Physicians at the Yale School of Medicine are calling for improvements in the way the Food and Drug Administration informs the medical community about research misconduct.
In an editorial published in the Journal of the American Medical Association Internal Medicine, physicians used data collected by journalist and New York University professor Charles Seife GRD ’95 to show that 57 of the 600 clinical trials published from January 1998 and September 2013 showed significant evidence of research misconduct. That misconduct ranges from outright fraud — such as a case in which a study reported an X-ray that never occurred — to instances where patients are not informed of the risks and benefits of a study. The FDA does not report research misconduct directly to the journals that publish the studies, said Robert Steinbrook, professor of medicine and associate editor for JAMA Internal Medicine. Though rare, misconduct does not get reported when identified because there is no systematic way for the FDA to inform the medical community, Seife said.
“The FDA needs to stop treating wrongdoing as confidential information,” Seife said. “Treating it as such is covering up fraud.”
Of the total instances of research misconduct, 39 percent involved falsification or submission of false information, 74 percent involved protocol violation, 61 percent involved inadequate or inaccurate record-keeping and 53 percent involved a failure to protect the safety of patients and obtain informed consent.
Seife said he believes the FDA is making it difficult for journals and researchers to find out when research misconduct happens. Officials at the FDA think the disclosure of such information would damage the reputation of drug companies, he added.
Steinbrook offered the example of a clinical trial in which a patient’s leg was amputated two weeks after being told they were doing well. Despite being detected by the FDA, that information does not appear in the medical journal in which the study was originally published or anywhere outside the FDA’s internal documents.
Steinbrook warned that the FDA’s refusal to ensure that fraudulent data is identified may have unintended negative consequences for public health. If the information from a study is being used to decide which treatment is best for a patient and this information is flawed, it can lead to bad decisions, he said, adding that some studies that the FDA has deemed completely unreliable are still in circulation.
Seife said he can point to several data points on FDA-approved labels that are based on fabricated data. Of the clinical trials the FDA inspected and found to have “objectionable conditions or practices,” only 4 percent saw the objections reported in peer-reviewed literature, he added.
“We were very broad in our interpretation of any mention — even if it was hinted at, we would include it,” Seife said.
According to Steinbrook, the FDA does not make evidence of proven research misconduct readily available to journal editors. While the information is available via Freedom of Information Act requests, editors and peer reviewers cannot feasibly sift through the hundreds of pages of documentation that would tell them if misconduct had occurred.
It is difficult to identify fraud using this process, professor of medicine and public health and associate editor at JAMA Internal Medicine Joseph Ross MED ’06 added.
“The FDA documentation is challenging to dig though. With every new drug approval, there is a dossier spanning hundreds of pages,” Ross said
He added that the question of whether or not misconduct has occurred does not produce a quick yes or no answer because there are so many complex questions involved.
Seife said he thinks the FDA’s lack of transparency may be partially explained by conflicting interests. The FDA has an interest in making sure drugs on the market are effective, but the agency is also under pressure from Congress, drug manufacturers and the public to approve drugs quickly, he added.
“People are hammering down the door of the FDA, asking them to ‘approve our drugs,’ ” Seife said.
When patients are sick and there is a drug with the potential to improve their condition, he explained, the agency is under pressure to make approval process swift.
Drug companies must pay the FDA in order for their drugs to go through the approval process, and members of the medical community believe that this makes the FDA more likely to approve drugs than they otherwise would be, Seife said. But he thinks that regulatory capture — when, over time, a regulatory body becomes more sympathetic to the industry it is overseeing — is a bigger problem.
The FDA did not respond to requests for comment.
Steinbrook said the FDA’s reluctance to be transparent about research misconduct leads to a broader problem of people not trusting the government’s medical advice.
Seife agreed, citing the growing anti-vaccine movement as an example.
“[The anti-vaccine movement], which I am very against, is predicated on the idea that the FDA is lying to us,” Seife said. He explained that the public’s trust in the agency has been undermined because of a track record that is far from perfect.
In his editorial, Seife recommended that journals require authors to disclose research misconduct, just as they already do with conflicts of interest.
“Fifteen years ago, no one really disclosed their conflict of interest, and now it is standard,” Seife said
Something similar could happen with research misconduct, he said.
FDA inspection involves officials visiting a trial site and auditing the records kept there.