In order to improve the integrity and transparency of medical research, there has been a growing movement toward prospective clinical trial registration. Universal registration of clinical trials can help detect and prevent selective (or distorted) reporting, which can include unspecified trial modifications and/or the addition of favorable (i.e., statistically significant) or removal of unfavorable outcomes within published studies. Our study published in Trials assesses trial registration among a large sample of clinical trials published in ten high-impact specialty society medical journals.
In 2005, the International Committee of Medical Journal Editors (ICMJE) adopted a policy to encourage prospective clinical trial registration. The ICMJE mandated that all clinical trials beginning July 2005 register at or before the time of first patient enrollment (i.e., “prospectively”) as a condition for publication in its member journals. Although thousands of journals now support the ICMJE manuscript publication guidelines, evidence suggests that journal editors and publishers often do not consistently adhere to the prospective trial registration policy.
Previous studies evaluating registration rates among trials published in the highest-impact general medical journals have suggested that that over one-quarter of published trials are registered retrospectively, which could potentially allow for study design modifications and outcome changing.
Although registration and prespecification of clinical research have become more common as a result of various initiatives, little is known about the registration of trials published among high-impact specialty society journals. These journals, which are generally administered by professional organizations (e.g., the American Neurological Association [ANA] and the American Association for the Study of Liver Diseases [AASLD]), often represent the views of their constituent specialists. Therefore, they are often the preferred source of clinical research and guidelines for specialists.
In our new Trials article, we aimed to evaluate adherence to ICMJE prospective trial registration policy. In particular, we were interested in determining how often trials were being registered late enough to potentially allow premature examination of collected data. Furthermore, we looked at characteristics associated with prospective registration and differences between registered and published primary outcomes.
We found that, among 487 trials published in the 10 highest-impact US medical specialty society journals between January 1, 2010 and December 31, 2015, nearly 10% were unregistered. Approximately 25% of all trials were registered retrospectively, of which the majority (68%) were registered late enough to have potentially permitted premature examination of primary outcome data. Overall, unregistered trials were more likely to report favorable results than were registered trials.
Journals, who ultimately disseminate research findings, play an important role in ensuring the integrity of published clinical research. Moreover, specialty society journals often publish trials that can have a significant influence on their targeted clinical readership. Therefore, these journals have the responsibility of minimizing reporting biases by ensuring adequate and timely registration of clinical trials.
In order to improve the current system, greater editor and journal scrutiny may be necessary when evaluating clinical trial manuscript submissions. Although journals generally require the disclosure of trial information as part of the peer-review process, studies suggest that potential differences between registered and reported materials do not influence decisions to publish manuscripts. According to a survey of editors from journals endorsing ICMJE guidelines, approximately two-thirds would consider publishing retrospectively registered trials.
Moving forward, it may be necessary to develop more detailed guidelines for journal editors, reviewers and authors when it comes to evaluating clinical trial registration. For instance:
- Within a submission, authors should be asked to clearly state the date or registration and if this is retrospective, it should be clearly indicated.
- During the peer-review process, editors and/or reviews should verify whether a trial has been registered prospectively or retrospectively.
- Editors and/or reviews should pay close attention to the significance of findings (i.e., reporting of favorable results) for unregistered or retrospectively registered trials.
- If journals elect to publish unregistered or retrospectively registered trials, steps should be taken to ensure that readers are aware of the fact that the trial was registered late, such as by including statements in the published article explaining why registration did not occur/was delayed.
- If journals elect to publish unregistered or retrospectively registered trials, editors should request original trial protocols, along with their revisions. When protocols are not provided, published statements should explain why they were not available.
Perhaps these efforts could help mitigate the potential negative impact of substandard prospective clinical trial registration. Overall, our study suggests that greater adherence to ICMJE’s prospective trial registration policy may be necessary to improve the integrity and transparency of clinical research.
Dr. Joshua D. Wallach is a research fellow at Yale University, working within the Collaboration for Research Integrity and Transparency (CRIT) and the Center for Outcomes Research and Evaluation (CORE).
Dr. Anand D. Gopal is a preliminary intern in Medicine at Pennsylvania Hospital of the University of Pennsylvania.
Dr. Joseph S. Ross is an Associate Professor of Medicine (General Medicine) and Public Health (Health Policy and Management)
The work discussed in this Blog was supported in part by the Laura and John Arnold Foundation to support the Collaboration for Research Integrity and Transparency (CRIT) at Yale.
This post originally appeared on BioMed Central Blogs Network