Unmasking 'Invisible' Drug Trials

Peter Doshi

Photo Courtesy of Peter Doshi

Publish your data, or else we will—that's the stark warning to drug companies in a new proposal released today. Peter Doshi (shown right), a postdoctoral fellow at Johns Hopkins University in Baltimore, Maryland, and his colleagues are fed up that only about half of all clinical trials are published. They want to change that, by convincing researchers and journals to print data that have been publicly released through other means, such as litigation and Freedom of Information Act requests, but, practically speaking, are sitting dormant in the filing cabinets or computers of individual scientists.

The unusual proposal is called RIAT, for Restoring Invisible and Abandoned Trials. It was published today in BMJ and also endorsed by PLOS Medicine. Doshi, who studies comparative effectiveness research, came up with the idea when his colleague, Swaroop Vedula, was analyzing reporting biases involving the drug gabapentin. Gabapentin's maker Pfizer had been sued for the way in which they marketed the drug for unapproved indications. During litigation, Pfizer had released thousands of pages involving gabapentin trials, and Vedula was poring through them. (One of the authors of the RIAT paper, Kay Dickersin, served as an expert witness against Pfizer in gabapentin litigation.) Pfizer had published only 12 of its 20 trials in gabapentin. But Doshi's center at Hopkins had the clinical study reports detailing the results of the other eight.

At the time, "it just hits me," Doshi says. "Why are we still referring to these as unpublished trials? Why aren't we publishing them ourselves?"

And so RIAT was born. Doshi can't say how much other unpublished drug data has been disseminated outside the pharmaceutical companies, but he knows it's substantial. An effort by the European Medicines Agency to share clinical trials data upon request has led to the release of 1.9 million pages, although it's since been curtailed by two lawsuits. The Hopkins group alone has 178,000 pages of data on drugs including paroxetine (the antidepressant Paxil), quetiapine (the antipsychotic Seroquel), and oseltamivir (the influenza drug Tamiflu), in addition to gabapentin. "We are hoping that others in a similar situation … are holding on to public data," Doshi says.

In their BMJ paper, he and his colleagues set forth a series of steps that researchers can follow to get the data out in the public domain. First, they need to confirm that the company hasn't already published it, or that if it is published the descriptions are misleading. In those cases, researchers should notify the company and give it 30 days to respond. If the company agrees to publish the work itself, it's given a year to do so. If that doesn't happen—or if the researchers are brushed off or get no response—they should contact a RIAT-friendly journal about publishing the work themselves.

The plan's success depends on journal participation, and Doshi is hoping that more will sign up. Its authors acknowledge that it's an unorthodox way to think about publishing. "Some people may think that publications based on clinical study reports with which the authors have no connection is equivalent to intellectual property theft, but you cannot steal what is already in the public domain," they write. That said, the authors can't definitively say if there are any legal issues that might stymie publication, even if the documents are public.

In an accompanying editorial, editors from PLOS Medicine and BMJ call the effort a "bold remedy," and "another step on the road towards a complete and unbiased account of the effectiveness and safety of medical interventions."

Follow News from Science