This is a website for an H2020 project which concluded in 2019 and established the core elements of EOSC. The project's results now live further in and

Data sharing done right

Rob Baxter writes about EOSC-hub’s recommendations to facilitate sharing and processing of sensitive data

Human health is a global issue. Disease knows no border. Today we have more digital health data than ever before, but sharing them across borders for the greater good of global research remains challenging. Public good must still be balanced against the privacy of individuals.

Making it easier to access and share research data of all kinds is the central principle of the European Open Science Cloud — and this includes health and other “sensitive” data that must only be shared under strict supervision and with the utmost care. Through EOSC-hub (and in EOSC more widely) we have an opportunity to lead the world on making sensitive data safely available for research.

In EOSC-hub we’ve just completed a set of draft recommendations for policies on data sharing (D2.8 First Data policy recommendations). In this we were lucky with our timing — both the EOSCpilot project and the EC Expert Group on FAIR data came out with excellent policy reports covering ethics, and the open and FAIR data agendas. We were able to take a “next-steps” approach and identify what practical actions EOSC-hub might take to advance both the FAIR agenda and the ethical sharing of sensitive research data for public benefit.

We adopted 11 key recommendations from EOSCpilot and translated them into 22 practical suggestions on data sharing across EOSC-hub, covering the spectrum of data from the open and public to the highly controlled and sensitive. These suggestions aren’t final yet but are up for consideration by strategists and technical system integrators in EOSC-hub. This time next year will see the final versions published.

Right now the recommendations fall under three broad headings:

  • Implement FAIR, taking a “Web first” approach to implementing the FAIR principles: data objects should be published on the Web in open, non-proprietary, machine-readable formats, well described and referenced by resolvable persistent identifiers.
  • Build technical expertise in safe data and safe settings, adopt the “Five Safes” principles of safe data, safe settings, safe projects, safe people and safe outputs, and work towards enabling continent-wide research that follows them.
  • Support the wider development of ethical and information governance frameworks. In EOSC-hub we have a terrific opportunity to engage with a wide set of stakeholders,including social science and statistical data service providers, and the emerging EOSC governance function, to build a strong consensus and strong processes for cross-border research using sensitive data.

Over the coming year we’ll chew these over as a project, figure out what’s feasible, what we can do now and what will have to wait. Sharing sensitive data across borders properly will be difficult, but the prize is surely worth it!


Rob Baxter is Programme Manager at EPCC, the University of Edinburgh, and leads the EOSC-hub task on Data Sharing Policies

News type: