Two leading medical journals announced on Thursday that they would begin publishing analyses that take advantage of the growing availability of unreported study data to rebut the conclusions presented in previous research findings.
The journals, BMJ and PLoS Medicine, made the commitment along with a small group of researchers who said they already have collected some 178,000 pages of companies’ previously confidential research documents related to published studies.
The effort is an attempt to take a more aggressive approach to the problem of studies in which an industry sponsor has either hidden or distorted unwelcome findings concerning a medical discovery, said one organizer, Peter N. Doshi, a postdoctoral fellow in comparative-effectiveness research at the Johns Hopkins University.
“We can’t have publications that are just about, ‘Trust me, here’s what I found,’” Mr. Doshi said.
Under their plan, Mr. Doshi and his colleagues will begin contacting companies to let them know about medical-trial data that contradict published research related to the trial. The companies will be given 30 days to commit to producing a new analysis for publication in a journal that reflects the newly discovered data.
If a company declines, Mr. Doshi and his team will work to find a researcher willing to produce an independent assessment of the data. BMJ and PLoS Medicine will consider publishing such analyses, their editors said in an editorial published on Thursday in BMJ.
The idea stems from years of failure by journals and others to take meaningful action against the problem of hidden or misrepresented clinical-trial results, the editors said.
“It is one of the leading scientific problems of our time, but few solutions have been put forward,” wrote the editors, Elizabeth Loder and Fiona Godlee of BMJ and Virginia Barbour and Margaret Winker of PLoS Medicine.
The Reality of WikiLeaks
There have been other attempts to crack down on the problem, such as asking journals to publish findings only when the trial and its planned objectives have been declared and registered in advance. But that approach has suffered from low rates of compliance.
Also on Thursday, officials of the National Institutes of Health outlined their latest efforts to crack down on the general problem of research results that do not prove to be reproducible. The steps, outlined by Story Landis, director of the National Institute of Neurological Disorders and Stroke at a meeting of an NIH advisory panel, include raising awareness of the problem within the research community and expanding training.
One member of the advisory panel, Reed V. Tuckson, executive vice president and chief of medical affairs at the United Health Group, said it wasn’t clear to him that the strategy was proving adequate.
Mr. Doshi said his collection of documents came from various sources, often evidence produced for court cases or the results of open-records requests posed to European science agencies. The U.S. Food and Drug Administration is rarely a source because it has strict rules against the release of data from medical trials, he said.
Although confrontational, the idea seems worthwhile if the data sources prove reliable, said Arthur L. Caplan, a professor of bioethics at New York University.
It’s also a warning to drug companies and medical journals—similar to that posed to other institutions by figures such as Julian Assange, the founder of WikiLeaks, and Edward Snowden, who blew the whistle on the National Security Agency’s snooping on private citizens—that they should be increasingly aware that, in today’s world of computerized databases, any hidden secrets may eventually become public, Mr. Caplan said.
“So get ready to deal” with that reality, said Mr. Caplan, head of the Division of Medical Ethics at NYU’s Langone Medical Center.
A Question of Retraction
One of the cases described by Mr. Doshi and his colleagues, writing in a separate article on Thursday in BMJ describing his plan, is a controversial study concerning the antidepressant drug Paxil, promoted by GlaxoSmithKline, its manufacturer. Critics have complained for years that the study dangerously misrepresented data and hid information indicating that the drug promoted suicidal behavior among teenagers.
The company, a group of university authors, and the journal that published the study, the Journal of the American Academy of Child and Adolescent Psychiatry, all declined to retract the study, even after federal prosecutors last year identified it as a focal point in their largest-ever health-care-fraud settlement.
Mr. Doshi said he didn’t expect his approach would necessarily force a retraction in such cases. But he said he hoped that producing another published analysis of the medical-trial data would mean there was another peer-reviewed piece of literature for doctors to consider when deciding whether to prescribe a particular drug.
At this early stage, it remains unclear how many researchers will voluntarily help produce such analyses, Mr. Doshi said. His hope, however, is that public and private agencies that finance research will see the value in supporting such work. “This is an organic movement that we’re trying to start here,” he said.
