STAYS AT SOURCE
The identifiable record
Names, contact information, local identifiers, raw clinical notes and re-identification keys remain within the originating custodian’s boundary.
DISCOVERY VAULT / PLANNED
Keep custody.
Unlock collaboration.
Make valuable research resources discoverable without turning patient records into a shared commodity. Vault connects a specific research purpose to a carefully prepared, anonymized view of the data.
Explore the custody modelREQUIRED DATA BOUNDARYArchitecture
Originating institution
Identifiable source records
Local custody and access controls
PREPARE · VALIDATE · APPROVE
Anonymized research view
Purpose-limited release
No patient identity or linkage key
CUSTODY BEFORE CONNECTIVITY
A deposit starts a responsibility.
It does not grant a download.
A useful research resource needs a steward, a purpose and a release decision. Those responsibilities stay visible as more collaborators join.
THE VAULT PRINCIPLEDiscover the resource.
Discover the resource.
Respect the person.
Vault is the bridge between an institution that holds information and a team that can create research value from an approved, anonymized derivative.
See the first pilot- 01
Register the research resource
Describe provenance, structure, permitted research purposes, responsible stewardship and retention. Review catalog metadata too: a resource description can itself disclose sensitive information.
- 02
Prepare inside the boundary
Filter direct identifiers, inspect free text and imaging metadata, assess indirect identifiers and combinations, and establish the applicable de-identification method. Keep linkage keys at the source.
- 03
Approve a specific research view
Bind the reviewed artifact to its purpose, recipients, location and expiry. A deposit receipt, sign-in or proof is never a substitute for permission and anonymization review.
- 04
Control outputs and future use
Record approved access and output release, enforce expiry and stop future use after revocation where applicable. Explain retained results and trained-model limitations before work begins.
THREE DIFFERENT THINGS
What stays.
What can be shared.
The medical sharing contract is anonymized-only. Pseudonymized records remain protected records when identity can be recovered through a key or other information.
REQUIRES RELEASE REVIEW
The research view
Only an anonymized derivative that passes the applicable assessment is eligible for a specifically authorised collaboration. A failed or uncertain assessment blocks sharing.
PROVE ONLY WHAT IS NEEDED
The bounded claim
Selective disclosure and zero-knowledge proofs can be designed to establish a defined eligibility or authority claim without exposing the underlying attributes. The proof scope needs its own validation.
BUILT FOR THE PEOPLE RESPONSIBLE
Stewards and researchers.
Working from the same terms.
Institutional data teams define the boundary. Academic and life sciences collaborators bring a precise question, a reproducible method and an output-review plan.
- For the data steward
- Retain custody, approve the anonymized view and know which research purpose each release serves.
- For the researcher
- Understand suitability and limitations before requesting a resource. Work with approved information rather than assuming a right to the source record.
- For the participant
- Keep participation choices specific. A resource listing or institutional agreement is not blanket permission to reuse an individual’s information.
- For the reviewer
- Inspect the release artifact, method, authority, expiry and output assessment as one accountable decision.
CLEAR ANSWERS
Understand the
boundaries.
Is shared research data anonymized?
Yes. Anonymized data only is the required model for research sharing in Hybrid Discovery. Identifiable patient records stay with the originating custodian; they are excluded from cross-clinic sharing and third-party AI requests. Medical sharing and anonymization services are planned and must pass validation before launch. No patient data is collected or shared by this website.
Can I deposit a medical dataset today?
No. Medical deposits, preparation, anonymization review and research access are planned integrations. This website has no upload or dataset-access endpoint.
Does encryption make a dataset anonymous?
No. Encryption protects access to information; anonymization addresses whether people can be identified. Both have different roles and must be evaluated separately.
ADVANCE DISCOVERY TOGETHER
Make the next dataset
a responsible collaboration.
Begin with one resource, one purpose and evidence that the boundary works.