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

Displaying 6 - 10 of 320 comments
  • Chris Winchester
    Oxford PharmaGenesis
    Role(s):
    • Researcher
    • Patient
    • Other
      • Medical communications professional
    Date Submitted: Monday, April 18, 2016 - 15:16

    Requirement To Share Data
    Comments: The ICMJE proposal has the potential to strengthen trust in the clinical research endeavour, or to damage it. The ICMJE has a duty to provide clear guidance on best practice for data sharing to investigators and authors around the world. We believe that such guidance should prioritize the unique needs of the patients who take part in clinical research, on whose participation and trust medical progress depends. The Declaration of Helsinki states that medical research participants ‘must be adequately informed of the aims, methods, sources of funding [and]… conflicts of interest [and] institutional affiliations of the researcher,’ as well as the anticipated benefits and potential risks of the study. Available information on patient views on data sharing come from a single study, which suggested that research participants and their families find sharing of individual-patient data acceptable for scientific purposes and to promote health, but are concerned about the media accessing and misinterpreting or misrepresenting their data (Merson et al., 2015 JEmpirResHumResEthics;10:251–63.). Shared individual-patient data are, however, being accessed by law firms, lay media and the general public (Bonini et al., 2014 NEnglJMed;371:2452–5.). A controlled access approach is needed. See http://www.pharmagenesis.com/wp-content/uploads/2016/04/Disclosure-of-individual-patient-data-time-to-put-the-patient-first.pdf

    6 Month Time Frame
    Comments: The primary ethical responsibility of investigators once a trial has been completed is to communicate its results quickly and accurately. The responsibility to make available individual-patient data is secondary. Sufficient time should be allowed for results to be published in several articles before investigators’ attention is focused on data sharing, especially given the complexities of ensuring that both are done ethically and in accordance with good practice. The ICMJE should set a timescale for responding to a request for individual-patient data rather than a timescale for data sharing per se. From an ethical perspective, not all requests for individual-patient data should be met. What matters is that requests are considered in a transparent and timely manner, according to a pre-defined plan for data sharing.

    Require a Data Sharing Plan
    • I agree with this general approach
    Comments: A data sharing plan should set out the criteria, process and timelines for considering requests for individual-patient data, and should make it clear that some requests may not be approved on ethical grounds. If oversight by an independent review panel is required by the ICMJE, then this should be specified.

    Providing Credit

    Other Comments: 1. We believe that uncontrolled public access to individual-patient data is unethical, so the ICMJE should make it clear that data sharing needs to be restricted to research purposes only. 2. We believe that multiple analyses of individual-patient data have the potential to distort the evidence-base, so the ICMJE should insist on their registration and disclosure. 3. We believe that the patient perspective has been largely ignored, so the ICMJE should call for it to be better studied and taken into account. 4. We believe that no method of de-identification is absolute and ‘future-proof’, so the ICMJE should make it clear that sharing of individual-patient data needs to be restricted to the minimum necessary. 5. We believe that current informed consent is inadequate, so the ICMJE should clarify what is required for genuinely informed consent. 6. We believe that patients deserve to have access to the data they help to generate, so ICMJE member journals should make a patient summary of results freely available. 7. We believe that both the benefits and risks of data sharing are poorly characterized, so the ICMJE should call on researchers to measure both intended and unintended consequences and review its policy accordingly.

    Position:
    Comments:
  • Professor John Zalcberg OAM
    Australian Clinical Trials Alliance
    Role(s):
    • Researcher
    • Clinician
    • Clinical trialist
    • Other
      • Chair, Australian Clinical Trials Alliance: Peak advisory body for public-good trials in Australia.
    Date Submitted: Monday, April 18, 2016 - 15:16

    Requirement To Share Data
    Comments: The Australian Clinical Trial Alliance (ACTA) recognises the significant potential benefits to be realised through clinical trial data sharing and in principle we support the aims of the ICMJE in proposing new journal publication requirements as a lever for change. However, we believe that any process that seeks to effectively mandate data sharing must appropriately consider and mitigate the substantial associated risks to trial participants, investigators and the clinical community - namely the potential for unscientific or premature post-hoc analyses that could undermine trust in clinical trials or result in inappropriate and harmful practice change. We also believe it is crucial that sharing of trial data only occurs in accordance with an agreed data sharing plan that respects the ethical, regulatory and legal rights and responsibilities of the participants, primary investigators and sponsors. ACTA recommends that a data sharing plan adopting a ‘controlled access’ model that requires data to be shared only for the purposes of replicating the published analysis or to conduct an a priori defined, hypothesis‐driven, secondary analysis is the most appropriate way to ensure that the integrity of the data, and any conclusions drawn from them, are maintained at the highest possible ethical and scientific standards.

