Update on JAMA 's Policies on Conflicts of Interest, Trial Registration, Embargo, and Data Timeliness, Access, and Analysis

Journal policies define the principles, standards, and philosophy of a scientific publication and accordingly are living documents that require periodic reevaluation and updating. The goal of these policies is to ensure that published reports are unbiased so that patients, physicians, and others can have confidence in and benefit from the information. Over the past 3 decades, scientific evidence has accumulated that bias can occur due to several major factors, including conflicts of interest, failure to register clinical trials, deviation from trial protocols, and failure to include all relevant data in reports of trials and other major studies.

During the past year, we have engaged in a process of review of many of our major editorial policies. This evaluation has included JAMA editors, contributing editors, Editorial Board members, the Journal Oversight Committee, and selected members of the medical research and publishing communities. There has been substantial discussion and debate, but we have pursued this approach to ensure we meet our obligations as journal editors. In this Editorial, we clarify and reaffirm several of our major editorial policies.

Medical journals, like other entities such as academic medical centers and professional societies, occasionally encounter challenges with defining and reporting conflicts of interest (COI). Relevance of specific relationships, time frame for disclosure of specific interests, and the expected degree of inclusiveness of disclosure information vary among organizations and among journals. To help ensure consistency of COI reporting for authors, we will continue to require that each author of every accepted manuscript complete the International Committee of Medical Journal Editors (ICMJE) Form for Disclosure of Potential Conflicts of Interest.^1,2 This form has been adopted by the ICMJE journals and has been an important step toward standardizing the requirements for COI reporting.

However, the reporting process still requires clarification. For instance, section 2 of the ICMJE form is detailed and specific, inquiring about 7 types of potential COI relevant to the work under consideration for publication, with the time frame for reporting being “from the initial conception and planning to the present.” Section 3 is less clear, inquiring about 13 types of potential COI involving relevant financial activities outside the submitted work, with a 36-month time frame for reporting. Section 4 (“other relationships”) is the most vague, asking, “Are there other relationships or activities that readers could perceive to have influenced, or that give the appearance of potentially influencing, what you wrote in the submitted work?” For example, suppose a physician serves on a committee of a professional society that is focused on a particular issue and that physician attended a single meeting with support from that society to discuss the issue. Should that physician declare a potential COI if he or she writes a Viewpoint about the topic that is related to the committee's focus? Although there could be debate about whether this represents a COI, we would expect the author to declare this as a possible COI. At times, a reader may perceive that the author has a COI that should have been reported in section 4, while the author does not. We will continue to work with the ICMJE, of which JAMA is a founding member, to make this form clearer and easier for authors to complete.

We will continue to be responsive when issues arise regarding the reporting of potential COIs. We will also continue to require that all disclosures of COIs are reported for each author in manuscripts that are accepted for publication in JAMA and include these disclosures in the published article.³ We believe it is critical that readers are aware of these possible COIs when reading the article and that the disclosure information should be available without having to search for the COI disclosures elsewhere, such as on multiple pages of forms posted on the journal website.

With the availability of modern search engines, it is possible for readers to check on the consistency between disclosures of COIs reported in a JAMA article and disclosures that authors have reported elsewhere or relationships that they may have with various entities. Over the past year, we have had to publish corrections and clarifications for a number of “undeclared” COIs for authors of published articles. Therefore, we encourage all authors to declare any and all possible COIs, and if there is any question, to err on the side of full disclosure.

JAMA supports the ICMJE policy of clinical trial registration prior to the enrollment of any participants.^1,4 In addition, JAMA requires, as a condition of consideration for publication, that all trials be registered in a public trials registry that is acceptable to the ICMJE.¹ This is to ensure that there is a complete public record of all trials, meeting ethical obligations to individuals who participate in research and helping ensure that all data are available to assess the evidence regarding a specific health issue or question.

We are adopting a somewhat more flexible policy with respect to late registration (ie, registration that occurs after participants have been enrolled), for a number of reasons. First, investigators continue to have some uncertainty about whether some types of interventions require trial registration, such as quality improvement investigations or educational interventions, trials in which the unit of randomization is a cluster rather an individual, and research that involves randomization of the same individual to various interventions, such as crossover trials. By our interpretation, the current ICMJE definition does require trial registration for these types of studies, but we recognize that confusion may occur.

Second, administrative errors do occur and mistakes may happen, leading to registration occurring after some patients have been enrolled in the study. Although the investigators are ultimately responsible for ensuring that trial registration is appropriate, complete, timely, and compliant with ICMJE guidelines,⁵ there have been several examples in which administrative errors have occurred at the level of the institution, the sponsor, and, rarely, the trial registry, resulting in delayed trial registration.

In the case of late registration, authors will have to provide detailed information regarding several specific issues that include, but are not limited to, detailed reasons for late registration as well as assurances based on substantial and convincing documentation that no change has been made in any of the primary outcomes after the first participant was enrolled. Clinical trials that were registered late will be evaluated on a case-by-case basis. Only those trials for which authors provide a sufficiently compelling explanation and justification for a delay in registration may be eligible for consideration by JAMA. In the event such a trial is published, details about the trial registration will be included in the article, along with an explanatory editorial note.

This modification of our current policy has been discussed at length with the Journal Oversight Committee, Editorial Board, and JAMA editors. We believe this approach is consistent with the initial intention of clinical trial registration in ensuring that information about all clinical trials is available in the public domain so that investigators, physicians, other health care professionals, and the public will be aware of the existence of all trials and, ultimately, of all trial results.

