Problems Spotted in Clinical Trials Can Go Unreported, Study Says
By Dennis Thompson
MONDAY, Feb. 9, 2015 (HealthDay News) -- Medical journals and the U.S. Food and Drug Administration rarely report violations of scientific conduct that federal regulators unearth during spot inspections of clinical trials, a new analysis shows.
In a study published in the Feb. 10 issue of JAMA Internal Medicine, a review of FDA inspection reports between 1998 and 2013 revealed nearly 60 clinical trials in which regulators had uncovered violations serious enough to earn the agency's most severe classification -- "official action indicated," or OAI, said study author Charles Seife, a professor at the Arthur L. Carter Institute of Journalism at New York University.
Seventy-eight articles were published based on data from these trials. But only three of them mentioned the violations that regulators found, Seife and a team of graduate students determined. The violations included fraud, incompetence or misconduct.
"Your physician makes decisions based upon what he knows from the peer-reviewed literature," Seife said. "If that literature is tainted in some way without the physician knowing, then the physician is making treatment decisions based upon bad data."
The good news is that such violations are rare, according to Seife. During the 2013 fiscal year, about 2 percent of the nearly 650 inspections the FDA carried out at clinical trial sites were classified as OAI, according to background information in the study.
But when these violations are uncovered, they can be serious enough to undermine the findings reported from the clinical trial:
- A patient in a trial had a foot amputated two weeks after receiving stem cells to treat poor blood flow in a leg. Despite this, the article from the stem cell trial claimed that all of the patients reported improvement in their limbs.
- The FDA deemed an entire trial devoted to the blood-thinning drug rivaroxaban unreliable, due to systematic and widespread scientific fraud. Despite this, the findings from the trial were published in a journal with no mention of the FDA's condemnation.
- A patient died in a chemotherapy trial due to a researcher's falsification of documents, and the researcher ended up in prison after pleading guilty to fraud and criminally negligent homicide. None of the articles published from this clinical trial mention either the fraud or the subsequent homicide case.
"Is this fraud that is definitely affecting all the drugs in your medicine cabinet?" Seife said. "No. But is this an issue that is affecting the quality of the peer-reviewed literature? You definitely can say that."
The FDA inspections uncovered one or more violations in the 57 published clinical trials identified by Seife and his students, including:
- 22 trials with falsified information.
- 14 trials where researchers failed to report adverse events.
- 42 trials with violations of the trial's protocols.
- 35 trials with record-keeping errors.
- 30 trials in which researchers failed to protect patient safety or acquire informed consent.
Seife said his team likely could have uncovered even more cases, but most of the documents his students obtained from the FDA were heavily redacted. "In some cases, you can't even tell which drug is being tested," he said.
The research revealed a need for improved transparency from FDA regulators, researchers and medical journals, said Dr. Robert Steinbrook, editor-at-large at JAMA Internal Medicine and an adjunct professor of internal medicine at the Yale School of Medicine in New Haven, Conn.
Steinbrook noted that the FDA set up an online database in 2014 that provides access to basic information about its clinical trial inspections from 2008 to 2014.
However, the database does not link to any inspection documents, and in some cases does not include enforcement action, he said. There's also no easy way to track a clinical trial listed on the database to a journal article based on that trial.
"When the agency does an inspection at a clinical trial site and issues an inspection report with serious findings, it should be easier to identify the specific trial and the specific publication, if there was a publication," he said. "We think the fact there is a database is a good step, and we suggest that the agency can now move beyond that and do some other things."
Researchers and the editors of medical journals also have an obligation to inform the public about such problems, either before or after publication, Steinbrook added.
"Not every FDA finding should be included in a report of a research study, but some are serious and would best be brought to the attention of editors and the public," he said.
For more information on clinical trials, visit the U.S. National Institutes of Health.