The What and How of Data Sharing

In a recent BMJ article, Milton Packer highlights the value of data sharing using iconic scientific figures: Copernicus, whose heliocentric theory of the universe was built using others’ data; his rival, Brahe, who hoarded data for fear of confirming Copernicus’ theory; and Kepler, whose famous laws of planetary motion depended on data sharing.  The National Institutes of Health (NIH) has long been a leader in data sharing, and there is a clear clamor for more and better data sharing by NIH and other federal agencies. There is little doubt that sharing biomedical research and health-related data plays a key role in advancing knowledge of human health and well-being. But data sharing is not without cost, and data shared in ways that are not useful to the research enterprise can waste, rather than maximize, resources. This is why we need help from you, the research community, to help us shape our data sharing strategies for the future.

Consistent with federal initiatives promoting open data and open science, NIH continues to be committed to ensuring that, to the maximum extent possible, the results of federally-funded scientific research are made publicly to support reuse, reproducibility and discovery. In order to move forward with ongoing commitments to the data sharing enterprise, we are considering priorities for data management and sharing and how to expand upon existing data sharing policies, such as our 2003 Data Sharing Policy. However, we recognize that many factors must be considered when determining what, when, and how data should be managed and shared including, for example, the purpose for sharing, supporting data re-use and reproducibility, maturity of the science, the infrastructure uniqueness of the data, and ethical considerations.

Today we are publishing a Request for Information (RFI) related to strategies for data management, sharing, and citation.  Through this RFI, we are seeking stakeholder feedback on considerations pertaining to what types of data should be shared, the costs and benefits of sharing different types of data, and standards for citation of data and software. Your feedback will help us to prioritize our thinking in data sharing stewardship and be considered as we move forward in developing new NIH policies in this area.

We need to hear from data users, data generators, and data scientists. By assisting us in this request for information, you can help ensure that NIH has the most robust set of information on hand when making future decisions in this important arena.  I encourage all interested stakeholders to review the RFI and provide us with their thoughts.  Comments on the RFI will be accepted until December 29, 2016.


    I am now-former Johns Hopkins research faculty, in part because an NIH worker would not share his data with me (after the paper using the data had been published for more than one year). Whatever is good for the goose is good for the gander: NIH policy on data sharing by grantees should apply to NIH staff. Email me for details.

Leave a Reply

Your email address will not be published. Required fields are marked *