While it may be a common complaint among researchers that IRBs overregulate, many IRBs opt to “uncheck the box” by building additional flexibility into the regulations that allows them to reduce investigator burden while still ensuring equivalent protections for human subjects.
But what is an IRB to do when federal agencies overregulate, requiring higher levels of IRB review than are necessary for the scope of a study, or for investigators to jump through extra-regulatory hoops, in order for funding to be released?
In the three years I’ve been working as an IRB Analyst, the majority of the pushback I’ve seen from federal agencies has been with respect to our not human subjects research (NHSR) determinations, though I have seen it occur with exempt studies as well. In these instances, a study will undergo IRB review and be determined NHSR or exempt, but the program officer (predominantly from the National Science Foundation [NSF]) will reject the IRB determination and let the principal investigator (PI) know that a higher level of IRB review is necessary in order for their funding to be released. We will then guide the PI to amend their study to undergo this higher level of review so they can get their funding and start their research with minimum further delay.
At our institution, we recommend that our investigators comply with these requests, to avoid jeopardizing their funding opportunities. However, our IRB posits that sometimes these requests create additional legwork for our investigators and IRB staff without increasing participant safety, and in some instances, have resulted in increased incidences of non-compliance at our institution. (Recently, a PI at our institution was able to argue up the ranks at NSF for our exempt determination to be accepted, though this approach did take a number of weeks. Anecdotally, our IRB postulates that complying with the funding agencies’ review change requests may result in faster approvals). In auditing these studies, we have found that for many we still agree with our initial NHSR or exempt determination, so it's unclear if the pushback from the federal agencies is response to reviewer error. And often their determinations do not seem to reflect agency-wide policy, as the federal agencies (including NSF) do accept the majority of our NHSR and exempt determinations. Rather, it seems to reflect varying interpretations of the regulations among program officers.
When it comes to NHSR studies needing to be reviewed as exempt, or exempt studies needing to be reviewed as expedited, our IRB found that the program officers’ rejection of the lower level of review in favor of the higher level of review has increased both investigator and IRB burden to accomplish, what we feel in many cases is the same level of human subjects protections. To illustrate, NHSR and exempt determinations have the same level of IRB oversight (none, once the initial determination is made). And while an exempt study being reviewed at the expedited level will now require signed consent and annual review, the study was initially determined to be no greater than minimal risk. So, in the opinion of our IRB, these additional requirements reflect additional burden without benefit of additional protections to the participants.
Within the last few months, we’ve also seen federal agencies (primarily the National Institute on Aging [NIA]) increasingly extending the NIH definition of clinical trial to include studies that our IRB found not to qualify (including studies limited to secondary data analysis that we had classified as exempt 4 and expedited 5), and then requiring (per NIH policy) Data and Safety Monitoring Plans (DSMP) for all of these studies. Treating retrospective studies as clinical trials and requiring DSMPs does not increase data protections in any meaningful way, but does require additional legwork on the part of researchers and result in increased review time for investigators and IRB staff.
While the above examples of overregulation are, at worst, inconvenient, we have seen their more problematic effects at our institution when it has resulted in investigator confusion that ultimately ended in non-compliance. In particular, this has occurred when federal agencies have required a full human subjects approval for studies that lack definitive plans for involving human subjects.
Under 45 CFR 46.118, the regulations stipulate that “Applications and proposals lacking definite plans for involvement of human subjects…need not be reviewed by an IRB before an award may be made.” Many studies that fall under the 118 determination need funding in order to begin preliminary research activities (hiring personnel, developing protocols and measures, pilot testing, etc). Ultimately, these investigators may find that their study is feasible and then seek additional funding (and IRB approval) for human subjects research activities, but 118 allows them to obtain funding to lay the research groundwork.
When these requests come through our IRB, we will grant an administrative approval that, under the regulations, allow the federal agencies to release funding to the investigators for these initial NHSR activities. However, some program officers have rejected our 118 determination on particular studies, and required instead a full human subjects research approval in order to release the preliminary funding; it can be difficult, in some cases, for the IRB to then have oversight over a project thatlacks definite plans for involving human subjects. The requirement for a full review becomes problematic for researchers, who are expected to prematurely communicate their plans for conducting human subjects research to the IRB, when they don’t have the funding to staff their study team, develop their instruments, or to see if research with human subjects is even feasible. It is also problematic for the IRB, who must make a risk determination about a study without fully developed methodologies, measures, recruitment documents or consent documents, and approve the under-articulated study, so that the PI’s funding can be released.
When we've been faced with approving such studies in the past, we have included a reminder in the approval letter stating that activities involving human subjects cannot begin until an amendment has been submitted containing the updated protocol, measures, recruitment and consent documents. However, some of our investigators have mistaken the approval letter with caveats as a full approval. And in these instances, we only discovered that human subjects research had taken place at the time of continuing review. By that point, the misunderstanding had become an issue of non-compliance, and our IRB required that the researchers scrap the data they had collected without IRB approval.
Luckily, these studies were minimal risk, and harm did not come to the participants. In some cases, the researchers were able to re-contact the participants with an IRB approved consent form, and re-do their data collection (though this, of course, requires that participants volunteer more of their valuable time). I did not hear that any of these investigators lost their federal funding. At our site, we are also working to make our IRB communications more clear so researchers won’t be able to mistake the IRB “approval” (for the purposes of releasing funding) with an IRB approval for research with human subjects.
But these examples of accidental non-compliance are extremely troubling because they resulted from an extra-regulatory requirement by the federal agencies. Had the 118 determination been accepted, I don’t think this confusion or the resulting non-compliance would have taken place. In the future, it may be helpful for our IRB to seek more information from program officers in order to clarify their reasoning in overturning some of our determinations in favor of a higher level of review. But we also feel it is important topush back on overregulationwhere extra-regulatory requests, made in an effort to offer greater protections, actually only increase legwork for PIs and IRBs and result in confusion and increased incidences of non-compliance that could result in harm to human subjects.
Sarah Luery is an IRB Administrator at the University of Southern California.