The NIH’s nearly 6,000-word final policy requiring the use of a single IRB for NIH-funded, multisite research leaves IRBs and research organizations with many unanswered questions, according to the comments the organization received since the Dec. 3, 2014, draft policy was published.

Some of the remaining questions include the following:

Will there be criteria or a template for sites serving as single IRB of record?

NIH could help research institutions with this policy change by providing a “set of template documents that specify criteria for IRBs to participate in the reliance program,” wrote Elliott M. Antman, MD, FAHA, immediate past president of the American Heart Association, in a comment dated Jan. 20, 2015.

Antman refers to a sample template on the Clinical and Translational Science Awards (CTSA) program site at The Master Reciprocal Common IRB Reliance Agreement (MRA) contains explanations for the following three elements:

  • common language and regulatory interpretation, including “uncheck the box” on the Federalwide Assurance (FWA); harmonization of subject injury language; insurance coverage, and privacy compliance requirements,
  • common processes and consistent approach, including identifying the reviewing IRB; defining chain of responsibility and communication; quality assurance mechanisms; aligning certification and continuing education requirements; controlling study activation; unanticipated problems, serious adverse events, and deviations; and requesting and agreeing to IRB reliance, and
  • common standard operating procedures (SOPs), including reliance agreement policy; audit SOP; SOP for reporting serious adverse events and unanticipated problems; SOP for data protection and incident responses; SOP for managing conflicts of interest, and SOP for training.

The Society or Clinical Research Sites (SCRS) also called for NIH to establish criteria for evaluation of central IRBs when selecting a single IRB for a specific multicenter clinical trial.

According to SCRS’ comment, dated Jan. 29, 2015, NIH should do following:

  • investigate compliance history of the single IRB,
  • review qualifications of board members, including therapeutic expertise,
  • request references and review organization’s history of working with institutions and/or sponsors,
  • evaluate IRB’s ability to step seamlessly into the process (including state laws and local considerations),
  • determine scope and associated costs of services provided,
  • establish communication process between institution, investigator, and IRB, and
  • assess operational processes (frequency of board meetings, document management, capacity, turnaround time, quality assurance processes), and inquire about technology used by the central IRB and compatibility with existing systems.

Will NIH provide templates and details for addressing local contextual issues?

As institutional comments noted, there is need for additional details and template forms for information about local contextual issues, including investigator competence, site suitability, state laws, and community standards. “In addition, provide details about when ad hoc members or consultants would be necessary to review local contextual issues,” wrote several IRB and research officials at Tufts University and Tufts Medical Center of Boston, in a comment letter dated Jan. 14, 2015.

How will research institutions and IRBs handle the differences between institutions’ electronic data systems and technologies?

Many IRB software systems are part of larger human research protection programs, including investigational drug pharmacy, biosafety program, radiation safety, and other offices, notes David L. Wynes, PhD, vice president for research administration at Emory University in Atlanta.

“Every time an external central IRB is used, the automated notification and information-sharing system is disrupted,” Wynes wrote in a comment letter dated Jan. 21, 2015.

“This not only increases the risk of noncompliance with institutional, regulatory, and sponsor requirement but also forces investigators to enter information into independent systems for these other units,” he wrote. “These systems can be adjusted via programming special workflows for a limited number of central or commercial IRBs and we have implemented these workflows to rely on WIRB and NCI’s [National Cancer Institute’s] CIRB, for example.”

But it’s not feasible to program unique workflows for a large number of new central IRBs, and this issue is not resolved in NIH’s final policy, Wynes says of the policy published June 21, 2016.

“The NIH policy is set on getting approval of protocols and not on the other parts of the relationship, which are extremely complicated and involved,” he says. “I think they have underestimated and not given appropriate weight to the importance of these other issues and the complexities that come with doing this with 20 to 100 IRBs; if I can do it with five IRBs, great, but that’s not the model that’s been put forward.”

How will NIH ensure central IRBs can effectively report noncompliance and unanticipated problems?

IRBs might find it challenging to manage multisite research, and this raises questions about how well they’ll handle noncompliance and unanticipated problems reporting, according to a comment dated Jan. 29, 2015, by the WIRB-Copernicus Group (WCG).

“Administrative processes must be flexible enough to accommodate the unique needs of each institution,” according to the comment. “For example, consent forms and other study materials will also need to reflect local differences, including site-specific legal and institutional requirements.”

NIH’s final policy doesn’t address these and other questions IRBs and research organizations have, says David Borasky, MPH, CIP, vice president of quality management at Copernicus Group, a WCG company in Princeton, NJ.

“It will be interesting to see what it looks like when NIH’s guidance comes out,” Borasky says. “From looking at the comments on the draft policy, institutions are well aware this will be a very challenging policy for them to implement if they want to be a single IRB.”

Even with nearly a year to get ready, it will be challenging for most IRBs to prepare. “In my own opinion, I don’t think a guidance document will solve these issues, although it will be nice if NIH puts out a list of recommendations for a single IRB,” Borasky says.