Centralization of IRB reviews have been increasing in the United States and elsewhere, but many questions about it remain. In the United States, a few centralized IRBs (CIRBs) have been established, but how they do and could operate remain unclear.
The resulting data have critical implications for future policy and research. Debates need to evolve beyond simply a binary discussion of whether CIRBs should replace local IRBs, to examine how and to what degree different models might operate, and what the relative advantages and disadvantages of each are. While some critics see CIRBs as panaceas, certain problems appear likely to continue. Careful consideration needs to be given to whether the advantages of local IRBs outweigh the problems that result, and whether a system can be developed that provides these benefits, while avoiding the disadvantages of local IRBs.
Centralization of reviews by Institutional Review Boards (IRBs) in the United States, and Research Ethics Committees (RECs), as they are called in many other countries, have been increasing in several areas of the world, but many questions about this trend remain. The European Union (EU) has sought to harmonize reviews of clinical trials across RECs in member countries, yet responses within individual countries have varied. Within EU countries, central, regional, and local RECs often exist, and tensions between these have arisen, including debates on how, and to what degree to unify processes and decisions.
In the United States as well, several centralized IRBs (CIRBs) have been established, generating controversies and questions about how they do and should operate. CIRBs can be either governmental bodies or private for-profit commercial entities. These two varieties differ in certain key regards—e.g., in purpose, and potential conflicts of interest. Yet similarities exist in that both have been established as alternatives to perceived limitations of local IRBs.
Critics have argued that the current system of review is flawed. In particular, documented discrepancies in IRB reviews of multi-site studies have led many critics to call for far more use of CIRBs—presumably through governmental or other non-profit mechanisms. Such discrepancies can require changes in protocols that delay research, and can make data from different sites difficult to compare.
Yet in the United States, CIRBs, though proposed over the past 15 years, have been instituted only to relatively limited extents, and resistance has arisen. Several models of CIRBs have been suggested for multi-site studies, generally allowing local IRBs to accept, reject, or amend these CIRB reviews. But it is not clear whether such increased centralization will occur, and if so, to what degree, how, and when. These issues are significant, since the impact, outcomes, and effectiveness of CIRBs will doubtlessly affect whether they are more widely adopted.
In a recent in-depth semi-structured interview study I conducted of views and approaches toward research integrity (RI) among IRB chairs, directors, administrators, and members, issues concerning CIRBs repeatedly arose. The study aimed to understand how IRBs viewed RI, which these participants defined very broadly. Interviewees revealed how they defined, viewed, and addressed integrity in research, and responded to violations of RI in a wide variety of ways, related to how they saw and approached the roles and responsibilities of IRBs; interpreted and applied federal regulations; viewed and interacted with researchers, federal agencies, institutions, and industry funders; and were affected by histories of violations of RI and audits at their own and other institutions, and psychological and personality issues on their IRB. They varied on whether additional guidelines and regulations would be helpful, and if so, what, where, and why; and how they related and responded to principle investigators (PIs), tried to improve these relationships, responded to tensions and complaints concerning IRBs, and interfaced with researchers in the developing world.
Issues frequently arose concerning how IRBs differed, and whether CIRB mechanisms might be advantageous or disadvantageous, and how, when, and why. Since the study used qualitative methods, it allowed for further detailed explorations of domains that emerged, shedding light on these issues.
A few studies have probed logistical aspects of IRBs (e.g., sociodemographics of members, and length of time that transpires before, though the quality of reviews was not assessed. Yet CIRBs have encountered resistance. Among medical school IRBs, 76% have never used a CIRB, and 73% thought there was no reason to do so since their local IRB worked efficiently. Surprisingly, no published studies have systematically examined how local IRBs view and experience CIRBs. Thus, this paper examines IRB chairs, staff, and members' attitudes and interactions with CIRBs - how these individuals have viewed and interfaced with these entities, and what they perceive to be the advantages and disadvantages of each of these two mechanisms for review.
Interviewees tended to have strong views concerning CIRBs, and highlighted advantages and disadvantages of both local and centralized processes. Overall, however, these local IRB chairs, administrators, and members overwhelmingly supported local over centralized IRBs. Specifically, these interviewees argued that local IRBs can better provide local knowledge of subjects and of PIs, and "curbside consults" with PIs that can facilitate mutual trust, potentially decreasing non-compliance. Local IRBs may also appreciate local institutional culture. While a few interviewees acknowledged that the current system can impede multi-site studies, and that CIRBs could potentially reduce duplication of efforts and streamline work, in the end all interviewees supported the continuation of local IRBs in some form.
