By Gary Evans, Medical Writer
In what amounts to a stunning compromise of research principles, a recent review of clinical trials over a five-year period found that results were never published for almost 90,000 participants.1
The authors of the study — who could not be reached for comment — looked at large, unpublished clinical trials on ClinicalTrials.gov from 2007 to 2012. The researchers found 67 unreported trials with a median enrollment of 765 people. The numbers ranged from 511 to 11,000 research participants, resulting in a total of 87,883 people with unreported results for a median of nine years after the trial was completed.
“I think there needs to be some mechanism of follow-up on these studies. They are experiments on people — and not reporting [them] is academic misconduct,” says Harlan M. Krumholz, MD, SM, director of the Center for Outcomes Research and Evaluation at Yale-New Haven (CT) Hospital.
Krumholz wrote an accompanying editorial2 on the study along with co-author Joshua D. Wallach, MS, PhD, an assistant professor of epidemiology at the Yale School of Public Health.
“When trial results are not made publicly available for years after study completion, patients, institutional review boards, clinicians, researchers, and the public must rely on incomplete evidence, which may lead to misconceptions about the efficacy and safety of interventions,” Krumholz and Wallach wrote.
IRB Advisor asked Krumholz whether IRBs should emphasize publication in research oversight by, for example, making reporting of findings part of the informed consent process.
“Yes,” he says. “For people not reporting results, future research studies should not be approved until the results are reported.”
Not reporting research data “biases the medical literature and potentially leaves the wrong impression of what is true,” Krumholz says. “It chills participant enthusiasm for research and betrays people’s trust in the system.”
The latter point is well taken considering the ongoing recruitment in the National Institutes of Health’s All of Us trial, which had pushed transparency and research results as a major benefit of participation in trying to reach 1 million people.
In addition to the common perception that the results were not what the researchers were looking for, there currently is little consequence for not publishing findings, Wallach says.
“To the best of my knowledge, no fines have been issued based on violations of reporting requirements,” he says. “For studies that are long unreported, there may be little incentive for authors to go back and share results.”
Lost Data in Key Fields
Beyond the ethical principles, such findings raise the question of whether meaningful data on a variety of diseases is lost to subsequent investigators. The unpublished trials reported by Tatsioni et al included research on neurological diseases, cardiology, infections, psychiatrics, and women’s health.1
“Without reported results and systematic reviews that can help guide practice and policy, we will not be able to evaluate and combine the results from all study sources,” Wallach says. “Participants are likely to assume that the evidence that is generated from a clinical trial will help inform practice, which helps justify the individual risks that may be involved with signing up for a study.”
The study by Tatsioni et al reinforces preceding evidence, suggesting unpublished research data is a longstanding problem. In particular, a previous study3 looking at large trials prior to 2009 found even more research participants whose results were never reported, Wallach says.
“In that study, they focused on trials with at least 500 participants, finding that nearly 30% remained unpublished — with an estimated total enrollment of nearly 300,000 patients,” he says. “This new study by Tatsioni et al focused on a more recent sample, and selected the largest studies that did not have results posted.”
- Tatsioni A, Karassa FB, Goodman SN, et al. Letters: Lost Evidence From Registered Large Long-Unpublished Randomized Controlled Trials: A Survey. Ann Intern Med 2019 May 7.doi: 10.7326/M19-0440. [Epub ahead of print]
- Wallach JD, Krumholz HM. Not Reporting Results of a Clinical Trial Is Academic Misconduct. Ann Internal Med 2019 May 7.doi: 10.7326/M19-1273. [Epub ahead of print]
- Jones CW, Handler L, Crowell KE, et al. Non-publication of large randomized clinical trials: cross sectional analysis. BMJ. 2013 Oct 29;347:f6104.doi:10.1136/bmj.f6104.