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

Displaying 291 - 295 of 320 comments
  • Frank Shann
    University of Melbourne
    Role(s):
    • Researcher
    • Clinician
    Date Submitted: Monday, January 25, 2016 - 19:36

    Requirement To Share Data
    • I agree with this general approach
    Comments: I am strongly in favour of mandatory data sharing - this measure is long overdue.

    6 Month Time Frame
    • I agree with this general approach

    Require a Data Sharing Plan
    • I agree with this general approach
    Comments: There will need to be a secure way to store data that is free for investigators from small institutions, especially those in low-income countries. Perhaps on clinicaltrials.gov?

    Providing Credit
    • I agree with this general approach

    Other Comments: I see an occasional problem in deidentifying data, for example when there are only a few patients in a subgroup and age and sex are required for the analysis. Even then, investigators using the data would rarely be able to identify individuals without access to the original records.

    Position:
    Comments:
  • Anil Makam
    UT Southwestern Medical Center
    Role(s):
    • Researcher
    • Clinician
    Date Submitted: Monday, January 25, 2016 - 11:22

    Requirement To Share Data
    • I agree with this general approach
    Comments: I would also argue researchers should share individual patient data for trials that are not published as well, but this is likely outside the scope of ICMJE

    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:
  • K.M. Venkat Narayan
    Emory University
    Role(s): Researcher
    Date Submitted: Monday, January 25, 2016 - 10:24

    Requirement To Share Data
    • I agree with this general approach
    Comments: This is important development and I think a good idea. So much effort and form filling is put into declaring conflict of interest, and its value in protecting scientific integrity is highly doubtful, in my opinion. On the other, if authors could make publicly available a final version of the dataset that they used for a publication, it would allow anyone to reanalyze and reproduce the results – this is what can improve science integrity, not ineffectual policies of declaring “conflicts of interest”. This whole drama of "conflicts of interest" is not only ineffectual but is also very subjective and political. This policy of data sharing should apply to all papers, not just clinical trials.

    6 Month Time Frame

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

    Providing Credit
    • I agree with this general approach

    Position:
    Comments:
  • Sean Coffey
    University of Sydney
    Role(s):
    • Researcher
    • Clinician
    • Clinical trialist
    Date Submitted: Saturday, January 23, 2016 - 22:33

    Requirement To Share Data
    Comments: I disagree with this statement. Anyone who has worked on a study of any size knows there are idiosyncrasies to data collection, and there is often a vast amount of knowledge specific to the trial required to analyse it effectively. Putting raw data "out there" simply makes it more likely that non-sensical re-analyses will be performed, with constant vigilance and time required by the original investigators in explaining why published re-analyses are or are not accurate. In addition, a well conducted clinical trial should generate multiple sub-analyses over the following years. For example, the Simvastatin and Ezetimbie in Aortic Stenosis (SEAS) trial, published initially in 2008, has continued to provide very useful answers to questions in aortic stenosis across multiple publications since then. To do this generally involves putting additional effort in during design and data collection, with the aim of completing the analyses at a later date. There will be little onus on a researcher to do this if the data is rapidly made public, and there is a likelihood that separate research teams will be working on the data.

    6 Month Time Frame
    Comments: The data cleaning involved in a trial of any size is onerous, and fully de-identified data is unlikely to able to be generated without significant effort, except for the most common diseases. A 6 month timeframe effectively excludes all but the best-funded industry trials. Academic trials are highly unlikely to be able to comply with such a stringent requirement.

    Require a Data Sharing Plan
    • I agree with this general approach
    Comments: This is very reasonable, and should allow for statements along the lines of "Investigators wishing to collaborate on subsequent sub-analyses are invited to contact us".

    Providing Credit
    • I agree with this general approach
    Comments: Those that generate the datasets should be offered co-authorship. In fact, going by ICMJE's recommendations for co-authorship, without collaboration there will not be anyone on the list of authors that fulfils the following: "Agreement to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved." Having both generated data-sets and analysed other people's datasets, the work involved, the expense, and the risk in the former far surpasses that of the latter. A citation does not provide anywhere near the level of compensation required.

    Other Comments: The effect of these recommendations overall are likely to make academic clinical trials a thing of the past. Why would I go to the time, effort and expense of designing and running a clinical trial, when I can just re-analyse someone else's data? Unfortunately, this means that pharmaceutical companies will frame the questions even more than they do now. Is this really what the ICMJE wish?

    Position:
    Comments:
  • richard harle
    retired public health nurse
    Role(s): Other
      • retired nurse
    Date Submitted: Saturday, January 23, 2016 - 10:46

    Requirement To Share Data
    • I agree with this general approach
    Comments: at long last we would have improved transparency and shorter response time for follow up on good research and exposure of poor research

    6 Month Time Frame
    • I agree with this general approach

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

    Providing Credit

    Position:
    Comments:

Pages