This is an open-access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in Journal of Participatory Medicine, is properly cited. The complete bibliographic information, a link to the original publication on http://jopm.jmir.org, as well as this copyright and license information must be included.
Sharing clinical trial data can provide value to research participants and communities by accelerating the development of new knowledge and therapies as investigators merge data sets to conduct new analyses, reproduce published findings to raise standards for original research, and learn from the work of others to generate new research questions. Nonprofit funders, including disease advocacy and patient-focused organizations, play a pivotal role in the promotion and implementation of data sharing policies. Funders are uniquely positioned to promote and support a culture of data sharing by serving as trusted liaisons between potential research participants and investigators who wish to access these participants’ networks for clinical trial recruitment. In short, nonprofit funders can drive policies and influence research culture. The purpose of this paper is to detail a set of aspirational goals and forward thinking, collaborative data sharing solutions for nonprofit funders to fold into existing funding policies. The goals of this paper convey the complexity of the opportunities and challenges facing nonprofit funders and the appropriate prioritization of data sharing within their organizations and may serve as a starting point for a data sharing
Sharing clinical trial data can provide value to research participants and communities by accelerating the development of new knowledge and therapies as investigators merge data sets to conduct new analyses (eg, meta-analyses, statistical modeling), reproduce published findings to raise standards for original research, and learn from the work of others to generate new research questions.
Nonprofit funders, including disease advocacy and patient-focused organizations, play a pivotal role in the promotion and implementation of data sharing policies. Funders are uniquely positioned to promote and support a culture of data sharing by serving as trusted liaisons between potential research participants and investigators who wish to access these participants’ networks for clinical trial recruitment. In short, nonprofit funders can both drive policies and influence research culture.
In the US Institute of Medicine (IOM) 2015 report,
Despite arguments to the contrary, in many cases, clinical trial participants are indeed willing to share their data for a wide range of uses, assuming that adequate security safeguards are in place [
Since the release of the 2015 IOM report, several efforts in the public sector have signaled that data sharing is increasingly regarded as a scientific responsibility rather than an optional activity [
More recently, for public comment, the National Institutes of Health (NIH) released an updated Draft NIH Policy for Data Management and Sharing, which would require NIH-funded grantees to submit a data management and sharing plan on how researchers intend to preserve and share their data [
In 2015, following the publication of the IOM consensus study,
Other recent publications have more specifically assessed and advised the practices of funders in data sharing. In one retrospective review published by Whitlock et al [
Although data sharing has gained momentum, major challenges remain, including transparency of data access procedures, reuse of consent provisions (ie, policies that accommodate participant consent regarding the volume, nature, and timing for secondary use of clinical trial data for future scientific research), governance of data sharing policies, availability of affordable technical infrastructure, alignment on data standards, and a highly fragmented landscape of data repositories [
The purpose of this paper is to detail a set of aspirational goals and forward thinking, collaborative solutions to data sharing for nonprofit funders to fold into existing funding policies. This paper was developed by the Sharing Clinical Trial Data Action Collaborative, an ad hoc activity associated with several forums at the National Academies of Sciences, Engineering, and Medicine: the Forum on Drug Discovery, Development, and Translation; the Roundtable on Genomics and Precision Health; the Forum on Neuroscience and Nervous System Disorders; and the National Cancer Policy Forum. The paper does not necessarily represent the views of any one organization, the Forums or Roundtable, or the National Academies and has not been subjected to the review procedures of, nor is it a consensus study report or product of, the National Academies.
