By: Sarah Wickline Wallan, MedPage Today
The vast majority of clinical trials fail to have their results posted to the main federal registry, Clinicaltrials.gov, within a year of completion as required by law, researchers found.
Out of more than 13,000 completed or terminated trials in the registry from 2008 to 2013 subject to the reporting requirement, result summaries for only 13.4% appeared there, according to Monique L. Anderson, MD, of Duke University, and colleagues.
Compared with NIH-supported trials, industry-funded trials were more likely to report results at the requisite 1-year mark (odds ratio 1.62, 95% CI 1.34-1.97), although more industry trials were exempt from the reporting, and these reporting differences vanished after 5 years, the researchers also found.
Trials supported by non-NIH government agencies or academic institutions were even less likely to report on time, compared with NIH-funded trials (OR 0.58, 95% CI 0.45-0.75) — and even after 5 years, these trials still lagged behind, the authors reported in the New England Journal of Medicine.
“The medical-products industry has been more responsive to the legal mandate of the FDAAA [Food and Drug Administration Amendments Act],” Anderson and colleagues wrote. “However, industry, the NIH, and other government and academic institutions all performed poorly with respect to ethical obligations for transparency.”
“The poor results reporting of NIH-funded clinical trials to ClinicalTrials.gov described in the new study is simply unacceptable,” Kathy Hudson, PhD, NIH Deputy Director of Science Outreach and Policy, said in an email to MedPage Today. “The study confirms the need for the NIH’s newly proposed policy that requires all NIH-funded investigators to submit clinical trial results information or risk losing their funding.”
Since 2007, as part of the FDAAA, sponsors of certain clinical trials have been required to register and report basic results on ClinicalTrials.gov.
This requirement applies to all non-phase I drug, including investigational new applications; medical device, including investigational device exemptions; or biologic clinical trials with at least one U.S. research site that had begun or were ongoing by the fall of 2007.
Unless a legal exemption exists, these types of trials must report prespecified primary outcome data within 1 year of the final data collection or termination of the trial. Penalties for failure to report include a “failure to submit” notification on the ClinicalTrials.gov, civil fines up to $10,000 per day, and loss of NIH funding.
Anderson’s team identified “highly likely applicable clinical trials (HLACTs)” that had begun by Jan. 1, 2008, and were either completed or terminated before Aug. 31, 2012 or filed for extension by Sept. 27, 2013.
Out of a potential 152,611 trials in the registry, from the U.S. and 185 other countries, 13,327 HLACTs were included in the analysis.
A sample review of 205 HLACTs revealed that 45% of industry-funded trials, 6% of NIH-supported trials, and 9% of government or academic institutions trials were not required to report results.
Only 1,790 (13.4%) of the trials reported results within 12 months of completion or termination. Even when the authors expanded the time frame to 5 years, only 38.3% of trials had reported results.
The primary purpose of 84.8% of the trials was “treatment.” For the remaining 15%, half of the trials listed “prevention,” and the other half listed diagnostic, screening, supportive care, health services research, or basic science as their primary purpose.
Broken down by type of intervention, 77.4% of the trials tested drugs, 11.9% tested devices, and 8.9% tested biologics. “Other” interventions like radiation or genetics comprised 1.9% of trials.
Industry backed 65.6% of the trials, the National Institutes of Health (NIH) funded 14.2% of trials, and other government agencies or academic institutions supported 20.2%.
Among the trials that reported results within the 12-month window, 17.0% were industry supported, 8.1% were sponsored by NIH, and 5.7% were funded by other government agencies or academic institutions.
For those trials that reported results within 5 years of completion or termination, 41.5% were industry supported, 38.9% were funded by the NIH, and 27.7% were supported by other government or academic institutions.
At the 5-year mark, after adjustments, little difference was found between the rates of reporting by industry-funded and NIH-supported trials (adjusted hazard ratio 0.97, 95% CI 0.89-1.07), but the difference still existed when compared with other government agency/academic institution-supported trials (adj.HR 0.62, 95% CI 0.56-0.70).
From 2008 through Sept. 27, 2013, 15,8% of the trials Anderson’s team included in the investigation had filed for a reporting extension. Among the extenders, 23.0% eventually reported results within 5 years.
The authors noted that phase II and phase III trials were more likely to submit exemption requests compared with trials in other phases. Compared with phase IV trials, phase II trials were less likely to report on time (OR 0.33, 95% CI 0.27-0.39), as well as phase III trials (OR 0.60, 95% CI 0.50-0.71).
Overall, the median reporting time was 17 months. For industry-funded trials, the median was 16 months. For NIH-supported trials, the median was 23 months. And for other government agency or academic institution trials, the median was 21 months.
Compared with trials with FDA oversight, those without it were less likely to report on time (adjusted OR 0.61, 95% CI 0.52-0.71). Timely reporting was associated with FDA oversight, later trial phase, and industry funding (P<0.001 for all comparisons).
In human trials, informed consent makes participants aware that the results of the trial will be reported, raising ethical issues with delayed reporting, the authors wrote.
For reference, the authors offered that Pfizer has said their results reporting takes anywhere from 4 to 60 hours to complete. “It is possible that the NIH and other funders have been unable or unwilling to allocate adequate resources to ensure timely reporting,” they wrote.
Anderson and colleagues also observed that penalties have never yet been imposed for violating the reporting requirement, because rulemaking for the reporting requirement is still underway (in fact, public comment on a proposed rule is open through March 23).
The authors noted study limitations included the possibility of outright missing data or misclassification of data.
To remedy low report rates, they suggested disallowing publication of trial results in peer-reviewed literature until results have been reported on the public registry.