Updated 3 min read

Why does open source matter for health data?

Trust in health data infrastructure usually rests on promises: a privacy policy, a certification logo, a vendor’s word. Open source replaces one class of promise with something stronger: code anyone can inspect.

The case in three points

Verifiability. When the software holding your medical record is open source, its privacy behaviour can be checked rather than merely promised: independent experts can read exactly what happens to the data. That is the oldest principle in security engineering: protection must not depend on the system staying secret. Open code is the precondition for such an audit, not a substitute for it; the evidence shows that openness alone does not make software more secure, review does.

No lock-in. Health data outlives every vendor. Open source makes platform independence possible: if the company or foundation behind the software disappears, the code survives, and open data formats keep your ability to migrate intact, which is why the DPG Standard checks both platform independence and non-proprietary data export.

A commons, not a moat. Closed platforms treat infrastructure as a competitive asset, which is one root of the fragmentation problem in health data. Open infrastructure lets clinics, researchers and app developers build on shared rails instead of rebuilding (and re-fragmenting) the same machinery.

What certification adds

Open-Pryv, the platform Health Data Safe builds on, has been recognised as a Digital Public Good by the Digital Public Goods Alliance (DPGA) since 19 April 2024. The DPGA is the multi-stakeholder alliance whose formation the UN Secretary-General welcomed in his 2020 Roadmap for Digital Cooperation. The DPGA Standard is a useful checklist of what “trustworthy open infrastructure” means: open licence, open standards, privacy compliance, platform independence, and alignment with the UN Sustainable Development Goals. HDS has adopted Open-Pryv as the technological foundation of its platform and is committed to its open-source development.

Open source with opinionated defaults

Openness alone does not protect anyone; defaults do. Open-Pryv ships twelve structural privacy defaults, privacy-protective behaviours that cannot be misconfigured away, starting with default-deny permissions: no data is accessible until the person explicitly grants access. The platform models 251 data-point types and keeps an invariant audit log, so consented research data is structured and every access is traceable.

Open code makes the guarantees inspectable; structural defaults make them hard to break, even by the operator.

Read the story of why HDS embraced Open-Pryv, or see how consent works in practice in granular consent explained.

Sources