The trusted source for
healthcare information and
Survey shows DSMBs' structure, operations
IRB members among those who say there should be tougher requirements
A data and safety monitoring board (DSMB) can be crucial to protecting participants in higher-risk studies, by reviewing the accumulating study data for potential emerging risks and if necessary, by recommending that a study be changed or even halted as a result.
While the FDA and NIH have created guidelines for setting up and operating DSMBs, they do not address all of the varied ethical issues that arise in trials that involve substantial risks, says Patti Tereskerz, JD, PhD, an associate professor of medical education and director of the Program in Ethics and Policy in Healthcare at the University of Virginia in Charlottesville.
Tereskerz says there's also a lack of information about how DSMBs are structured and how they operate. She and her colleagues set out to fill that void by surveying past and current members of DSMBs, as well as investigators and IRB community members, to see not only how DSMBs operate, but how participants thought they should be operating.
What they found were some significant differences between what participants thought should be occurring and how DSMBs actually worked – in training, in membership, in conflict of interest rules and other areas.
She says IRBs should take note of these discrepancies when they're reviewing data and safety monitoring plans for clinical trials.
"They should look carefully where there is a difference between what investigators and statisticians thought should be happening and what is happening," Tereskerz says. "And maybe they should push for some more of the requirements that these folks think should be in place."
She notes that the IRB community members surveyed also differed from investigators and statisticians in how they thought DSMBs should be run, nearly always recommending tougher requirements than the other groups.
"I think that's an important take-home message – that IRB members wanted more stringent requirements."
Results from the survey were published in a recent issue of the journal Accountability in Research.
IRBs as public proxy
Tereskerz says that the community IRB members included in the study actually were intended to serve as a proxy for public opinion, rather than for any technical expertise. In fact, nearly 30 percent of the 95 community IRB members surveyed had never heard of a DSMB prior to this study.
But because of their service on IRBs, they were likely to be much more conversant with research ethics issues than the average person and so could credibly represent the public's opinion, Tereskerz says.
"We realized if you just did a survey of the general public, they probably wouldn't have any idea of what you're talking about," she says. "We wanted someone who would be representative of the community, but at the same time had sufficient knowledge of what data and safety monitoring involved and that's how we came up with IRBs."
Those who actually served on DSMBs and the investigators who worked with them were asked about the structure of those boards DSMBs:
Were members trained, and were they paid? Did membership include a community representative, an attorney or a bioethicist?
Were minutes kept of meetings?
Were there conflict of interest rules? Were sponsors and statisticians blinded? Was an independent biostatistics center used?
Were stopping rules established? Were there rules in case members or the sponsors disagreed?
Tereskerz's team also asked all of the participants, including the community IRB members about their opinions regarding all those issues.
In every case, actual DSMB practices failed to live up to the ideals set by the participants.
For example, while all the participants believed that people should be trained or required to serve an apprenticeship before serving as a DSMB member, only 15 percent of those who served actually were trained or apprenticed prior to their most recent DSMB service.
All but one of the 267 participants thought a DSMB should have conflict of interest rules, but only 55 percent of members reported that their most recent DSMB had such rules.
Two-thirds of those surveyed thought there should be a bioethicist on a DSMB, but only 16 percent reported that a bioethicist had served on their own DSMB.
Compared to the other survey participants, community IRB members were more likely to want members to be trained, an independent biostatistics center to be used and for there to be a community member on the DSMB.
They were less likely to ask for sponsor and statistician blindedness. Tereskerz says she's unsure why the issue of blindedness was not as important to IRB members as it was to statisticians and investigators. She does note that investigators are acutely aware of conflicts of interest that can arise in an industry-sponsored study.
"I'm not sure that IRB members are as aware of those issues in terms of conflict of interest," she says.
Conflict over IRB participation
Tereskerz notes that others involved in the study were somewhat critical of her plan to interview IRB members for this survey.
"I actually got pushback about including IRB members – I had to fight for that," she says. "The argument was that they don't have the technical knowledge, investigators should decide.
"And I said, wait a minute, this is research, this is about trust. You're talking about clinical trials, adverse events, stopping or not stopping a study," Tereskerz says. "If you want people to enroll in research, you have to have trust."
Tereskerz would like to see IRBs use this information to ask more pointed questions about data and safety monitoring plans during review. But ultimately, she says, the best way to address this issue would be to strengthen guidance about the structure and operation of DSMBs.
"You can use this data to say look, there's a discrepancy between what experts think should happen and what is happening," she says. "Don't we need to modify or begin to produce guidance that addresses some of these issues?"
Tereskerz PM, Guterbock TM, Kermer DA, et al. An Opinion and Practice Survey on the Structure and Management of Data and Safety Monitoring Boards. Account Res 2011 Jan;18(1):1-30.