In the interest of timeliness, research reports submitted to JAMA should be based on the most current data possible. Manuscripts based on data from clinical trials should be reported as soon as possible after the trial has ended, ideally within 1 year after follow-up has been completed. For cohort studies, the date of final follow-up should be no more than 5 years before manuscript submission, and data used in case-control or cross-sectional studies should have been collected as recently as possible, but no more than 5 years before manuscript submission. Because manuscripts in which the most recent data have been collected more than 5 years ago ordinarily will receive lower priority for publication, authors of such manuscripts should provide a detailed explanation of the relevance of the information in light of current knowledge and medical practice.⁶ Likewise, meta-analyses and systematic reviews should be based on the most recent literature available, with searches including articles published within 6 months of manuscript submission.

In addition, we will continue to require that authors have complete access to the data on which the study is based and that data analysis is conducted under the direction of the author who takes responsibility for the data. For all reports containing original data (regardless of funding source), at least 1 named author, such as the principal investigator, and no more than 2 authors, must indicate that she or he “had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.”^6,7

In recent years, the ICMJE and major journals have paid increasing attention to the issue of industry influence in biomedical research. For instance, according to current ICMJE Uniform Requirements,¹ “Editors may request that authors of a study funded by an agency with a proprietary or financial interest in the outcome sign a statement, such as ‘I had full access to all of the data in this study and I take complete responsibility for the integrity of the data and the accuracy of the data analysis.’ Editors should be encouraged to review copies of the protocol and/or contracts associated with project-specific studies before accepting such studies for publication. Editors may request a statistical analysis of all data by an independent biostatistician. Editors may choose not to consider an article if a sponsor has asserted control over the authors' right to publish.”¹

Given the requirements for attesting to data integrity, there are many institutions and their investigators that engage in “best practices.” For example, the Thrombolysis in Myocardial Infarction (TIMI) group and the Duke Clinical Research Institute, regardless of the source of study funding, ensure that the investigators have full access to all study data and conduct their own analyses independent of any funding agency or entity. We would encourage all institutions and investigators to adopt these practices. Few issues are as important in medical journalism as ensuring data integrity and the veracity of scientific research.

We reaffirm our policy of independent statistical analysis for industry-analyzed studies. For all reports containing original data, regardless of funding source, analysis of the data must be conducted by a statistician at an academic institution, rather than by statisticians employed by the sponsor or by a commercial contract research organization. The biostatistician should have full access to the entire raw data set and must be a faculty member at a medical school or academic center (such as a university) or an employee of a government research institute, such that the academic organization has oversight over the person conducting the analysis.⁸

We recognize that the vast majority of investigators in industry are well-qualified scientists and are above reproach, as are researchers and statisticians at commercial research organizations who often analyze industry-supported trials. Yet some evidence suggests that, perhaps because of the financial or proprietary interests, as noted by ICMJE, some trial reports and other studies that have been authored or analyzed by investigators who have funding from, affiliations with, or are employed by commercial entities may have introduced bias in the way data are analyzed, presented, or interpreted.^9-13

The policy of requiring independent statistical analysis will be reconsidered annually because mandatory trial registration, changes in the way industry conducts trials, and evolving ethical standards governing research may influence this policy. Future evaluation may include audits and perhaps research projects that will allow us to further analyze our approach to these studies. We will continue to seek input and advice from our Editorial Board and from research leaders in industry, government, and academia. We welcome discussion on this policy.

All information in manuscripts under consideration by but not yet published in JAMA, including related materials and communications and the fact that the paper was submitted for consideration, is strictly confidential. For accepted manuscripts this includes the date of publication. This policy of confidentiality applies to authors, members of the editorial staff, peer reviewers, and funders. Information contained in or about accepted manuscripts cannot appear in print, audio, video, or digital form or be released by the news media until the specified publication date and time of release of the journal embargo.^6,14 Unauthorized prepublication or release of information contained in accepted manuscripts and those under consideration by JAMA may result in rejection of the paper.

However, a standard abstract presentation of data contained in a manuscript at a bona fide scientific meeting is acceptable, as is publication of abstracts in print and online conference proceedings and posting of slides or videos from the scientific presentation on the meeting website. Authors considering presenting or planning to present the work at an upcoming scientific meeting should indicate the name and date of the meeting in JAMA 's manuscript submission form. For accepted papers, the editors may be able to coordinate publication with the meeting presentation or release by the sponsor by prior agreement.

Authors who present information contained in a manuscript that is under consideration by JAMA during scientific or clinical meetings should not distribute complete reports (ie, copies or PDFs of manuscripts) or full data presented as tables and figures to conference attendees or journalists. This policy extends to all types of scientific communication, such as webinars, which are becoming common at national scientific meetings.

For data from reports for which authors are considering submitting a manuscript to JAMA but have not yet done so, media coverage of presentations at scientific sessions will not jeopardize consideration. However, publication of full reports in proceedings or online, issuing detailed news releases reporting the results of the study, or participation in formal news conferences without prior discussion with the JAMA editorial staff may jeopardize chances for publication. Authors should contact us with any questions about these issues.

A critically important editorial responsibility is to review major editorial policies regularly and to update them when substantial evidence suggests a modification is necessary. You have our pledge that we will fulfill this responsibility. In surveys and discussions with authors and readers, the most consistent and resounding comment involves the high integrity of JAMA, and the trust that if a study is published in JAMA, the results are true. Integrity is critically important in medicine and public health. We will not waiver from our commitment to ensure that what we publish is valid and that the interpretation of the data by all authors is consistent with the evidence and does not go beyond it. These editorial policies are designed and intended to uphold that commitment and to ensure the scientific validity and integrity of all articles published in JAMA.

Editorials represent the opinions of the authors and JAMA and not those of the American Medical Association.