When it comes to accessing clinical trial data, for me “transparency” means above all clarity and openness, a willingness and indeed a desire to share data from clinical trials.
I speak as chairperson of the Steering Committee of ClinicalStudyDataRequest.com (CSDR) , which is a group of 13 major pharmaceutical companies dedicated to facilitating greater access to clinical trial data to promote scientific research. As such, I applaud ICMJE’s efforts in support of data sharing.
CSDR has been a leader in the broader movement by pharmaceutical companies and other institutions that fund trials who are together transforming the data disclosure landscape. CSDR offers access to detailed information from clinical trials in ways unheard of just 5 years ago.
By being transparent and inviting researchers to access clinical trial data, CSDR hopes to foster research by giving scientists access to a rich set of data. The fruits from this research, once published, are intended to advance patient care, contributing to medical knowledge more generally.
CSDR currently lists more than 3,400 clinical trials and this list is expanding quickly—often as a result of enquiries from researchers. Researchers are encouraged to submit requests for the detailed, anonymized, patient-level data from these trials. They can request access to trials across multiple sponsors, and access is free. Each request is reviewed independently by a process overseen by the Wellcome Trust. On approval and receipt of a signed data-sharing agreement, the data are made available in a secure environment—one that facilitates the research while at the same time preserving the privacy of patients who were part of the trials.
The available clinical trials span a wide range of disease areas. Today there are more than 360 diabetes and 240 asthma trials, for example. So far there have been 335 requests from researchers for trial data. A publication plan is required so that results can be shared broadly. Some results may be published in journals associated with the ICMJE or in other journals that may encourage sharing of patient-level data. Those data may in turn be shared on CSDR—a cycle of knowledge that, in the end, will help inform the scientific community and ultimately benefit patients.
Executive Deputy Editor, Annals of Internal Medicine
Secretary, International Committee of Medical Journal Editors
The International Committee of Medical Journal Editors (ICMJE) recently announced a requirement for data sharing statements as a condition for consideration at our member journals of manuscripts describing the results of clinical trials, effective July 1, 2018. In addition, clinical trials that begin enrolling participants on or after January 1, 2019, must include a data sharing plan in the trial’s registration.
By adopting these criteria, ICMJE aims to help fulfill our ethical obligation to trial participants who put themselves at risk on the promise that we will make the most of the information gathered in order to improve clinical care. Although sufficient infrastructure is not yet in place for the ICMJE to mandate data sharing at this time, requiring data sharing statements reflects our commitment to help move the scientific community toward creating the environment in which sharing is expected. ICMJE is excited to engage multiple stakeholders such as Medical Publishing Insights & Practices (MPIP) and the International Society for Medical Publication Professionals (ISMPP) in this endeavor.
MPIP has inquired about the content of data sharing statements that would meet the ICMJE requirements. Data sharing statements must indicate the following items: whether individual deidentified participant data (including data dictionaries) will be shared; what data in particular will be shared; whether additional, related documents will be available (eg, study protocol, statistical analysis plan); when the data will become available and for how long; by what access criteria data will be shared (including with whom, for what types of analyses, and by what mechanism). Illustrative examples of data sharing statements (including example mechanisms by which data will be made available) that would meet these requirements are in the Table of the editorial announcing these requirements at http://www.icmje.org/news-and-editorials/data_sharing_june_2017.pdf .
It is important to note that the examples provided represent some, but not all, possible statements that would fulfill the ICMJE requirement. These may range from unrestricted, publicly available access to deidentified data without restrictions, to a statement indicating that data will not be shared. The important thing is to be transparent about what will be done regarding each of the items noted above.
MPIP also asked how journal editors will evaluate data sharing statements, and how the information provided will be factored into editorial decisions. All ICMJE editors will check to be sure that the data sharing statements include the items outlined above and in our editorial. Although all data sharing statements must address these items, editors may differ in the data sharing plans they find acceptable. Some ICMJE editors already require that data will be shared, as indicated in their information for authors. Others do not currently have such a rule (although they may adopt it in the future) and the degree to which a more or less restrictive data sharing statement will influence editorial priorities may vary. As a general rule, all of the editors would have a preference for plans indicating that data will be responsibly shared. Peer reviewers will see the data sharing statements in the submitted manuscript and, like editors, may vary in the weight they place on more or less rigid data sharing plans when making their recommendations to editors.
MPIP also inquired when it might be acceptable to plan not to share deidentified individual patient data. Such situations might include when it is impossible to adequately assure the confidentiality of trial participants or where national regulations prohibit sharing. As stated above, however, the ICMJE does not currently require that data be shared (although some individual journals may). ICMJE does require a transparent statement of what will be done.