ABSTRACT

The Yale University Open Data Access (YODA) Project acts as a data intermediary to facilitate the sharing of clinical research data between members of academia, government, and private industry for the purposes of conducting meta-analyses, replicating trial results, building upon prior findings, and conducting secondary analyses. Notable data partners include Johnson & Johnson, Medtronic, Queen Mary University of London, and SI-Bone, Inc (YODA, 2014). The YODA Project emerged to address issues in the clinical trial research community with regard to data transparency and integrity. A linchpin of its work is clearly defined, stakeholder-driven policies and procedures for data access and use, which have helped inform other data-sharing efforts and propelled the field toward embracing open access principles. Guided by a commitment to transparency, YODA has demonstrated the benefits of open access, promoted the responsible conduct of research, and eased a key barrier to sharing data between researchers and industry collaborators.

BACKGROUND

The YODA Project was conceived in 2011 as an effort to address issues within the clinical trial research enterprise. At the time, the program’s co-directors had identified, through the course of their research, a number of challenges with regard to research integrity, transparency, and dissemination. For instance, their research demonstrated widespread problems with selective publication, as it took up to five years for two-thirds of completed clinical studies to be disseminated, and the remaining one-third were not disseminated at all. Simultaneously, Medtronic approached the co-directors to solicit an independent study replicating the results of the company’s proprietary recombinant human bone morphogenetic protein-2 marketed for back pain (Ross et al., 2018). The opportunity helped position the YODA Project as a coordinating body for independent study reviews, offering private entities a mechanism to validate their research externally.

Box

Case Study at-a-Glance: Yale Open Data Access (YODA) Project.

An important emerging dimension was a focus on data sharing. Following the conclusion of the Medtronic studies, the YODA Project convened a multi-stakeholder advisory group, consisting of regulators, industry leaders, clinical trial researchers, and patients, to discuss the merits of sharing data and what the infrastructure needs would be. The group contemplated a range of considerations, including the concern that the published data could be misused for commercial or litigious purposes or data dredging. However, the consensus among patients was that as long as their identities were protected, they supported the practice. In this spirit, the group decided to adopt an open data policy, and a significant outgrowth of the advisory group’s engagement is the development of the YODA Project’s Policies and Procedures.

The YODA Project requires data partners to make available raw, de-identified datasets, as they would with the FDA. In addition, the data are not limited to Phase 3 randomized control trials, but instead include all clinical trial data, regardless of whether they have been published or disseminated (YODA, 2021). The YODA Project’s data-sharing pathways emulate National Heart, Lung, and Blood Institute’s Biologic Specimen and Data Repository, or BioLINCC (NHLBI, 2021). In the same vein, requesters are granted access to the data files along with the study protocols and reporting analysis plans through a, secure enclave (YODA, 2019). Early on, the YODA Project addressed the burden of data preparation by deciding against accommodating custom data requests, such as imaging data, which are difficult to transform into a tabular format. When asked if the decision foreclosed any opportunities, Joseph Ross, Co-director of the YODA Project, who participated in this interview, responded by saying that it strives to be a one-stop, nimble shop for data use. Entertaining specialized data requests would be too cumbersome for the initiative and detract from its appeal. He pointed out, though, that, if there is a need for specialized datatypes, the YODA Project could facilitate contact between the requester and data provider.

DESCRIPTION

Requesters initiate access by submitting an online application. Along with submitting a research proposal, they are required to review the YODA Project’s policies and procedures, complete a conflict of interest disclosure, and view a training video on the data use agreement (DUA). Data inquiries are thoroughly reviewed by a team of clinical investigators associated with the YODA Project. A key criterion of the evaluation process is whether the requested datasets match the intent of the project. Proposals also undergo blinded reviews by the data partners to ensure that the provided data are rendered appropriately for the request.

• Receipt of the data is contingent upon signing a DUA. DUAs are notoriously complex, nuanced, and can undergo multiple iterations; the more bespoke these agreements are, the more time they take for approval. To solve this issue, the YODA Project created three templates: one designed for U.S.-based researchers, another for researchers affiliated with the U.S. government, and the last one for foreign research entities. YODA’s DUA circumscribes the following elements (YODA, 2019).
• The data must be used in service of the project goals as described in the research proposal.
• Data users agree to allow YODA to publicly post project proposals on its website in recognition of open science.
• Requesters are forbidden from redistributing or publicly posting the data online on their own channels.
• Beyond the completion of a project, investigators can access the data for up to five years.
• The YODA Project permits investigators to disseminate their findings in a peer-reviewed journal, as a preprint, or at scientific conferences.
• In pursuit of transparency, the YODA Project calls for data users to share a copy of their abstract with the project’s data partners.
• Data requesters also are required to report their results if they choose not to publish their findings. Access to the data is pro bono. The YODA Project is supported by grants provided by participating pharmaceutical and biomedical device companies.

The YODA Project leaders noted that the data-sharing mandates issued by the European Medicine Agency and the 2007 FDA Amendments Act, collectively, have put “wind in their sails.” Additional impetus came from the Pharmaceuticals Research and Manufacturers of America and the International Council of Medical Journal Editors, which issued consensus statements in support of data sharing, signaling a shift in attitudes toward research integrity and transparency.

Despite initial hesitancy from the health care community to open access to the data, the YODA Project’s first successful demonstrations were vital to engendering stakeholder confidence. As YODA’s codirector noted, once it was shown that the “sky did not fall when others had access to the data,” the community gained a sense of appreciation that compelled other companies to join or feel more comfortable in their own data-sharing efforts. Notably, Johnson & Johnson’s involvement, given its standing in the industry, helped to elevate the organization’s profile.

The YODA Project’s transparent and public processes are paramount to upholding relationships with data partners. The intent is to sustain engagement with data partners and foster an ethos of open science collaboration in the health care ecosystem that extends beyond the interactions occurring within the YODA Project. The YODA Project has started to witness signs of further adoption of open access data policies through changes mandating the sharing of data from research funded by the National Institutes of Health and the Patient-Centered Outcomes Research Institute. These changes contribute to the sentiment that data sharing, including patient access to data and cooperation between researchers, are becoming more of a common expectation (Dey et al., 2017).

FUTURE DIRECTIONS

The YODA project recognizes that the sharing of clinical research data is a cultural issue in that there are variable norms, incentives, and expectations across the research community with respect to the inherent value of sharing, and that stakeholders hold differing views on its relative benefits. Data has served as an important scientific currency in academia; consequently, there may be persisting reluctance to share it without an explicit mandate to do so. Acknowledging the potency of policy levers, the YODA Project believes that a mandate and portal akin to ClinicalTrials.gov for the sharing of privately sponsored research could compel the field to move toward this goal. While current data-sharing conditions may not be ideal, the YODA Project’s leadership offers optimism for the field, imparting the advice that having a vision and unwaveringly advocating for that vision is essential. Equally significant is the imperative to cultivate trust from the outset by responding to concerns from stakeholders and being transparent about not only the approach to a policy or process, but the rationale behind the approach. Understanding the “why” can accrue trust about the “what.”