The Crucial Role of the Little-Known Panels
Like other large clinical trials involving potentially high-demand or controversial products, studies of COVID-19 vaccines in most countries are supervised by groups of independent observers. Known in the United States as data safety and monitoring boards (DSMBs), and elsewhere as data monitoring committees, these panels consist of scientists, clinicians, statisticians, and other authorities with no ties to the sponsor of the study.
The six trials funded by the federal program known as Operation Warp Speed (including those of newly approved Moderna and frontrunner AstraZeneca) share a DSMB, whose members are selected by the National Institutes of Health; other companies (including Pfizer) appoint their own. The panel's job is to monitor the safety and efficacy of a treatment while the trial is ongoing, and to ensure that data is being collected and analyzed correctly.
Vaccine studies are "double-blinded," which means neither the participants nor the doctors running the trial know who's getting the real thing and who's getting a placebo. But the DSMB can access that information if a study volunteer has what might be a serious side effect—and if the participant was in the vaccine group, the board can ask that the trial be paused for further investigation.
The DSMB also checks for efficacy at pre-determined intervals. If it finds that the vaccine group and the placebo group are getting sick at similar rates, the panel can recommend stopping the trial due to "futility." And if the results look overwhelmingly positive, the DSMB can recommend that the study sponsor apply for FDA approval before the scheduled end of the trial, in order to hurry the product to market.
With this kind of inside dope and high-level influence, DSMBs could easily become targets for outside pressure. That's why, since the 1980s, their membership has typically been kept secret.
During the early days of the AIDS crisis, researchers working on HIV drugs feared for the safety of the experts on their boards. "They didn't want them to be besieged and harassed by members of the community," explains Susan Ellenberg, a professor of biostatistics, medical ethics and health policy at the University of Pennsylvania, and co-author of Data Monitoring Committees in Clinical Trials, the DSMB bible. "You can understand why people would very much want to know how things were looking in a given trial. They wanted to save their own lives; they wanted to save their friends' lives." Ellenberg, who was founding director of the biostatistics branch of the AIDS division at the National Institute of Allergy and Infectious Diseases (NIAID), helped shape a range of policies designed to ensure that DSMBs made decisions based on data and nothing else.
Confidentiality also shields DSMB members from badgering by patient advocacy groups, who might urge that a drug be presented for approval before trial results are conclusive, or by profit-hungry investors. "It prevents people from trying to pry out information to get an edge in the stock market," says Art Caplan, a bioethicist at New York University.
Yet the COVID crisis has spurred calls for DSMBs to come out of the shadows. One triggering event came in March 2020, when the FDA approved hydroxychloroquine for COVID-19—a therapy that President Donald J. Trump touted, despite scant evidence for its efficacy. (Approval was rescinded in June.) If the agency could bow to political pressure on these medications, critics warned, it might do so with vaccines as well. In the end, that didn't happen; the Pfizer approval was issued well after Election Day, despite Trump's goading, and most experts agree that it was based on solid science. Still, public suspicion lingers.
Another shock came in September, after British-based AstraZeneca announced it was pausing its vaccine trial globally due to a "suspected adverse rection" in a volunteer. The company shared no details with the press. Instead, AstraZeneca's CEO divulged them in a private call with J.P. Morgan investors the next day, confirming that the volunteer was suffering from transverse myelitis, a rare and serious spinal inflammation—and that the study had also been halted in July, when another volunteer displayed neurological symptoms. STAT News broke the story after talking to tipsters.
Although both illnesses were found to be unrelated to the vaccine, and the trial was restarted, the incident had a paradoxical effect: while it confirmed for experts that the oversight system was working, AstraZeneca's initial lack of candor added to many laypeople's sense that it wasn't. "If you were seeking to undermine trust, that's kind of how you would go about doing it," says Charles Weijer, a bioethicist at Western University in Ontario, who has helped develop clinical trial guidelines for the World Health Organization.
Both Caplan and Weijer have served on many DSMBs; they believe the boards are generally trustworthy, and that those overseeing COVID vaccine trials are performing their jobs well. But the secrecy surrounding these groups, they and others argue, has become counterproductive. Shining a light on the statistical sausage-makers would help dispel doubts about the finished product.
"I'm not suggesting that any of these companies are doing things unethically," Weijer explains. "But the circumstances of a global pandemic are sufficiently challenging that perhaps they ought to be doing some things differently. I believe it would be trust-producing for data monitoring committees to be more forthcoming than usual."
Building Trust: More Transparency
Just how forthcoming is a matter of debate. Caplan suggests that each COVID vaccine DSMB reveal the name of its chair; that would enable the scientific community, as well as the media and the general public, to get a sense of the integrity and qualifications of the board as a whole while preserving the anonymity of the other members.
Indeed, when Operation Warp Speed's DSMB chair, Richard Whitley, was outed through a website slip-up, many observers applauded his selection for the role; a professor of pediatrics, microbiology, medicine and neurosurgery at the University of Alabama at Birmingham, he is "an exceptionally experienced and qualified individual," Weijer says. (Reporters with ProPublica later identified two other members: Susan Ellenberg and immunologist William Makgoba, known for his work on the South African AIDS Vaccine Initiative.)
Caplan would also like to see more details of the protocols DSMBs are using to make decisions, such as the statistical threshold for efficacy that would lead them to seek approval from the FDA. And he wishes the NIH would spell out specific responsibilities for these monitoring boards. "They don't really have clear, government-mandated charters," he notes. For example, there's no requirement that DSMBs include an ethicist or patient advocate—both of which Caplan considers essential for vaccine trials. "Rough guidelines," he says, "would be useful."
Weijer, for his part, thinks DSMBs should disclose all their members. "When you only disclose the chair, you leave questions unanswered," he says. "What expertise do [the others] bring to the table? Are they similarly free of relevant conflicts of interest? And it doesn't answer the question that will be foremost on many people's minds: are these people in the pocket of pharma?"
Weijer and Caplan both want to see greater transparency around the trial results themselves. Because the FDA approved the Pfizer and Moderna vaccines with emergency use authorizations rather than full licensure, which requires more extensive safety testing, these products reached the market without the usual paper trail of peer-reviewed publications. The same will likely be true of any future COVID vaccines that the agency greenlights. To add another level of scrutiny, both ethicists suggest, each company should publicly release its data at the end of a trial. "That offers the potential for academic groups to go in and do an analysis," Weijer explains, "to verify the claims about the safety and efficacy of the vaccine." The point, he says, is not only to ensure that the approval was justified, but to provide evidence to counter skeptics' qualms.
Caplan may differ on some of the details, but he endorses the premise. "It's all a matter of trust," he says. "You're always watching that, because a vaccine is only as good as the number of people who take it."
Scientists have been using MRI images (like the ones pictured here) to train artificial intelligence to detect cancers earlier and with more accuracy. Here, MIT's AI system, MIRAI, looks for patterns in a patient's mammograms to detect breast cancer earlier than ever before. news.mit.edu