AMIA Supports New NIH Data Policy, Encourages Phased Implementation

Wednesday, December 12, 2018

Nation’s biomedical and health informatics experts want more thorough evaluation of grant applications’ data plans

In comments submitted earlier this week, the American Medical Informatics Association (AMIA) supported an expansive update to a 2003 National Institutes of Health (NIH) data sharing policy. The organization of biomedical and health informatics experts recommended a phased approach to implement new data management and sharing requirements for NIH-funded research and noted that quality data management and sharing plans are prerequisite to the NIH’s goals of making research data Findable, Accessible, Interoperable and Reusable.

The NIH established its current data sharing policy in 2003, requiring 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. Additionally, several NIH Institutes and Centers (ICs) have developed their own policies with varying requirements since 2003.

A Request for Information (RFI) issued by the NIH in October sought feedback from stakeholders on how their existing policy should be updated and how new provisions of a Data Management and Sharing Policy (DMSP) should be implemented. Notably, the RFI asked for input on a DMSP that would apply to “all intramural and extramural research, funded or supported in whole or in part by NIH, that results in scientific data, regardless of NIH funding level or mechanism.”

AMIA commented that quality data management and sharing plans are “prerequisite” to achieve the vision of FAIR data principles and such a scope should be the long-term goal of the NIH DMSP. AMIA also noted that a key deficiency in the current policy was that grant applications are not scored on the quality of their data sharing plans. This has led to suboptimal and incomplete sharing plans, and likely has contributed to increased scrutiny over whether and how NIH-supported research data is made available to other researchers and the broader public.

While supportive of an NIH-wide DMSP, subject to IC-specific grant-types and awardees, AMIA strongly recommend that the DMSP encourage, if not require, ICs to factor the quality of data management and sharing plans into the overall impact score through the peer review process, especially for those grants that are supported at high levels or support programmatic priorities. Further, AMIA also recommended that the NIH incentivize deposition of scientific data in NIH-endorsed databases and knowledgebases by allowing such applicants to comply with streamlined DMSP requirements.

“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.”

In addition to improving the “FAIR”ness of data funded by the NIH, AMIA said the DMSP should improve institutional support and professional advancement for experts managing and sharing scientific data.

Specifically, AMIA applauded the NIH for suggesting that reasonable costs associated with data management and sharing could be requested under the budget for the proposed project. AMIA recommends that the DMSP establish a standard way to account for data management and sharing costs as both Direct and Facilities & Administration costs.

AMIA encouraged the NIH to develop a draft DMSP for further input and that the draft should facilitate implementation of the NIH Data Science Strategic Plan, especially aspects of the Strategic Plan that seek to credit experts who manage and share valuable data sets / software for their work.

“If data is seen as valuable, experts who enable FAIR data should also be valued,” AMIA comments said. “The NIH should support certifications for experts that manage and share scientific data…and there is a need for R&D on data management tools to facilitate compliance with the DMSP.”

Click here for AMIA’s full set of recommendations.