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

Displaying 31 - 35 of 320 comments
  • European Association of Hospital Pharmacists (EAHP)
    European Association of Hospital Pharmacists
    Role(s): Other
    Date Submitted: Monday, April 18, 2016 - 08:11

    Requirement To Share Data
    • I agree with this general approach
    Comments: As noted in the ICMJE proposal article, there can certainly be said to be a consensus now on the need to share clinical trial results in an open, transparent and accessible manner. This is evidenced both by the growth of coalitions such as the Alltrials campaign, and by movement in public legislation, such as the EU’s 2014 Clinical Trials Regulation.

    6 Month Time Frame
    • I agree with this general approach
    Comments: It would be useful if, after a period of time following implementation of the new approach (3 years?), the timescale be reviewed to see if it is feasible and appropriate to reduce the period of time, to potentially 3 months.

    Require a Data Sharing Plan
    • I agree with this general approach
    Comments: This approach is supported. The sharing of information is an important part of the value of any trial. Ultimately it would be for the great benefit of society if all data on trials be shared in a single searchable portal and we trust ICMJE may support moves to help this occur.

    Providing Credit
    • I agree with this general approach
    Comments: As stated in the proposal document, “data sharing is a responsibility”. As such, authors should consider it a duty to share clinical trial data. A qualifier as to what is meant by “substantial credit” and in what cases it applies might be helpful therefore (i.e. what was above and beyond the call of duty).

    Other Comments: EAHP suggest ICMJE join the Alltrials campaign. www.alltrials.net

    Position:
    Comments:
  • Zora Djuric
    University of Michigan
    Role(s):
    • Researcher
    • Clinical trialist
    • Clinical trial participant
    Date Submitted: Monday, April 18, 2016 - 07:57

    Requirement To Share Data
    • I agree with this general approach

    6 Month Time Frame

    Require a Data Sharing Plan
    • I agree with this general approach
    Comments: This is acceptable if aggregate data only I shown, not at the individual level.

    Providing Credit

    Other Comments: 1) Providing individual level data would violate my IRB approvals for conducting clinical studies. 2) It may be possible for patients to identify themselves from a few demographic descriptors and the fact that they participated in a specific trial at the institution. These are serious issues!

    Position:
    Comments:
  • Andrew Udy
    Australian and New Zealand Intensive Care Research Centre, Monash University
    Role(s):
    • Researcher
    • Clinician
    • Clinical trialist
    Date Submitted: Monday, April 18, 2016 - 04:33

    Requirement To Share Data
    Comments: Along with many in the scientific and lay communities, the Australian and New Zealand Intensive Care Research Centre (ANZIC-RC) recognizes the potentially significant benefits of sharing clinical trial data. However, if unregulated, this also poses significant risks. Specifically; multiple, unplanned, and biased post-hoc analyses of clinical trial data may significantly undermine the primary, planned analysis and published results, thereby reducing the community’s confidence in this process, and potentially resulting in widespread, ill-informed practice change. To mitigate against such risks, the ANZIC-RC would support a ‘controlled access’ model, where by sharing of clinical trial data only occurs in accordance with an agreed data sharing plan, that respects the ethical, regulatory, and legal obligations of the primary investigators, and the privacy and informed consent of clinical trial participants. We would argue that data be should shared for the purposes of replicating the primary analysis only, or to conduct pre-specified, hypothesis driven, secondary analyses. Of concern to non-industry academic centres such as ours, is that the proposed model of data sharing will require specific infrastructure and resources. Funding agencies will therefore have to allocate significant capital to establishing a data sharing governance structure.

    6 Month Time Frame
    Comments: The ANZIC-RC would not be supportive of an inflexible six-month timeframe for sharing individual patient data from clinical trials. Investigators often have a number of pre-specified post-hoc exploratory analyses planned, which take considerably longer than six-months. Researchers should be allowed to complete and publish these analyses, prior to any data sharing commitments. While a six-month timeframe may be appropriate where the aim is to replicate the primary analysis only, we consider that in other circumstances, this time period should vary. The ANZIC-RC would support a system that allows for such timeframes to be specified in a data sharing plan, and that this could be modified, depending on the data being shared. We suggest that reasonable timeframes would be a maximum of six months for requests to replicate the primary analysis, and a maximum of two years for other requests.

    Require a Data Sharing Plan
    Comments: The ANZIC-RC deems a data sharing plan as essential, in order to protect the interests of all parties actively involved in clinical trials, not least of which includes participants, researchers, trial sponsors, industry, and the community. If deemed a necessary component of trial registration, the ANZIC-RC considers a ‘controlled access’ model as the only robust governance process that would be appropriate. Inherently, this must comply with relevant ethical, legal, and regularly requirements, and should detail at least the following: a) the submission process, b) how this is adjudicated, and by whom, c) the stipulation and limitations of any data access, d) how these any additional results will be disseminated, e) any fees or costs involved, and f) an independent appeal process in the event of a dispute between the primary researchers and those seeking access to trial data.

