- The American Medical Informatics Association (AMIA) has issued comments in support of an expansive update to NIH’s data sharing policy, and recommended a phased approach to implement new requirements for NIH-funded research across the healthcare industry.
NIH established its current data sharing policy in 2003. The policy requires grants that receive more than $500,000 annually to develop data sharing plans. However, these plans are not factored into the overall score of the grant application.
Moreover, several NIH Institutes and Centers (ICs) have developed their own policies with varying requirements since 2003, making it difficult for researchers to understand the requirements for data sharing across the landscape.
In October 2018, NIH issued a request for information to update its existing data sharing policy. The agency asked stakeholders for input on new provisions of a Data Management and Sharing Policy (DMSP) should be implemented.
In response to NIH’s request, AMIA voiced its support of a new DMSP, and applauded the NIH for including some of AMIA’s own data sharing principles.
“AMIA enthusiastically supports development of a pan-NIH Data Management and Sharing Policy (DMSP) and we commend the NIH for initiating this effort,” the organization wrote in its comments.
“We are pleased to see several elements of AMIA’s Data Sharing Principles & Positions incorporated in the Proposed Provisions, including a reliance on FAIR (Findable, Accessible, Interoperable, and Re-usable) data principles, and acknowledgment that the DMSP should support underlying infrastructure and curation activities with funding.”
The organization recommended a phased compliance timeline based on funding levels, which would only apply to new research funded after the DMSP is final. AMIA also suggested that the DMSP encourage NIH institutes and centers to factor the quality of data management and sharing plans into overall impact scores, especially for those grants that are supported at high levels.
Additionally, AMIA suggested that the draft DMSP should enhance data sharing and management in a way that facilitates learning health systems and continuous discovery. The Association advised that the new policy include data standards and best practices to improve interoperability and support further use of research data.
“AMIA recommends the DMSP seeks to improve the interoperability and supplemental uses of research data writ large by encouraging the use of established biomedical data standards and adherence to data management and data sharing best practices,” the organization said.
“Over time, better use of and refinement of data standards, buttressed by systematic scoring of plans, will optimize scientific data for continuous learning and discovery.”
To help both small and large grantees more easily comply with the DMSP, AMIA suggested that the new policy incentivize the deposition of scientific data and tools into NIH-approved data repositories and knowledge bases.
The Association advised that the DMSP should increase institutional support and professional advancement for experts managing and sharing scientific data.
“If data is seen as valuable, experts who enable FAIR data should also be valued. The NIH should support certifications for experts that manage and share scientific data. We also see a need for R&D on data management tools to facilitate compliance with the DMSP,” AMIA wrote.
The DMSP should also facilitate implementation of NIH’s Data Science Strategic Plan, especially the aspects of the plan that seek to credit experts who manage and share valuable datasets for their work.
AMIA also called for more specificity and clarity around concepts in the DMSP.
“Data management is distinct from data sharing. The processes and activities that support data management and sharing are also different. AMIA recommends the NIH develop a DMSP that specifies these distinctions,” the organization said.
AMIA called on NIH to expand its current list of concepts to include definitions for “data management” and “covered data,” and to refine definitions for “metadata” and “scientific data.”
In addition, the organization said that while it does support the scope of a DMSP that covers all grants and contracts, AMIA recommends that NIH convene stakeholders with individual ICs to operationalize the DMSP.
Finally, AMIA suggested that NIH evaluate and consider these comments before creating another draft DMSP.
“We offer AMIA and its members as resources during subsequent work on the DMSP. We strongly recommend the NIH develop a subsequent draft DMSP based on stakeholder feedback to the concepts in this RFI. Another comment period will provide NIH with valuable insights before issuing a final DMSP,” the association stated.
With these comments, AMIA expects to help NIH enhance data management and sharing.
“Data management and sharing are vital to continuous discovery and learning,” said AMIA CEO and President Douglas B. Fridsma, MD, PhD, FACP, FACMI.
“Setting expectations for how NIH-supported research should be managed and shared, using the long-established peer-review process to ensure best practice, will greatly improve the downstream utility of our national investment in NIH research.”