Throughout this paper,
To discuss and advise on the development of data sharing goals for nonprofit funders of clinical trials, an ad hoc group of individuals with relevant expertise representing nonprofit organizations and other stakeholders, including patient and disease advocacy representatives, researchers, regulators, and drug developers, was convened in July 2016 (n=9) and November 2017 (n=39) in Washington, DC, associated with the Sharing Clinical Trial Data Action Collaborative of the National Academies of Sciences, Engineering, and Medicine. Invited participants were identified by project coleaders, Sharon Terry (Genetic Alliance) and Timothy Coetzee (National Multiple Sclerosis Society), and with the research assistance of National Academies staff, to represent a subset of nonprofit funders of clinical trials interested in exploring and implementing data sharing policies. Over the course of planning for the 2016 and 2017 discussion meetings, 70 individuals were invited and 48 accepted and participated in either or both the 2016 and 2017 meetings.
The 2016 meeting focused on gathering feedback on a draft set of data sharing principles for nonprofit funders drafted by Sharon Terry, Tim Coetzee, and National Academies staff working with Action Collaborative and drawn largely from themes in the 2015 IOM report,
The National Academies provided a neutral venue for nonprofit funders to have a candid conversation about the opportunities and challenges of taking up formal data sharing policies within their organizations. Participants in this activity had an opportunity to share perspectives, sharpen questions, spark new ideas, and explore possible solutions. The participants did not cover a full representation of the nonprofit community. The goals included in this paper are not binding agreements but rather indications of support for advancing clinical trial data sharing among the nonprofit community.
Patient communities and the lay public should have a voice in the development of data sharing policies regarding what and how data will be shared with others, including, but not limited to, IPD, participant-generated data (eg, patient-reported outcomes, patient data acquired from wearables), lay summaries, and resulting publications [
Embedded in this goal is the collective duty of nonprofit funders to prioritize educating patient communities about data sharing and build an informed public trust around the value of clinical trials and data sharing.
Nonprofit funders should strive to prioritize data sharing in the earliest conceptions of a clinical trial. The ultimate utility of shared clinical trial data often hinges on the degree to which sharing was planned for from the beginning—a principle also endorsed by the multistakeholder task force on data sharing, as described by Ohmann et al [
Nonprofit funders can incorporate data sharing principles into grant application policies. For example, grant provisions could require that the proposed clinical trials include a plan for how and when data from the trial will be shared and that clinical trials be registered on a public data sharing platform (eg, ClinicalTrials.gov). In addition, nonprofit funders may set expectations for awardees to deposit clinical trial data in public databases and publish results of the trial regardless of the outcome.
Nonprofit funders should work with grantees and patient communities to ensure that data access policies (1) facilitate the appropriate use of shared data, (2) enable research participants’ data access preferences, (3) protect participant privacy, and (4) mitigate risks to the scientific integrity of investigators and sponsors that share data (ie, reducing the likelihood of misuse or misanalysis of shared data)—all without unduly restricting access to shared data. This goal echoes broader data sharing goals of making data access processes more explicit and transparent, and funder actions made to achieve this goal would additionally demonstrate support and alignment with FAIR principles [
Data access policies may vary based on the type of data being collected in a clinical trial and the preferences of the community. For instance, IPD with imaging and genetic information may warrant a third-party intermediary to review requests, as opposed to open access for anyone. It is difficult to make one-size-fits-all decisions on who should be authorized to have access to data, but variations are possible. For instance, platforms and technology exist to enable individuals to decide who should access their data.
Nonprofit funders could encourage or require grantees to provide the following information and respect the following policies as objective signals of the FAIR use and intent:
Establishing a plan or proposal that states the purpose of the data request (eg, to support hypothesis generation or protocol development)
Providing evidence via a standard biographical sketch of the qualifications of the requestor
Using data use agreements that may help ensure that data requesters follow the plan stated in the original request and do not attempt to use data in harmful or malicious ways, such as reidentifying participant data or using data for commercial or litigation purposes
Sharing clinical trial data, particularly of sensitive nature, in a way that it can be housed, accessed, and analyzed behind a firewall or other secure mechanism
Third-party review teams can vet data requests to provide an independent, transparent, accountable, and efficient data access review process (
Can the data requested support the stated purpose of the requestor?
Does the request have a public health or health goal and address real patient needs?