    Providing Credit
    Comments: The ANZIC-RC believes that at least one author from the original study should be offered co-authorship on any secondary analyses using shared data, in order to recognise the contribution made by the primary investigators. In the event that the original study author(s) do not wish to be acknowledged as co-authors, any resultant manuscript should clearly state this. Furthermore, where the primary author(s) do not take up the opportunity to be co-authors, they should be afforded the opportunity to provide contemporaneous commentary to any resulting publication, thereby outlining their views regarding the strengths and weaknesses of the secondary analysis.

    Position:
    Comments:
  • jack Cuzick
    Wolfson Institute of Preventive Medicine, Queen Mary Univ London
    Role(s):
    • Researcher
    • Clinical trialist
    Date Submitted: Monday, April 18, 2016 - 03:13

    Requirement To Share Data
    Comments: Principle is OK, but there several issues. One is ongoing trials with long term follow up where seeking consent is likely to be almost impossible. The require should only be required for new trials and changes to this proposal are needed for ongoing trials. Concern about mischievous use of data by those with vested interests. We are open to requests for data but reserve the right to say no if there are issues. Data should not be freely available to anyone, but an independent review body could be created under the control of either the ethics committee or if academic, the university or hospital review committee. We prefer in many cases to provide tables for the requested analyses rather than raw data to avoid misuse and loss of control of trial data.

    6 Month Time Frame
    Comments: 6 months ok but full release of IPD not always desirable or ethical.

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

    Providing Credit
    • I agree with this general approach
    Comments: This is key and needs to be developed more. Too many overviews do not include initial trial leaders as authors and they are the ones who need full credit for their work. An offer of authorship should be the norm.

    Other Comments: In general provision of tables and specific analyses is desirable rather than uncontrolled provision of IPD. Trialists should be able to interact with external requesters to ensure there is no misunderstandings about what the data measure and how they were collected. Some sort of independent body should adjudicate disputes about what data is provided and how it is used.

    Position:
    Comments:
  • Catrin Tudur Smith
    MRC NWHTMR, University of Liverpool
    Role(s):
    • Researcher
    • Clinical trialist
    Date Submitted: Monday, April 18, 2016 - 01:47

    Requirement To Share Data
    • I agree with this general approach
    Comments: We fully support the intention of introducing a requirement for authors to share de-identified IPD. However, it is important that all IPD from the trial be made visible and available rather than simply the IPD underlying the results presented within the published article. Unfortunately, as it stands, this policy will not help the research community to identify other valuable data that were collected during the trial, will only partially increase transparency and will not tackle the important problem of selective reporting that is well known to be prevalent in the medical literature (1,2). (1) Dwan K, Altman DG, Clarke M, Gamble C, Higgins JP, Sterne JA, Williamson PR, Kirkham JJ. Evidence for the selective reporting of analyses and discrepancies in clinical trials: a systematic review of cohort studies of clinical trials. PLoS Med. 2014 Jun 24;11(6):e1001666. (2) Dwan K, Gamble C, Williamson PR, Kirkham JJ; Reporting Bias Group. Systematic review of the empirical evidence of study publication bias and outcome reporting bias - an updated review. PLoS One. 2013 Jul 5;8(7):e66844.

    6 Month Time Frame
    • I agree with this general approach
    Comments: We commend the ICMJE for defining a timeframe within which IPD should be shared. However, this rather short timeframe could be problematic for the academic, publicly funded trials community, as secondary analyses of the trial data will often be undertaken shortly after publication of primary analyses. We would suggest that a maximum of 12 months be allowed with authors encouraged to work towards a shorter time frame wherever possible.

    Require a Data Sharing Plan
    • I agree with this general approach
    Comments: We support the requirement to make data sharing plans clear and available during trial registration. Also important is to define data sharing plans within the trial protocol (3). (3) Tudur Smith C, Hopkins C, Sydes MR, Woolfall K, Clarke M, Murray G, Williamson P. How should individual participant data (IPD) from publicly funded clinical trials be shared? BMC Med. 2015 Dec 17;13:298

    Providing Credit
    • I agree with this general approach
    Comments: This could be achieved by assigning an appropriate citation identifier to the dataset, publishing details about the dataset in an appropriate journal (4) and imposing a requirement to cite the data appropriately in any subsequent secondary analyses. Rather than being cautious about ‘giving away their data’, the academic community should view data sharing as a positive and direct pathway to creating ‘impact outside of academia’, a measure of research attainment that is being increasingly focussed upon by higher education and research funders. The fourth requirement is logical but could be impractical particularly when synthesising multiple independent clinical trial datasets, each of which has an associated published primary analysis. The aim of secondary analyses will often be completely different to the original aims of a clinical trial (eg researchers may wish to use IPD from multiple clinical trial datasets to develop a prognostic model) and it would be unrealistic to expect a researcher to have to list how the analyses differed to each of those previous analyses. (4) Hrynaszkiewicz, I., Khodiyar, V., Hufton, A. & Sansone, S.-A. 2016. Publishing descriptions of non-public clinical datasets: guidance for researchers, repositories, editors and funding organisations. bioRxiv. doi: http://dx.doi.org/10.1101/021667

    Position:
    Comments:

Pages