A panel of experts charged with watching patient safety during a 1999 Vioxx study may have had conflicts of interest that prevented it from putting patient interests first. A later study showed the painkiller could cause heart problems and death. That 2004 study was immediately stopped by its safety panel, and Vioxx was removed from the market worldwide.
To learn more about data safety monitoring boards and their role in protecting patients who participate in drug studies, NPR turned to statistician David DeMets. He's the chairman of the Department of Biostatistics and Medical Informatics at the University of Wisconsin, and the author of several books on the subject, including Data Monitoring in Clinical Trials: A Case Studies Approach (Springer 2005).
DeMets says the current watchdog system for patient safety is a good one, but there are practical limits on the extent to which drugs can be monitored:
What exactly is a Data Safety Monitoring Board?
One of the reasons we conduct clinical trials is to see if a new intervention -- a drug or a medical device, for example -- works, or sometimes whether it works better than an old intervention.
So suppose we wanted to see which of two drugs (a new one and an older one) works better to treat high blood pressure. We would give people in the trial one of the two drugs, and then check their blood pressure to see whether the drug works. We would also collect information about any side effects that are observed. Participants getting the new drug would be compared with those getting the older standard drug.
You can imagine that when you have hundreds or thousands of people in a trial, you get a lot of data. So the idea is for a group of experts to look over all this data periodically and make sure that it is safe and ethical for the trial to continue. This group of experts -- made up of physicians, statisticians and sometimes medical ethicists -- is called a Data Safety Monitoring Board (DSMB).
What sorts of decision can a DSMB make?
After the DSMB members look over all this data, they can recommend that the trial be continued as planned, be modified or be terminated.
Trials aren't always terminated because a drug causes harm. They can be terminated early because one drug works a lot better than the other, one drug has too many side effects, or if we can't answer the question of which drug works better.
Who chooses the DSMB members?
Either the trial sponsor -- which is usually the drug company or federal agency paying for the trial -- or the steering committee of the trial.
The steering committee typically is made up of senior participating investigators from universities or clinical centers and representatives of the sponsor. This team is generally responsible for the design and conduct of the trial. Whether the safety-panel appointments are made by the sponsor or the steering committee, both need to agree on the DSMB membership. The appointed DSMB members must be independent of the trial, i.e. not participating, and have no obvious conflicts of interest.
What are examples of conflicts of interests that should be avoided by DSMB members?
Conflicts of interest aren't always easy to define. They can be both financial and intellectual. If you work for the sponsor, even if you're not involved in the study, that would be an obvious conflict to avoid. If you were on the speakers circuit for the sponsor on its product, that might be viewed as a conflict.
Where it gets more challenging is when you have a grant or contract with any company or agency to do research in the same area as the one under investigation in the clinical trial. You would have to take a careful look at those relationships.
On the one hand, you want DMSB members to have experience and knowledge about the research, which means they are probably conducting research on this topic funded by some sponsor, perhaps even a competitor.
On the other hand, if the relationship with the sponsor is too close, the DSMB member might not be perceived as being independent. Where to draw the line is not always clear.
If absolute separation were demanded, the level of expertise on the DSMB might suffer and that would not be good for trial participants, investigators or the sponsor.
Are DSMBs required for drug studies in the United States?
The Food and Drug Administration (FDA) does not require formal DSMBs for all U.S. clinical trials, but it does require all trials to have some form of a safety-monitoring plan. For trials without an independent, formal DSMB, the safety plan is overseen by the trial investigators.
The FDA currently recommends DSMBs for trials where a lot is at stake: Patients are seriously ill, or we're trying to determine whether the new drug prevents events like death or cancer recurrence. Many, but not all, trials meet those criteria.
Should the FDA make its guidelines mandatory, and require DSMBs for all trials?
Probably not, for practical reasons. We don't have enough people who have the experience and training to serve on all the DSMBs that would be needed. We'll get there with time, but not right now.
So we're picking and choosing the trials for now, and I think the FDA has given a pretty good suggestion of when a formal, independent safety committee would be a good idea.
Do trial participants need to take additional steps to protect themselves?
Patients might ask whether a DSMB has been appointed for their trial or at least be aware of what the safety-monitoring plan is. Trial participants should also know that they have the right to withdraw from a trial at any time, a point made in the consent form.
Considering what some see as a failure of the Merck safety panel overseeing the 1999 Vioxx VIGOR trial, does this system of protecting patients in clinical trials need reform?
I don't know enough details about the VIGOR trial to comment about it. However, based on my experience over the past 30 years of being on DSMBs, I think they are doing a good job at what they're designed to do.
A lot of the controversy and discussion going on now is about what you do when the trial is over and the therapy is approved. How do you track the long-term incidence of serious adverse effects? We're all struggling to sort that out, and I don't think anyone as yet has the perfect solution.
One of the leading suggestions is to have a post-approval system that requires serious adverse events to be reported to the FDA once the drug is on the market. Suggestions have been made that it should be mandatory that for the first 10,000 patients or even 100,000 patients, you must have a follow-up so you know something about side effects that are too rare to be seen in studies done before the drug is marketed. In those studies, drugs are tested, at most, in a few thousand patients.
Right now, the system is passive. Once a drug reaches the market, we have to rely on someone having enough interest or concern to fill out a report on an adverse event and send it in to the FDA.