What technology infrastructure is used to provide data access, and who is the data steward?
What data standards are used to share the final data set?
What data versioning processes does your data steward recommend?
What data variables and data types are shared and over what length of time?
What security measures are required for your shared data?
What is detailed in the data documentation (ie, data dictionary, schemas, example analyses, and use)?
Requiring that data be shared but not providing a place to host shared data is an impractical mandate, and the recognition of the general need to properly structure and build a suitable and sustainable infrastructure for data sharing has remained an important recognition [
Many nonprofit funders acknowledge that developing and maintaining the technology to support a data repository is beyond their skill set. Thus, nonprofit funders will increasingly be looking to external platforms—and partnerships with technology companies—for hosting shared data. Nonprofit funders should work together to collectively form partnerships and provide support (eg, funding, guidelines, and requirements) for data sharing platforms in accordance with the needs and goals of their communities. Although each community will still have unique needs and expectations, there are significant similarities among the desired specifications of a data sharing infrastructure and much to be gained from ensuring that data are not unnecessarily siloed.
Within their organizations, nonprofit funders should develop and implement data sharing guidelines and/or policies that detail requirements for storage, curation, standards, documentation, sharing, and archiving of data produced by grantees. Nonprofit funders are encouraged to use the goals outlined in this paper as a framework for such guidelines or policies. When possible, nonprofit funders should support the training of grantees on the tools and methodologies for sharing and appropriately analyzing data.
Standards relevant to research data sharing encompass a number of types, including common data models; transport or exchange formats; metadata standards and analysis standards; data elements; terminologies and vocabularies, ontologies, and code lists. Organizations such as the Clinical Data Interchange Standards Consortium, TransCelerate, and the Critical Path Institute continue to develop and maintain the data standards, tools, and methods needed to support clinical trial data sharing [
Nonprofit funders should, at a minimum, require that investigators have a data sharing plan in place before enrollment of the first participant that dictates when, what, with whom, how, and under what circumstances the data will be shared. Nonprofit funders can direct grantees to applicable guidelines for specific provisions of data sharing plans, in accordance with the needs of each nonprofit funders’ organization, in general, and the clinical trial at hand, in particular [
The need to improve incentives has also been emphasized by others, and nonprofit funders are in a unique position to incentivize researchers by promoting the recognition of investigators who share their data [
Data are valuable and often one of the greatest assets a nonprofit organization has. However, nonprofits are encouraged to maintain a clear focus on their mission and resist the temptation to unnecessarily keep data internal to their organization. Nonprofit funders should consider how to incentivize grantees to use externally authored shared data sets to improve the proposed clinical research design (eg, to better estimate the effect size of the proposed treatment;
Publish a list of top data sharers and shared data users for usage analytics, either within their organization or in partnership with other organizations.
Highlight the success stories of data sharing or the use of shared data by grantees in articles in high-impact publications.
Provide credit and/or rewarding grantees who share data and use shared data.
Engage in open science efforts that incentivize the use of publicly available data, crowdsource challenges in medicine, and share the data and insights gleaned from the work (eg, DREAM Challenges) [
Educate both patient and research communities on the benefits of data sharing to create an expectation of data sharing in clinical research.
Implementing data sharing involves costs, but the precise cost is often a challenge to unearth. Nonprofit funders face difficult decisions, including whether to prioritize and pay for data sharing, thereby funding fewer new grants for research. Nonprofit funders could include, as a provision in grants and contracts, the allocation of funds to support sharing the data produced by the funded research. This could entail a line item dedicating a set amount of funding, or a percentage of overall funding, to data sharing efforts. Even if they do not provide the full funding for a particular trial, organizations may use their leverage to insist on such terms in grants and contracts before promoting trials to their networks. Organizations are broadly encouraged to incorporate data sharing costs into their respective funding models, maintaining a diverse, mission-aligned portfolio. If practical, and conducive to the organization’s policies, charging a fee for access to data could help defray the cost of data sharing.
