Comments on the ICMJE’s Proposals for Sharing Clinical Trial Data

Displaying 21 - 25 of 320 comments
  • Debra Barton
    University of Michigan
    Role(s): Researcher
    Date Submitted: Monday, April 18, 2016 - 10:42

    Requirement To Share Data
    Comments: I could not disagree with this condition more. There are already required data sharing policies and plans based upon levels and sources of funding. So this already exists and I do not know anyone who does not comply with their data sharing policies. Making this a mandatory requirement of article submission with the idea of submitting de-identified raw data for others to analyze any way they see fit does not advance science. Data are collected within a context and the study design based on a specific research question. To allow that data to be used in an unsupervised, scientifically un-reviewed way is unethical and will lead to poor science.

    6 Month Time Frame
    Comments: In addition to my comments above, some of us have post docs, pre docs and junior faculty who use our data for secondary analyses to advance their careers. This policy would put this at risk - one would not know what is happening with the data. In addition, making data available to others who may not handle it appropriately and then might publish something contrary to the initial results would then require a process to resolve the discrepancy and expose mishandling. Resolving this would take time and resources that would need to be paid for. Who would take the responsibility for this; who would be the arbiter? Who would pay?

    Require a Data Sharing Plan
    • I agree with this general approach
    Comments: Including the data sharing plans so people know how to appropriately request data for legitimate scientific reasons makes sense and would be helpful.

    Providing Credit
    Comments: There is absolutely no way to enforce this. Therefore, it is not operational.

    Other Comments: If such a policy were put in force to require de-identified datasets be submitted, I simply would never publish in any of the journals affiliated with this.

    Position:
    Comments:
  • Jacques Demotes
    ECRIN-ERIC
    Role(s): Other
      • Research Infrastructure Director
    Date Submitted: Monday, April 18, 2016 - 10:23

    Requirement To Share Data
    • I agree with this general approach
    Comments: We strongly agree with this initiative. The question is not if but how to responsibly share data with appropriate control. Sharing clinical trial data strengthen responsible and safe clinical research. ECRIN’s Scientific Board has endorsed transparency by requesting access to de-identified individual patient data (IPD). The individual participant data (IPD) sharing will be especially valuable for teams planning re-analyses, systematic reviews and meta-analyses. IPD sharing will also benefit group analyses or network meta-analysis, a method that addresses multiple interventions. The findings generate new hypotheses that should be tested in new randomized clinical trials.

    6 Month Time Frame
    • I agree with this general approach
    Comments: We consider that six months is maybe a short timeframe and would question whether twelve months would make it easier especially for the under-resourced academic investigators and sponsors

    Require a Data Sharing Plan
    • I agree with this general approach
    Comments: In the longer term, data sharing should be mandatory also for non-interventional studies; (epidemiologic cohorts, registers of patients).

    Providing Credit
    • I agree with this general approach
    Comments: In general the current authorship rules act against large collaborative work, and tend to fragment consortia and reduce collaboration and sharing. There is an urgent need for an in depth revision of authorship rules in medical research.

    Other Comments: The ICMJE initiative should also enforce access to other trial-related information (trial protocol, data management plan, statistical analysis plan, informed consent form, regulatory approval, ethical approval, sample CRF, statistical report, cleaned database with IPD, etc). Without this information a proper and valid re-, meta- or secondary analysis of clinical trial data is not possible. We propose the development of a metadata repository for data objects from journals, databases, repositories, webpages, etc, that would provide a single point for researchers and reviewers to track clinical research data objects of interest, allowing identification of clinical studies (eg. via clinical trial registry IDs), display a selection of data objects related to a specific study, metadata of a data object, including the nature of the object (e.g. name, type, provenance) and the mode of access (public, restricted, private). A link to the data object can be provided if possible, including the link to IPD. In the H2020 CORBEL cluster project, ECRIN is developing policies and procedures for safe sharing of IPD and is also piloting ideas for a metadata repository (www.corbel-project.eu). Pilot work with respect to such a meta-repository has also been done by OpenTrials.

