Earlier this year an elite group of medical journals proposed it would require data sharing by authors within 6 months of publishing clinical trial reports. At Springer Nature we, in general, support the International Committee of Medical Journal Editors’ (ICMJE) ambitious proposals but feel more detail, guidance and a transitional period will be needed if implementation of the policy is to be effective.

Sharing and connecting clinical research data and publications is central to a number of projects and journals (such as Trials) at Springer Nature, including the Linked Reports of Clinical trials initiative and efforts at Scientific Data to better connect sensitive/clinical data with the peer-review process and open access articles.

Sharing and connecting clinical research data and publications is central to a number of projects and journals at Springer Nature.

Our full response to the proposals is here. A summary of the key points:

• We recommended that the effort often involved in anonymising individual patient data (IPD) should be better recognised, and that patient consent for data sharing be more explicitly addressed. Also, to what data the policy applies to could be better defined. Defining what data are relevant to support published claims is notoriously difficult.
• On the 6-month deadline for data sharing proposed, while we support a deadline, it may conflict with other mandates – such as the NIH policy for sharing genomic data from clinical trials that might require more rapid data sharing.
• In clinical research there is insufficient use and awareness of suitable repositories for clinical data (although they are now being catalogued by Scientific Data), which would enable compliance with the policy. This contrasts with clinical trial registration policy, where there are numerous established registries to enable compliance.
• We support the requirement to share data management plans (DMPs) but where and how DMPs are shared needs flexibility – journals, by utilising extended methods sections, supplementary material or innovative articles types are alternative venues to trial registries.
• As a means to provide credit to those who share data, the clinical trial community should embrace the growing culture and practice of data citation. By making data available in repositories (public and restricted access) that can assign data digital object identifiers (DOIs) to deposited datasets, data can be formally cited in reference list/bibliographies. Leveraging established mechanisms of citation facilitates data linking to publications and tracking of reuse, and provides a means to reward clinical data sharing through citation – the established currency of scholarship.

Overall, this ICMJE effort to increase transparency on data sharing in clinical research should be applauded.

Overall, this ICMJE effort to increase transparency on data sharing in clinical research should be applauded.

The journal research data policy landscape is currently too complex and sometimes opaque, and efforts to harmonise standards across hundreds – if not thousands – of journals should improve the experience for authors, readers and reviewers, and increase the visibility and reliability of research.