Mechanisms by which nonprofit funders could
Withholding a portion of the allocated funding until certain benchmarks of data sharing plans are realized
Including a neutral third party as an honest broker to administer the sharing of data in a responsible manner
Promoting the clinical trial to the nonprofit funder’s patient or disease advocacy networks contingent on inclusion of language stipulating a data sharing plan
Requiring a data sharing plan as part of any funding request, which includes an appropriate level of detail demonstrating specific steps to comply with the funder’s data sharing requirements
In addition, nonprofit funders may facilitate the adoption of data sharing agreements by offering examples of organizations providing funding to support data sharing, such as NIH (eg, NIH grantees are permitted to charge the salaries of administrative and clerical staff as a direct cost to help investigators meet their responsibilities under the NIH policy on reporting research results) [
As part of the decision-making process surrounding funding or support of clinical trials, nonprofit funders should request that prospective grantees with a history of sharing data provide evidence and impact of previous data sharing and, to the extent possible, invite grantees to provide evidence of the impact of earlier data sharing. Such evidence could include new collaborations, publications, novel analysis or findings, or the evidence that emerges from secondary monitoring of usable spaces to see who is active in the community and contributing to knowledge generation. Nonprofit funders might consider conducting pilot projects to evaluate the feasibility and identify the challenges of including prior data sharing as an impact measure.
The goals in this paper convey the complexity of the opportunities and challenges facing nonprofit funders and seek to provide a starting point for a data sharing
There are costs associated with data sharing activities; therefore, it will be important to ensure that the aspirational goals expressed in this paper do not create an undue burden on researchers, nonprofit funders, and trial participants. Some of the data sharing approaches that have been developed by the participant and advocacy community may help relieve investigators of administrative burden by streamlining data storage, curation, sharing, and archiving processes. In addition, data sharing platforms have the potential to improve aspects of clinical trial research that go beyond data sharing alone (eg, enhancing participant recruitment, engagement, retention, and encouraging collaboration).
Nonprofit funders, along with other key stakeholders, play an important role in ensuring the responsible sharing of clinical trial data. The goals in this paper offer a path forward for nonprofit funders to continue to take steps, even if incremental, toward a more robust data sharing ecosystem.
Alzheimer’s Association
Center for Open Science
Genetic Alliance
Critical Path Institute
Doris Duke Charitable Foundation
Food Allergy Research & Education
Geoffrey Beene Foundation Alzheimer’s Initiative
Global Healthy Living Foundation
Michael J. Fox Foundation for Parkinson’s Research
Multiple Myeloma Research Foundation
National Health Council
National Multiple Sclerosis Society
National Psoriasis Foundation
New York Stem Cell Foundation
Open Humans Foundation
Pancreatic Cancer Action Network
Parent Project Muscular Dystrophy
Parkinson’s Foundation
Parkinson’s UK
Susan G. Komen
The V Foundation for Cancer Research
Vivli
findable, accessible, interoperable, and reusable
Institute of Medicine
individual participant data
National Institutes of Health
Promise for Engaging Everyone Responsibly
The authors would like to thank the following individuals for their participation in meetings and discussions of the Action Collaborative: Caren Heller, Crohn’s & Colitis Foundation, Simon Noble, CHDI management/CHDI foundation; and Heather Pierce, Association of American Medical Colleges.
Author BAN is the Executive Director of the non-profit Center for Open Science that operates the Open Science Framework (http://osf.io/) and has a mission to increase openness, integrity, and reproducibility of research; Author IS is the Co-Founder and consultant to Vivli. Author AFH has received research funding from American Regent, Astra Zeneca, Boehringer Ingelheim, Merck, Novartis, and Verily and is also a consultant to Amgen, Astra Zeneca, Boehringer Ingelheim, Boston Scientific, Bayer, Merck, and Novartis. All other authors have no conflicts to declare.