Common compliance problems: IRB issues

OHRP study points to IRB failures

Add up the number of citations of noncompliance related to IRB issues that the Office of Human Research Protections (OHRP) at the Department of Health and Human Services in Rockville, MD, handed out between October 1998 and June 2002 and the numbers are pretty staggering. Of the 1,120 citations given to 155 institutions, 1,014 of them say something about IRB noncompliance and deficiencies.

Those figures were reported last fall in a review published in IRB: Ethics and Human Research.1 The review was done partly in response to a request from the Institute of Medicine.

During the time frame studied, 269 compliance oversight letters were sent to those 155 institutions. Only 13 of the institutions received no citations of noncompliance. Among the citations that were issued:

  • 27% related to problems with IRB approved informed consent documents and processes.
  • 25% related to the IRB initial review process.
  • 10% related to problems with continuing reviews, with half in this category failing to conduct continuing reviews at least every year.
  • 8% found problems with written IRB policies and procedures, and nearly as many found issue with IRB records.

Of the 155 institutions that got the citations, more than half of them had deficiencies in the initial review processes by the IRB, and 55% had issues with written IRB policies and procedures. Just over half the institutions cited had problems with informed consent processes and documents. Continuing review was a problem in 45% of the institutions, and 37% had issues with IRB records. A quarter of the institutions had problems with IRB review of protocol changes. At the very bottom of the list, 5% had overburdened IRBs.

Not written down? Didn’t happen

Despite the numbers, things are probably getting a little better, says Kristina Borror, PhD, director of the Division of Compliance Oversight at the OHRP. "In 2002, there were more total findings issued than in 2001, but 2003 looked lower." That said, there were 37 letters issued in the context of a single case. "The previous highest one month total we issued was in April 2002, when 30 letters relating to two cases were sent."

Besides, she adds, "I’m pretty sure that there are problems at every institution," she says. "There are always ways in which we can do better."

That so many of the problems related to having good written IRB procedures, and that so much of the total number of citations related to IRB issues just shows areas where improvements can and should be made, probably for every single research institution, she says.

"I think one of the real messages here is that you have to document things. If you don’t document it, you can’t show you did it," says Borror. Furthermore, if you do document things, "you are less likely to forget to do things."

Borror says the most common problem in written procedures that she has seen is that institutions must have written procedures for determining which projects require review more than once a year, and which need verification that no material changes have occurred. "People don’t know what that means," she says."

Another issue is that some institutions will have written procedures that cite regulations rather than explaining how they will apply those regulations. They simply iterate what the law says, Borror explains, rather than explain what the institution does to comply with the rules.

There is help, though. On the OHRP web site, there are some model procedures, as well as guidance on preparing written procedures. All of the examples, she says, are taken from institutions that have done a great job at fulfilling those requirements.

The documents and guidance can be found at

Borror says she has some suggestions for staying in compliance. First and foremost, pay attention to the informed consent process. In the 330 citations related to informed consent reported in the study, the problems ran the gamut from problems of description — everything from explaining risks and discomforts to benefits, alternatives, and even the purpose of the study — to the forms being too complex for easy understanding. Twenty-six of the citations related to a complete failure to obtain informed consent.

"It’s not just about making sure that the document has all the right elements, but that the person you are recruiting understands the research, understands it is voluntary, and understands what participating means for them," Borror says. "You have to be absolutely sure that there is no undue influence from the coordinator or researcher."

Every year, there are study subjects who report that they felt coerced, she continues. "It can be very subtle. Highly educated people can think that since they are being asked to participate, it must be a good thing for them. They have to do this because they think it’s the right thing to do, not because the researcher does."

It isn’t just small practices with little research experience that have problems with informed consent, either. In a compliance letter issued Feb. 26, University of California San Francisco researchers were brought to task for having informed consent documents that incorrectly stated the research would not lead to a change in routine care for the patients agreeing to be part of the protocol.

Compliance letters are always available at the OHRP web site and make for informative and instructive reading, particularly as they include the corrective actions taken by the institution and approved by the OHRP.

Second, Borror admonishes all investigators and their staffs to follow the rules. Believe it or not, "there are cowboys out there who do research without IRB review, without true informed consent. They don’t think of it as research, but as innovation or a change in techniques. But if they are collecting data on outcomes, then it is research."

Surveys and retrospective chart reviews also count. If you aren’t interacting with subject or providing an intervention, that doesn’t mean it’s not research, she says.

Borror hopes to keep collecting the letters of citation so that investigators and institutions can get a good sense of what some of the most common lapses are. In the meantime, she suggestss looking monthly at the determination letters (see the web site at lindex.htm) for guidance.