IRB differences hamper multicenter trials
Approval times range from one day to 303 days
IRB processes are so varied from hospital to hospital that they can unnecessarily impede multicenter trials, according to a survey that looked at the responses of 68 hospitals to one study.
Different IRBs looking at an identical study protocol varied widely in their responses to it, according to the survey, which was published last year in the Journal of Nursing Scholarship.1
The time needed from submission to approval ranged from one day to 303 days. While most of the IRBs gave the study expedited review, others gave it exempt status and 17% required a full board review.
Expedited reviews actually took longer than full board reviews — a median of 30 days for expedited reviews, compared to 25 for full board. It was a finding that surprised one of the study’s authors, Elaine Larson, RN, PhD, FAAN, CIC, associate dean for research and professor of therapeutic and pharmaceutical research at the Columbia University School of Nursing in New York City.
"We probably need to get rid of the term expedited,’" she says with a laugh.
The survey stemmed from an earlier multi- site study in which Larson’s group looked at the effect of ICU nurses’ working conditions on patient and health care worker safety. Her group collected survey data from ICU nursing staff, retrospective data from Centers for Medicare & Medicaid Services on patients discharged from the same ICU, and hospital administrative data such as nurse-patient staffing ratios.
Since the focus of the study was health care-associated infections, participating hospitals had to have infection control departments, and were sought out through channels that emphasized infection monitoring and control.
In the end, 68 hospitals enrolled in the study. Mean bed size was 465. Nine of the hospitals were associated with academic health centers. Two hospitals didn’t have IRBs; in those cases, a hospital administrator approved the study protocol. In two cases, two hospitals in the same health care network required a single IRB submission.
Larson’s group noted that based on criteria by OHRP, their study did qualify to be expedited, despite the fact that 14 hospitals considered it exempt and 12 required full board review.
The result, she says, is that the study took longer than expected, particularly figuring in the one institution where the wait was more than nine months.
"It put us a bit behind, of course," Larson recalls. Noting the difficulties they were encountering, group sought its own IRB’s approval for the survey, which analyzed the responses they received from the other IRBs they worked with.
Larson’s group encountered a wide variance in IRB requirements. About a third required that the principal investigator listed on the application be from within the institution, while 26.5% required evidence of human subjects research training.
Forms for 44.1% of the hospitals were available on a web site. The forms themselves ranged in length from one to 31 pages, and anywhere from one to eight copies were requested.
The survey reinforces other studies that have pointed out a wide variability in the human subjects review process at different institutions.
Larson notes that she serves on an IRB and understands the commitment to getting human subjects protection right. But she worries that IRBs are becoming overcautious, which has the potential for keeping valuable research from moving forward.
"There are some times if you don’t respond quickly, you lose a funding opportunity," she says. "Certainly our top priority is to protect the patient’s safety, but added to that now is a concern that maybe an IRB will accidentally miss something, so that it almost becomes like a major scientific review — even if it’s a study that’s reviewed and funded by NIH.
"And of course, NIH grants now have a huge review for human subjects protection, too. I feel like we’re doubling the effort sometimes," Larson adds.
Larson suggests areas in which IRBs could streamline the process without endangering human safety. For example, she’d like to see IRBs be more willing to accept review by other IRBs when it’s appropriate.
"If there’s a multisite study, and one IRB, particularly a federal IRB, reviews a project, then the other IRBs should certainly look at local issues; but I just can’t see why you would need to have two or three or four or 68 reviews of the identical protocol," she says.
While it’s important for individual IRBs to make sure that the local needs of their patients and their institutions are met, there should be sufficiently clear overriding principles about human subjects protection that collaborative agreements can be possible, Larson says.
"Even academic health centers that have two or three hospitals linked with them often don’t accept each other’s IRBs. That, I think, is unfortunate."
Larson says she’d also like to see IRBs attempt to standardize their processes, so that investigators working with multiple institutions would find roughly the same processes in place. She’s hopeful that the recent move toward accreditation of human subjects protection programs will help in this respect.
"I think that accreditation has a lot of promise, because I would hope that not only would there be a rigorous public review, but there would be some standard operating procedures that would be appropriate," Larson says.
Marjorie Speers, PhD, executive director of the Association for the Accreditation of Human Research Protection Programs, agrees that standardization of processes is one goal of accreditation.
Speers says that in general, there are two kinds of variation in IRB practices — unnecessary variations, which stem from IRBs being unfamiliar with regulations and how to use them; and necessary variations, which take into account the fact that institutions are different, serving different patient populations.
"The kinds of issues that the authors are talking about in this paper — making determinations about exemption or expedited review — relate more to the kinds of variation that can be reduced," she says. "They have more to do with how you interpret the regulations, what you’re comfortable doing, or the time it takes to conduct an expedited review versus a full board review."
Speers says she’s seen other instances in which expedited reviews were found to take longer than full board reviews. She says there are a few explanations for why that might be happening.
"The institution may not have enough people to do expedited reviews, so you have a couple of people with a big workload," she says. "Another issue could be that the people doing expedited review are less confident as individuals to make certain decisions on a protocol than they would be if they were on the full board."
While accreditation ensures that an institution complies with all the necessary regulations, it also can help streamline the way in which protocols get approved, Speers says.
"It means they have a full system in place, with policies and procedures and programs that meet all the regulatory requirements," she says. "We look at the knowledge and understanding of IRBs and investigators — they have to be competent to do their respective jobs. So when you have a good system in place and you have knowledgeable people working within the system, then that leads to a more efficient and streamlined review process."
1. Larson E, Bratts T, Zwanziger J, Stone P. A survey of IRB process in 68 U.S. hospitals. Journal of Nursing Scholarship 2004; 36(3):260-264.