by László M. Szabó, Esq., Director of the Office of Research Regulatory Affairs at Rutgers, The State University of New Jersey
PRIM&R is pleased to continue our series of posts from the PRIM&R Blog Squad at the 2013 Advancing Ethical Research (AER) Conference. The Blog Squad is composed of PRIM&R members who will blog here, on Ampersand, to give our readers an inside peek of what's happening at the conference on November 7-9 in Boston, MA.
I attended an all-day pre-conference program on Central IRB Models: Benefits, Challenges, and Role in Clinical Trial Networks. The program was introduced and moderated by P. Pearl O’Rourke, MD. In her introduction, Pearl discussed that central IRBs are highly encouraged by the National Institutes of Health. Generally, there are three tiers for central/shared IRB models: tier 1—reliance on central IRB; tier 2—designation of central IRB; and, tier 3—reliance on local IRB. Funding preference generally reflect the numerical ordering of the tiers. Thinking of it on a spectrum, models for central IRBs range from solely local IRB review, to shared review by local and central IRBs (existing in the middle of the spectrum and sometimes referred to as a "federated" model), to sole review by a central IRB.
The reason for central IRBs is fairly commonsense: avoid the cat herding and gridlock that would inevitably take place if several IRBs needed to review a single protocol (especially as every and each IRB could endlessly revise a consent form). Such an approach would raise issues such as: which IRB stamps the consent forms (e.g., all? each? each for their own site?); how are adverse events reported; how is continuing and/or serious non-compliance dealt with; and how is it possible to decrease arbitrariness in the review process (because it's not possible to completely eliminate arbitrariness). IRB review, much like the judicial process, carries with it an inevitable range of human arbitrariness in the decision-making process. This often results in a range within which IRBs change protocols without altering anything truly substantive. Sometimes this is based on group think, IRB institutional cultural practices, or legal liability hyper-aversion. The end result is often increased cost and latency in the review process. I'm sure we've all reflected on how to overcome such matters.
Central IRBs offer a promising solution to this, but, of course, present their own challenges: they standardize review to harmonize practices, but the standardization can overlook or impact certain concerns that a local IRB may not be comfortable with. Similarly, as a matter of IRB culture, many IRBs are reluctant to relinquish review of a study conducted by one of "their" principal investigators. Equally relevant from the local IRB's perspective is how reliable is the central IRB? And, how can their reliability be assessed: AAHRPP accreditation? Reputation in the field?
Another aspect, thinking about it from the perspective of a director who has to consider staffing and costs, is whether central IRBs offer any cost sharing. The data on decreasing the costs associated with IRB review (e.g., staffing, software, etc.) through centralized IRB review is unclear. A common theme appears to be that costs remain, which appears to be due in some part to many institutions opting for a shared/federated model (near the middle of the spectrum I noted above), for fear of wholly relinquishing control over a study.
Julie Kaneshiro, MA, provided the Office for Human Research Protection’s (OHRP’s) perspectives on central IRBs. Julie noted that for shared/federated models, ensuring review of local context is a significant area where a local IRB has to ensure that it is discharging its “Common Rule” obligation. Although "local context" does not appear in the Common Rule, the phrase is a term of art grounded in the Common Rule's sections 45 CFR 46.107(a), (d), (f), and 45 CFR 46.111. It is helpful to keep in mind that a central IRB can also discharge the Common Rule's local context review obligation, just as any IRB reviewing research taking place elsewhere. Finally, Julie noted that the regulations provide IRBs with a significant amount of flexibility in opting to select central IRBs. Given the flexibility in the regulations, as well as the complexity in shared/federated models, central IRBs perhaps are most effective at the end of the spectrum: sole review by a central IRB. This flexibility was reflected to some degree in OHRP's 2009 and 2011 advanced notices of proposed rulemaking, which discussed the use of centralized IRBs. Along these lines, I have found it valuable to frequently remind myself, and be reminded, that the regulations do not just dictate obligations (which are too easily the reflexive focus of IRBs myopically focused on compliance), but also a range of freedom of actions that IRBs have the regulatory discretion to exercise. Often, IRBs incorrectly see this discretion as vagueness or lack of guidance from OHRP. It is important to keep in mind that the regulations and OHRP provide a framework for obligations and authority—not a fact-driven solution applied to each situation. They provide the recipe, we're the cooks.
Next, Emily Chi Fogler, Esq. presented on the legal aspects of reliance agreements. She explained that, as with contracts in general, memorializing the operational aspects and foreseeable scenarios will work to alleviate future contractual wrangling and confusion that can later stall progress on the study. Along these lines, areas to consider: scope (single study or all), updating information and communication, HIPAA, local research context (including site-specific, such as financial conflicts of interests), cooperation (especially for investigations), record keeping, and termination. A valuable aspect of the same agreement Emily discussed is that it requires the investigator who is the principal investigator (PI) to sign a commitment statement. That document essentially summarizes the reliance agreement, and spurs to get awareness and compliance ownership by the PI.
Finally, Elizabeth Hohmann, MD discussed some of the challenges and elements of implementing a central IRB for clinical trials. Elizabeth gave some first-hand and pragmatic examples of implementation, focusing on: if finances permit, procuring an electronic protocol system to handle large loads (her central IRB, NeuroNEXT, oversees about 7,000 protocols), distributing some responsibility to engaged sites in strategic ways so that staff sizes are kept in check, and keeping NIH (or whatever the funding entity is) apprised of what works and what doesn't.
The program was really well suited to IRB chairs and directors of IRB offices, as well as any staff involved in managing or developing a clinical trial network. The handouts also contained an excellent bibliography (actually, more talks should). A few articles of note:
- Check DK, Weinfurt KP, et al. (2013). "Use of Central Institutional Review Boards for Multicenter Clinical Trials in the United States: A Review of the Literature." Clinical Trials. 10(4): 560-7.
- Klitzman, R. (2011). "How Local IRBs View Central IRBs in the US." BMC Medical Ethics. 12:13.
- Silverman HC, Hull SC, et al. (2001). "Variability Among Institutional Review Boards' Decisions Within the Context of a Multicenter Trial." Critical Care Medicine. 29(2): 235-41.
Takeaway: central IRBs are here to stay, they're likely to increase, and become required for more studies. The framework and resources (such as PRIM&R) to manage these relationships is, largely, there. Cost savings are unclear.