    6 Month Time Frame
    Comments: ACTA believes the ICMJE proposal must acknowledge the rights of investigators to conduct and complete their own secondary pre-specified and post-hoc exploratory analyses. A six month timeframe may, in some cases, be appropriate for sharing data solely for the purpose of replicating the results presented in the article, but the fair and reasonable timeframe for sharing data will differ with each trial and may be significantly longer - particularly for large, definitive phase III trials or where all trial endpoints are yet to be reported. We recommend that investigators be allowed six months to publicly declare their timelines for pre-specified subsidiary analyses and their intentions to undertake their own post-hoc exploratory analyses not specified in the published Statistical Analysis Plan. We also recommend that investigators be allowed to specify, with appropriate justification, the maximum timeframe following which data will be shared (subject to the controlled access terms outlined in the data sharing plan) with an allowance for this to vary, without risk of prejudice by journals, depending on the nature of the data being shared.

    Require a Data Sharing Plan
    • I agree with this general approach
    Comments: ACTA believes that a detailed, published, data sharing plan is essential to protect the interests of trial participants, investigators, sponsors and the community. While the vast majority of requests for data or analyses will be collaboratively approached, scientifically valid, ethically appropriate and mutually beneficial, unfettered open access would potentially be subject to abuse; hence we believe that models that incorporate clear governance controls are critical. ACTA strongly recommends the ICMJE promote a ‘controlled access’ model of data sharing that incorporates a robust governance process and complies with all relevant local ethical, legal and regulatory requirements. Such a plan would need to clearly outline: • Submission process 
 • Who adjudicates the request and how trial participants views are represented • Assessment criteria • Limitations of use including an analysis agreement 
 • How the results will be published and authorship assigned • Control of IP
 • Fees involved 
 • An independent appeal process if data/analysis request is refused. 
 We believe using existing trial registries to serve as the repository for data sharing plans is a sensible approach. However, urgent scoping work is needed to understand how readily this could be implemented and overseen by different registries, and with what resource implications.

    Providing Credit
    • I agree with this general approach
    Comments: We strongly agree that appropriate recognition of the primary investigators is a key issue and also agree that collegial, open and respectful collaboration, whilst likely to be the norm, will not always ensue with these endeavours. ACTA suggests that agreeing arrangements for co-authorship on subsequent publications produced through secondary analysis of shared data (or of analyses performed by the trialists at the request of an external party) will need to be part of the governance process clearly set out in the data sharing plan, including a right of review/appeal through an independent party. We believe that a collaborative approach should accommodate listing of an agreed number of original authors, or a group author (eg. The XXX Study Investigators) with appendix listing of all contributors. We suggest that in the event that the original study authors do not wish to be acknowledged as co-authors the manuscript should clearly state this. Importantly, we recommend that where primary author(s) do not take up the opportunity to be co‐authors on secondary publications they should be given the opportunity by the journal to provide contemporaneous commentary to any resulting publication outlining their views regarding the strengths and weaknesses of the secondary analysis.

    Other Comments: ACTA is a large national “network of networks” representing more than 10,000 clinician researchers. These collaborative clinical trials networks have collectively designed and undertaken more than 1,000 studies involving 1,000,000+ participants in the last decade. More than 100 of these were landmark trials published in ICMJE journals that directly influenced clinical practice. One of the key concerns raised by ACTA members is the substantial cost burden the ICMJE proposal will impose and its potential to significantly impact the feasibility of conducting important public good trials. While it may be reasonable to assume that the cost of facilitating individual applications for data/analyses would be covered by those requesting access, complying with the ICMJE proposal will require a substantial investment in the underlying infrastructure necessary to develop and administer data sharing capability. This will be an important consideration for funders supporting public good trials as part of the business of providing evidence-based healthcare - but it is likely (for most) to be set against a backdrop of increasingly constrained budgets for supporting high-quality clinical research. ACTA believes the next iteration of the ICMJE proposal must include a thorough evaluation of the resource implications and clear recommendations about who should bear these costs.

    Position:
    Comments:
  • Martine Clozel M.D. Chief Scientific Officer & Guy Braunstein M.D., Head Global Clinical Development
    Actelion Pharmaceuticals Ltd.
    Role(s): Clinical trial sponsor / funder
    Date Submitted: Monday, April 18, 2016 - 15:05

    Requirement To Share Data
    • I agree with this general approach
    Comments: Actelion agrees with this general approach. We currently facilitate the analyses of research proposals from study investigators on an ad hoc basis and are evaluating a more formal process for data sharing. We believe a formal process is needed to prevent post-hoc analyses that lack scientific rigor. The process should ensure that data sharing addresses research questions that benefit clinical care and advance medical knowledge. The process should also confirm the feasibility of the proposed analyses i.e. an evaluation of whether the research question can be addressed based on the data-set. Contrary to the ICMJE suggestion that authors should share clinical trial data, we believe this to be the responsibility of the study sponsor.

    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
    Comments: Actelion believes that researchers who publish articles based on data generated by others should acknowledge the study sponsor in their publications. Our experience using clinical trial data to address secodary research questions from researchers (in our case, study investigators) is one of complexity. The work typically involves extensive discussion between the researcher posing the question, the external medical experts involved in the study, the clinical trial lead and the clinical trial statistician. We acknowledge that collaboration may not always be possible. As a safeguard against misunderstandings, we would advocate that study sponsors be given a courtesy review of publications from secondary analyses.