Importantly, these interviewees also suggested that the quality of CIRBs can range considerably. Many of these IRBs had had experience with the centralized reviews of the NCI for adult oncology trials, and of the Pediatric Oncology Group, and of for-profit IRBs, and felt that the nature and outcome of centralized IRBs can vary, depending on who happens to be members of the committee. Of note, both local and centralized IRBs appear to have unintended consequences, both good and bad, that have received little attention, and need to be more fully taken into account.
These interviewees felt that the local knowledge of subjects and PIs, long-standing informal interactions with these PIs, and consequent trust and IRB "comfort" are important. PIs may interact more fully and informally, and hence in many ways effectively, with local than centralized IRBs. In general, interactions can depend on both the quantity and quality of information exchanged. Communication early in a process can lead to avoidance of later difficulties (e.g., expending efforts on what turn out to be unfruitful approaches in addressing research ethics issues).
IRBs here believe that accessibility to PIs may enhance trust. Local IRBs can potentially play vital roles in providing "curbside consults" that may establish and maintain mutual trust that may, in turn, help ensure the protection of human subjects. Hence, informal interactions in an institution, established over time, can be critical. However, future research is needed to explore how PIs themselves view these issues—whether they agree, and what factors they feel increase or decrease their trust of IRBs.
Though usually seen in debates as fixed, objective, bureaucratic entities, IRBs emerge here as engaged in complex, dynamic social systems, shaped by particularized, individual relationships with PIs, institutions, and communities.
These social bonds include not only colleagues and PIs, but patient populations as well. IRB chairs and members, many of whom are themselves physicians, felt a sense of commitment to "their patients" - i.e., those of their institution - and felt that entrusting these patients to a distant, "anonymous" CIRB represented an abdication of this responsibility. This sense of commitment highlights how belonging to a community can generate deeply personal social, as well as moral, bonds that can potentially enhance protection of subjects. Belief in this sense of local bonds may fuel local resistance to CIRBs. But presumably, an ethical obligation to safeguard research subjects as much as possible should exist on CIRBs as well—i.e., be the same for participants elsewhere, too. Researchers should presumably not protect subjects at their home institution more than participants at another medical center.
The close knowledge of PIs and patients that local IRBs offer does not appear to have been the intent of the regulations. Rather, such IRBs were established to reflect local community laws and values. But between IRBs, discrepancies that arise may result instead from differences in local knowledge, or institutional history and culture, and personalities and idiosyncrasies of particular IRB chairs and/or vocal members, or may be far more complex and seem relatively random—i.e., not systematically resulting from clear, objectifiable factors. Further research is needed as well to probe more fully these potential underlying causes of discrepancies.
High levels of communication and trust between IRBs and PIs, and IRB knowledge of details about subject populations, and commitment to these individuals, are nonetheless valuable. But difficult tradeoffs then ensue. Local IRBs may enhance reviews (e.g., providing local knowledge of PIs) and/or generate costly and unnecessary discrepancies. However, it is not clear how to balance these advantages (which were not the intent of the regulations) against the disadvantages of local IRB reviews (e.g., inter-IRB discrepancies). Whether these advantages of local IRB review offset the resultant problems, and whether a system can be implemented that yields these benefits, while avoiding the disadvantages of local review, is not evident. Needs emerge to ensure that details of local knowledge are sufficiently incorporated into reviews, but how to do so such that reviews do not also become duplicative is unclear.
Within the United States, it is unknown, too, how often community values significantly differ between IRBs in ways that do or should shape specific IRB decisions. No systematic data have ever been published concerning this question. The fact that institutional and personality issues can play roles is not surprising, given the complex dynamic social processes involved. But while interviewees, in arguing for local IRB review, frequently aver that variations exist in community values, the examples that the men and women provide here appear instead often to reflect differences in local knowledge, and potentially other factors. Further research is thus vital on other, larger samples concerning how often, in what ways, and to what degree inter-IRB discrepancies result from differences in local values vs. local knowledge and other causes.
