Charles Seife (Sigrid Estrada, New York University)
10 February 2015. An investigation by a New York University journalism professor shows serious violations in clinical trials revealed in Food and Drug Administration inspections are largely not mentioned in reports of those trials in scientific journals. The findings of Charles Seife and his team from NYU’s journalism school were published yesterday in the journal JAMA Internal Medicine.
FDA routinely inspects clinical trial sites to verify data in reports from researchers, but also in response to complaints about studies, concerns raised by sponsors, or early closings of study sites. Inspections can be announced or unannounced, and often involve visits to trial sites and inspections of documents. In drug trials, trial managers are required to keep records for two years after a drug is approved or a clinical trial is discontinued and FDA notified of its closing.
Inspections result in one of three results:
– No action indicated, where no violations are found
– Voluntary action indicated, where inspectors find violations of good practice, but are not serious enough to require sanctions
– Official action indicated, where violations are considered serious enough to draw sanctions
In fiscal year 2013, FDA inspectors made 644 inspections of clinical trials and found 56 percent with no violations, 42 percent required voluntary actions, and 2 percent drew official-action sanctions.
Seife and colleagues looked for instances where serious — official action — sanctions resulted from inspections, and whether those violations were reported in the published literature. The investigators gathered evidence from FDA public archives covering January 1998 through September 2013, collecting documents such as warning letters, disbarments, and notices of hearings or disqualifications. These archive searches were supplemented with Google searches of the fda.gov domain and Freedom of Information Act requests.
The collections yielded 421 reports of inspections resulting in official-action sanctions. The NYU team was able to link the 421 inspection reports with 101 clinical trials, of which about two-thirds (68) were reported in scientific journals. Of those 68 trials, 57 could be linked directly to specific violations cited in inspection documents. Those violations include under-reported adverse events, recruitment guideline violations, and other forms of scientific misconduct.
Among those 57 clinical trials, the investigators found more than half (53%) of the violations were related to the safety and welfare of patients in clinical trials, including issues with informed consent. Some 4 in 10 violations (39%) were for submission of false information, and another 25 percent resulted from problems in adverse events reports. Most violations were for failure to follow the trial’s plan or protocols (74%) and inadequate or inaccurate record-keeping (61%).
FDA inspectors generally provide clinical trial site managers with details of violation reports, known as form 483, soon after the inspections. Despite these notices given to researchers, few published results of clinical trials in scientific journals mention these violations. Seife and colleagues found the 57 clinical trials with specific FDA violations yielded 78 publications in peer-reviewed journals. And among those 78 articles, only 3 mentioned the FDA violations, despite in 59 of those 78 cases, the inspections were completed 6 months or more before the journal publication.
In the JAMA Internal Medicine article, Seife cites four examples of violations that went unreported in the peer-reviewed journal articles, mainly falsifying reports of adverse, blinding and randomization errors, and systematic discarding of medical records. In one case, a falsified lab test result of impaired kidney functions led to a patient’s death, and a subsequent guilty plea to fraud and criminally negligent homicide. None of those violations or eventual court actions were reported in the journal articles for those clinical trials.
Seife acknowledges that the cases he and his team reviewed are not meant to be representative of all clinical trials, and were limited by the availability of records and in some cases heavy redactions of documents to protect confidential information. Nonetheless, the lack of violation reports in published findings are a problem that should be addressed by FDA and authors of journal articles. Seife and colleagues recommend immediate publication of violations in trial records on clinicaltrials.gov, or in a separate online database. In addition, authors of journal articles should be required to reveal inspection violations, much like current disclosures of financial conflicts of interest.
Read more:
- FDA Issues Draft Wellness Device Guidance
- FDA Issues Draft Industry Guidance on Social Media
- FDA Proposes Accelerated Medical Device Approval Process
- Biotech Group Issues Clinical Trial Data Sharing Guidelines
- Many Large Clinical Trials Remain Unpublished
* * *
RSS - Posts
You must be logged in to post a comment.