    Other Comments: 1) Actelion commends the ICMJE on their efforts to encourage data sharing. 2) We believe that the rigor applied to the initial analyses of clinical trial data should apply to secondary analyses. Requests to conduct secondary analyses should include a robust plan including appropriate objectives to be evaluated and a statistical analyses plan. 3) We suggest that provision be made for exceptions, for example to protect the privacy of patients with rare diseases, to allow flexibility with timelines for publications relating to medication not already approved. 4) We are interested to hear of how the journal editors see their potential role in evaluating the secondary analyses of clinical trial data. 5) We encourage the ICMJE to also increase transparency around the publication of scientific and medical research, e.g. by disclosing the names, affiliations, and potential conflicts of interest of peer-reviewers, by sharing peer-review and or editor comments on accepted and rejected manuscripts, or by other means within the scope of their role as journal editors.

    Position:
    Comments:
  • Brian M. Bot
    Sage Bionetworks
    Role(s):
    • Researcher
    • Clinical trialist
    • Clinical trial participant
    • Patient
    Date Submitted: Monday, April 18, 2016 - 14:27

    Requirement To Share Data
    • I agree with this general approach
    Comments: The signatories support the sharing of all data underlying any scientific publication, to the broadest level allowed given participant consents. These data provide the basis for the claims made within the publication and should be available as part of the scientific record. In order to maximize the ability to assess the merits of claims made by the authors, data should be made available during the review of the primary article. Having these data available to the entire scientific community post-publication, as is proposed here, is also vitally important. First, release of these data will enable direct assessment of the individual study in order to determine how it advances the field. Second, release of these data will allow researchers maximize the benefit of data donation through aggregation and further analysis.

    6 Month Time Frame
    Comments: The signatories suggest that the ICMJE require that all IPD be made available immediately upon primary publication. Compliance with such requirements is quite low without some sort of formal incentive or deterrent (in this case: publication, or withholding of publication). If ICMJE has strenuous objection to data release at the time of publication, the signatories suggest that the primary data be made available to the public in an embargoed manner. That is, these data are released to journal editors by authors as a requirement at the time of primary publication and are subsequently released more broadly after the decided upon embargo period (e.g. 6 months).

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

    Providing Credit
    Comments: The signatories think it is of paramount importance that data generators receive appropriate credit for their contributions to science - both as primary researchers, as well as for broadly affecting the research ecosystem by the use of their work in a secondary manner - however, we do not believe that novel attribution credits are necessary. Traditional literature citation and, more specifically, digital object identifiers (DOIs) are mechanisms to provide credit to researchers for their contributions to the scientific community. Given the difficulty of gaining traction for new credit mechanisms, the signatories strongly suggest that the ICMJE adopt data citation practices already in use within the broader scientific community, i.e., DOI for data citation as adopted by data journals.

    Other Comments: The signatories applaud the ICMJE for taking on this extremely important topic. It is our belief that the broad availability of clinical trial data will both improve the public trust in published trials as well as maximize the long-term impact of the data donated by the research participants who have so unselfishly contributed to these trials. Signatures in support of this response are being collected here: http://www.ipetitions.com/petition/icmje-community-response

    Position:
    Comments:
  • Kathy Rowan
    Intensive Care National Audit & Research Centre (ICNARC)
    Role(s):
    • Researcher
    • Clinical trialist
    Date Submitted: Monday, April 18, 2016 - 14:16

    Requirement To Share Data
    • I agree with this general approach
    Comments: Yes, this should happen. This is an important initiative that reinforces openness and transparency in the conduct of clinical trials and, by encouraging collaborative working, maximises the benefits to patients from data collected. However, the devil is in the detail - who the "others" are may require greater specification, what data are identifiable and/or "deidentified" may require further agreement and the time and resources to prepare the data for wider consumption are not insubstantial (there needs to be clear requirements for the sharing of data that enables the correct understanding and interpretation of data by secondary analysts). To maintain the confidence of all those involved in clinical trials it is important that there are clear and robust protocols in place for sharing "deidentified" patient data.

    6 Month Time Frame
    • I agree with this general approach
    Comments: Often, dissemination according to a specified publication plan can take a lot longer than six months. Some further agreement around the definition of "following publication" as time zero is required - six months may be far too short a period (perhaps there should be flexibility as to the length of that timeframe?) In addition, this period will likely fall outside the original research grant period and makes obtaining the time and resources to prepare the data an even greater challenge.

    Require a Data Sharing Plan
    • I agree with this general approach
    Comments: Providing the issues above are addressed and agreed, "a plan for data sharing as a component of clinical trial registration" is important. Without doubt, an effective plan would and should provide assurance that "deidentified" individual patient data will be shared in a structured and timely manner and that data are shared such that that patient anonymity and data integrity are maintained. Perhaps, there should also be appropriate agreements and plans in place for those using data for secondary analysis to ensure that all potential conflicts of interest are stated and that their results and analyses are made available for peer-review and publication.

    Providing Credit
    • I agree with this general approach

    Position:
    Comments:

Pages