These interviewees' reasons for favoring local IRBs reflect attitudes (e.g., local knowledge of PIs and study populations), rather than objective evidence of improved human subject protection. Yet some of these attitudinal factors may influence protection of human subjects (the goal of IRB review) not directly, but indirectly. Systematic studies demonstrating the degrees to which IRBs in fact reduce concrete harms to subjects are still lacking. Granted, such evidence may be difficult to collect since such injuries related to deficiencies in research ethics may be relatively rare and hard to measure. Subjects may be hurt due to serious, unanticipated adverse events in a study protocol (e.g., previously unknown side effects of an experimental drug), but these may not be due to deficiencies in research ethics per se.
Nonetheless, enhanced local knowledge of subject populations and PIs (e.g., awareness that certain populations are vulnerable, and that certain PIs have been relatively more "cavalier" in the past) is important as it can facilitate human subject protection by raising caution in reviewing certain protocols. Such increased cognizance may heighten human protection indirectly—even if not directly and measurably—by decreasing the likelihood of threats to research ethics occurring.
Realtionship with PI create bonds
Relatedly, at first glance, two findings here may appear contradictory: on the one hand, interviewees support local IRBs, not because such local entities reflect community values, but because such local IRBs can facilitate interactions with PIs and provide local knowledge of PIs and subjects. Yet on the other hand, these interviewees are nonetheless wary that CIRBs may insufficiently attend to local values. In the end, however, these two sets of beliefs are not mutually exclusive. Rather, the existence of significant differences in local community values may be somewhat illusory, more than empirically-based. Despite this absence of empirical evidence, interviewees still support this rationale, because it is stipulated in 45-CFR-46, and appears to justify their existence, efforts, and expense.
These interviews also underscore several potential disadvantages of CIRBs. While some critics of IRBs may see these centralized entities as panaceas, certain problems appear likely to continue. CIRBs may be as subjective as local IRBs—i.e., shaped by the particular views, biases, or predilections of whoever happens to be the chairs and membership (e.g., how "nit-picky" vs. "user-friendly" and pro-research these individuals are).
These interviewees express, too, deep wariness and suspicion of CIRBs, and of "anonymous government bureaucracies," reflecting the same apprehension that underlies broader political caution of federal government involvement and notions of "states' rights." Yet as a result, local IRBs may resist efforts at centralization, potentially impeding the spread, efficiencies, and acceptance of such efforts. Given tensions due to interviewees' concerns—whether valid or not—that CIRBs will insufficiently incorporate local knowledge, other two-tiered models may be advantageous. Such mechanisms might still include some local review (e.g., having local reviewers contribute to CIRB discussions). But how such mechanisms would operate—e.g., in multi-site studies being reviewed by a CIRB - and how much such "dual reviews" would decrease expenses and eliminate discrepancies, are unclear.
Granted, as members of local IRBs, these interviewees may, in certain ways, be biased in their assessments. The IRB chairs and administrators here receive support—even if it is only part of their salary—for their work in the status quo, which they thus may be invested in maintaining. Future research on IRB members' views on these issues should take this potential limitation into account, and explore it more fully.
Policy questions surface as well concerning whether, how much, and in what ways to arrange for central and regional IRBs to interact with PIs closely, frequently, and effectively. At local IRBs, these interactions may be based on long-standing relationships and mutual trust, allowing for informality that can facilitate communication. Whether centralized IRBs can also establish such relationships is critical. To do so at a federal level may be impossible, since these interactions may then diminish "informality". The responses of a federal bureaucrat would presumably be "official" in a way that local "curbside consults" are not. But potentially, CIRB reviews can incorporate additional local knowledge in varying ways and to varying degrees, and hence, review options need not be simply dichotomous—local vs. central, all or none. Rather, enhanced local input and reviews could be made available to CIRBs.
Many of the characteristics of CIRBs explored here apply, too, to regional IRBs, which appear more common in other countries, and seem a logical middle-ground—permitting "local knowledge" and trust, while still streamlining the review process. But the current data suggest possible limitations of such a regional system as well. Vast differences can arise in regional IRBs, depending on exactly how regional these entities are (e.g., whether they cover a city, state, or large part of an entire country). These interviews suggest needs for additional research and guidance on, and possible further establishment of, such regional IRBs. The specific degrees to which such regional organizations indeed provide the advantages and avoid the disadvantages of both local and central IRBs require close investigation.
