In July, PRIM&R collaborated with CITI Program to host the advanced-level webinar Data and Safety Monitoring: Advanced Issues and Case Studies. Expanding on introductory knowledge in the module Data and Safety Monitoring in Human Subjects Research, part of CITI Program’s Biomedical Basic course, this webinar described ways in which the IRB, the data and safety monitoring board (DSMB), the investigator, and the sponsor can work together to ensure scientific integrity and subject safety in clinical trials. Summaries of government and non-government organizations’ guidance as well as interactive case studies offered strategies for handling complex situations that may arise during data and safety monitoring, including when and how to report adverse and unanticipated events, when a DSMB is needed, and what is considered a conflict of interest.
After the webinar, presenters Stephen Davis, MPA, MSW and Susan Ellenberg, PhD responded to the attendee questions time didn’t permit us to address live. We’re pleased to share those answers with the readers of Ampersand.
How much information about the results of monitoring data quality should an IRB expect to see in a study’s annual report?
Stephen Davis (SD): In my experience, it is not unusual for trial coordinators to upload study monitoring visit reports that are prepared by the visiting monitor, and some human research protections programs (HRPPs) require such documentation as part of the renewal. Sponsors also often request that these reports be shared with the IRB. There may be several of these depending on the visit intervals. These reports will often denote monitoring deficiencies (e.g., discrepancies between the source documents and case report forms, incorrectly completed forms, etc.). Otherwise, there may be no additional information related to data quality.
How much authority does a site IRB have over the membership of a DSMB appointed by a study sponsor for a multisite study?
SD: In theory, the IRB could request that a DSMB appointed by a study sponsor contain additional members to address specific technical and ethical issues, and not allow a trial to be conducted at the local site if it is not satisfied with the DSMB membership composition. In practice, however, I have found that sponsors meticulously vet DSMB members to ensure that the composition is both comprehensive and competent, and this is usually not an issue.
I have heard that DSMBs are generally fully unblinded to the interim treatment comparisons. Won’t they be more objective if they just see the data by treatment group but don’t know which the experimental group is and which is the control group?
Susan Ellenberg (SE): It is very important that DSMBs know which treatment arm is which. The DSMB is the only group with access to the interim data and therefore the only group able to assess the safety of study subjects based on those data; they are also free of any major conflicts of interest. Keeping them blinded to actual study treatments may limit their ability to recognize emerging problems. In many trials, differing rates of specific adverse events will reveal the actual treatments; trying to maintain blinding of a DSMB by using different codes for safety and efficacy outcomes will prevent the DSMB from making benefit-to-risk assessments, which is their primary responsibility. Additionally, there have been cases where errors in coding have been discovered when patterns of adverse events shown to the DSMB are implausible based on known mechanisms of the treatments under study. Blinding of the DSMB in any way reduces its ability to ensure subject safety and study integrity.
The IRB—and for FDA-regulated trials, the FDA—must be aware of the DSMB’s recommendations, even if the sponsor does not agree. The sponsor may have reasonable justification to not follow the DSMB recommendations, but that does not obviate the responsibility to report to the IRB and to FDA who have overlapping trial oversight responsibility.
SE: I agree that there is a rationale for such communication, but it is problematic as the investigator, who is typically the one to convey recommendations to the local IRB, will not be aware of the recommendation if it is not accepted. If the trial is continuing (and, as the commenter notes, that might actually be the right decision), knowledge that there was a DSMB recommendation that was not accepted may well influence the future conduct of the study, leading to biased or otherwise misleading results. In some cases, the sponsor may provide a satisfactory explanation to the DSMB for its nonacceptance of the recommendation; otherwise a DSMB, having pledged confidentiality, has little recourse other than resignation (which would not release them from their confidentiality agreement). There is ongoing debate about the recourse a DSMB has if a sponsor takes action that the DSMB disagrees with. HHS and FDA regulations do not address this question specifically. I note that as the use of single IRBs becomes more common, it may happen that sponsors circulate recommendations to the IRB rather than investigators, but it is still not clear that the sponsor would be required to report rejection of a DSMB recommendation to the IRB.
PRIM&R thanks Mr. Davis and Dr. Ellenberg for sharing their expertise!
The recording of those webinar is available for individuals to purchase in PRIM&R’s online store. If you would like to purchase the webinar for group viewing, please download the order form (PDF) and send it to firstname.lastname@example.org