guarding the privacy of patients and providers.
The report also emphasizes the growing role of DSMBs in protecting participants. While DSMBs are a welcome addition to the repertoire of participant protections, the report offers not guidance as to when DSMBs should be formed. While it is customary to do so in large, long-term, randomized drug trials, there is no cut-off in the report as to the size and/or duration of studies that should have DSMBs. Furthermore, additional forms of research, including non-randomized prospective studies and some behavioral research, would seem to be good candidates for DSMBs.
Recommendations 4.14 - 4.15: Local vs. Central IRB Review
Central IRB review serves to address two overarching concerns. As mentioned earlier, local IRBs may not be adequately qualified to review today's increasingly complex research. Also, participant protections are not furthered when hundreds of IRBs review the same research protocol independently. One IRB might notice a problem and remedy it at their site, but a less-scrutinizing IRB may not and the problem would remain unresolved. While local input may occasionally be important, central IRB review by more experienced and less pressured IRB members can help shore up participant protections and insure a sufficient level of protections at all research sites. Here, again, however, communication between central and local IRB is key (see comments to Recommendation 2.3 above).
Recommendation 4.16: Compensation for Research Injuries
Research-related injuries should most certainly be compensated by research sponsors. Sponsors receive enormous profits from subjecting individuals to health risks and should therefore be obligated to compensate them in the event of adverse effects. Neither the participants themselves nor the research institutions should have to bear these costs. It is offensive that participants volunteering to participate in research designed not to benefit them may face significant medical bills while research sponsors sit idly by. The implication in the way this recommendation is drafted is that it would be a no-fault system. This is not preferable to the deterrent effect of a fault-based system because it is blind to negligence on the part of the sponsor or investigator. Moreover, compensation beyond just medical and rehabilitation costs is sometimes necessary. Whereas funds could be set aside by research sponsors for prompt compensation for medical and rehabilitation costs, patients and their families should retain the right to sue for pain and suffering and negligence beyond those costs.