Questions emerge, too, as to which studies should be reviewed through central, regional, or local boards. While some critics recommend CIRBs for multi-site studies (where the costs of multiple, conflicting IRB reviews appear highest), single-site studies might then be left wholly to local review, and prey to local biases, and potential impediments. Subjects in single-site and multi-site studies may then receive different kinds or degrees of protection. Hence, it may make sense to consider CIRBs for single-site studies as well, though doing so may be harder to justify. These issues require further examination and discussion as well.
Given interviewees' resistance to CIRBs as structural solutions in efforts to improve IRBs, altering local IRB attitudes may be more achievable, and hence in need of further consideration. For instance, helping IRB members to become more cognizant of their own potential biases and assumptions that are not evidence-based might be beneficial.
These data have several additional implications for future research as well. Several "myths" about both local and central IRBs—concerning their respective advantages and disadvantages - may exist. As policy-makers and others contemplate increased CIRBs, it is critical to have as full an understanding of the potential pros and cons of these entities as possible. Studies are needed to explore more fully, among larger samples, how IRB chairs, members, and staff view and interact with local and centralized boards; what the roles and effects of informal communication are; how frequently and in what ways local IRBs in fact alter CIRB recommendations; how to develop and assess such models of CIRBs to enhance trust and overcome wariness; and how to guide the development of any such policy initiatives.
Further research is crucial, too, to assess how often IRBs interpret and apply regulations differently in ways that reflect psychological and institutional issues vs. differences in community values, and whether and how educational or other interventions can make IRB interpretations more uniform and less discrepant due to personalities and local institutional histories. Further data on actual IRB practices vs. assumptions can help guide these debates.
This study has several potential limitations. These data are based on in-depth interviews with individual IRB members and chairs, and did not include direct observation of IRBs making decisions, or investigation of IRB written records. Future research can also observe IRBs and examine such records. However, such additional data may be difficult to obtain since, anecdotally, IRBs have at times required that researchers obtain consent from all IRB members, as well as from the PIs and funders of protocols.
These interviews probed respondents' views at present and in the recent past, but not prospectively over time to assess whether they changed their views, and if so, why. Future research can explore these issues as well.
It is possible that response bias may have occurred here as well—i.e., that individuals who agreed to participate may differ systematically from those who did not. Those who responded may be more or less favorably inclined toward the status quo. However, there is no evidence that such differences exist either way. Further research can, however, also examine these possibilities among additional, larger samples of IRB chairs and members.
This study also did not assess views of CIRB chairs or members, but future research can examine their perspectives as well. In addition, among interviewees here, experiences with CIRBs varied in amount and type—e.g., from for-profit IRBs to governmental, from extensive to minimal interactions (i.e., from dealing with one to more than one CIRB, concerning one to more than one protocol), and from harmonious to problematic and conflictual. A few interviewees did not have direct experience, but knew of other IRB members or chairs at their own or at other institutions who had had such interactions.
Nonetheless, the attitudes of all of these interviewees are vital, since these may reflect prevalent views among local IRBs more generally, and highlight wider concerns, fears, and misconceptions. These perspectives are all critical, and need to be addressed, as efforts to centralize IRBs will no doubt continue.
In sum, these data highlight the intricate, complex tradeoffs and uncertainties that may be involved in assessing the advantages, disadvantages, and varieties of CIRBs, which should be considered in ongoing discussions concerning these and other approaches to improving human research subject protection. While some critics herald CIRBs as a virtual cure-all to current problems with IRBs, interviewees here underscore the importance of advancing the current debate from whether local or centralized IRBs should be instituted, to more nuanced analyses of how these and other models may best be used. Careful ongoing investigation and discussion of these issues are vital.
Editors Note: This article has been cut for space. We urge you to refer to the complete study at http://www.biomedcentral.com/1472-6939/12/13
By Robert Klitzman, Department of Psychiatry, Columbia University, 1051 Riverside Drive #15, New York, NY, e-mail: firstname.lastname@example.org. BMC Medical Ethics 2011, 12:13 doi:10.1186/1472-6939-12-13