    Position:
    Comments:
  • Somenath Roy
    Vidyasagar University
    Role(s): Researcher
    Date Submitted: Monday, April 18, 2016 - 10:22

    Requirement To Share Data
    • I agree with this general approach

    6 Month Time Frame
    • I agree with this general approach

    Require a Data Sharing Plan
    • I agree with this general approach

    Providing Credit
    • I agree with this general approach

    Position:
    Comments:
  • Gérard Krause
    German Society for Epidemiology
    Role(s): Researcher
    Date Submitted: Monday, April 18, 2016 - 10:08

    Requirement To Share Data

    6 Month Time Frame

    Require a Data Sharing Plan

    Providing Credit

    Other Comments: Thank you for providing the opportunity to comment on your proposal 'Sharing Clinical Trial Data'. In the German guidelines for good epidemiological practice, secure longterm storage of all relevant data sets and syntax is already recommended (DOI: 10.1055/s-2004-813850). However, further recommendations on how to archive epidemiological data sets and to which extent they can be made accessible without restricting personal rights of study participants are urgently needed. Ensuring data protection will be of particular relevance to our discipline, since epidemiological surveys often cover a large amount of person-related information, making it easier to trace back individuals behind the data. We would like to point out to some initiatives aiming to provide easy and cost-efficient access to reasearch data. DataCite (www.datacite.org) is an international collaboration that implemented a registration service for DOIs for scientific data sets, a major prerequisite for a consistent and reliable data storage, enabling researchers to identify and cite data sets. (see also www.da-ra.de). Quantification of DOI-referencing for datasets has been suggested as a starting point for alternative ways of providing academic reward to those collecting and managing primary data.

    Position:
    Comments:
  • Professor Sir Rory Collins
    Nuffield Department of Population Health, University of Oxford
    Role(s):
    • Researcher
    • Clinical trialist
    Date Submitted: Monday, April 18, 2016 - 10:07

    Requirement To Share Data
    Comments: We do not support this proposal as it might result in less rather than more transparency. An alternative solution would be to make additional tables of all of the recorded adverse events available as a supplement to the trial publication (for example as were provided in the HPS2-THRIVE trial http://www.nejm.org/doi/full/10.1056/NEJMoa1300955 and the HOPE3 trial http://www.nejm.org/doi/full/10.1056/NEJMoa1600176). This approach addresses concerns about unpublished data in a highly cost-effective way, and gives credit to the investigators.The current proposal has the following limitations: (1) By only requiring the IPD to be made available for results that have been published, it does not address the issue of data that aren’t published (2) Making IPD data available in a format that is accessible to other researchers is costly/time-consuming, and such data may be of value from only a subset of trials (and thus is not very cost-effective) (3) The proposal may result in trialists holding back data/results so that they don’t have to make it public until they have had a chance to explore it all (so inadvertently delaying access) and may disincentivise researchers from generating data in the first place.

    6 Month Time Frame
    Comments: In the alternative proposal we suggest above, it would be reasonable to make the tables of all AEs that were recorded available at the same time as, or soon after a publication (and certainly within 6 months).

    Require a Data Sharing Plan
    • I agree with this general approach
    Comments: We agree with this proposal.

    Providing Credit
    Comments: See alternative proposal suggested above: if the approach was to make detailed web-tables available that supported the analyses and, indeed, extended the data made available to all of the recorded AEs (and any other tabulations that might help others) then the credit would go to the people who generated the data by referencing these tables as the source.

    Other Comments: Making datasets usable by other researchers is not a trivial task. To be usable, data needs to be carefully curated with associated metadata and detailed manuals on how to use the data. Even with these efforts, our considerable experience analysing others’ IPD for collaborative meta-analyses indicates that support from the researchers who conducted the trial is required more often than not. We therefore encourage the Editors to seek advice from people who have experience both of generating the data and of using other people’s data before proposing a course of action that is evidence-free.

    Position:
    Comments:

Pages