AuroraMed v0.5.0 is a synthetic-data pilot. See exactly what's built →

Privacy fundamentals

The privacy model

US-style compartmentalized privacy by default, with optional strict privacy for Latin America.

How to read status labels: In the pilot (v0.5.0, synthetic data) Simulated in the pilot (no real external system) Designed, not yet built Open decision Not offered / no claim made Planned partner integration

C0021What privacy model does AuroraMed follow?

US-style compartmentalization is the default: role, relationship, purpose, context and risk must all line up before access. Latin American customers can switch on a stricter optional tier. Legal minimums in each country still apply regardless of tier.

Designed, not yet builtPermalink

C0022Is strict privacy a paid add-on?

For Latin America it is decided as a switchable, optional paid tier. It goes beyond legal minimums. The baseline tier has to satisfy local law on its own, pending counsel review.

Designed, not yet builtPermalink

C0023What does the strict tier add?

The design adds keys held only in the clinic, cloud sees ciphertext only, per-patient keys everywhere, patient-signed consent for every cross-site access, and enhanced audit visible to patients. It is designed, not built.

Designed, not yet builtPermalink

C0024Does a cheaper plan weaken privacy or security?

It must not. The specification says tiers change price and scale, never the legal-minimum privacy and security floor.

Designed, not yet builtPermalink

C0025What is a compartment?

A slice of the record with its own keys and policy, such as mental-health, substance-use, HIV, reproductive, genetic, minors, or VIP. The general chart shows that restricted data exists and needs a logged reason to open.

Designed, not yet builtPermalink

C0026Can staff see VIP patients freely?

In the pilot, VIP and restricted patients are hidden from anyone without an explicit care assignment. Break-glass is possible but raises a security alert and a mandatory review.

In the pilot (v0.5.0, synthetic data)Permalink

C0027Who can see a patient's Social Security number?

The pilot masks it. Revealing it needs a stated purpose, is audited, and shows for only a few seconds.

In the pilot (v0.5.0, synthetic data)Permalink

C0028Is patient data ever sold or used to train AI?

The design forbids monetizing patient data and secondary use without explicit consent and legal basis. AI inference is designed to run on site or in an approved region, and patient data may not train models without separate written approval.

Designed, not yet builtPermalink

C0029Where is patient data stored?

In the pilot, on a cloud hub, with an on-site hub (Node 22 and SQLite) that replicates with it, and only synthetic data. The design goes further: a per-clinic closed on-site server as the operational primary, with an encrypted backup copy in the cloud. That inversion is not built.

Designed, not yet builtPermalink

C0030Is patient data stored on the ledger?

Never. Only hashes, consent records and audit records go on the federated ledger, so deletion and correction rights can be honored. The ledger is designed, not yet built.

Designed, not yet builtPermalink

C0031Can a hash on the ledger identify someone?

The design uses salted commitments with per-record salts kept off the ledger, because plain hashes of low-entropy data can be guessed. Destroying a salt breaks linkability.

Designed, not yet builtPermalink

C0032Do registration staff see clinical charts?

No. In the pilot registrars deliberately have no chart access, only demographics and registration functions.

In the pilot (v0.5.0, synthetic data)Permalink

C0033Do IT administrators see patient data?

The design prevents it: administration is separated from data, database administrators cannot decrypt patient columns, and key release needs an attested policy decision.

Designed, not yet builtPermalink

C0034How is minimum necessary enforced?

Access is role AND relationship AND purpose AND context AND risk, evaluated server-side. Roles such as billing or quality see only the minimum needed for the work item.

Designed, not yet builtPermalink

C0035Are messages and alerts kept on site?

In part. The on-site hub runs messaging, Memo Desk and code alerts locally when the internet is down and replicates to the cloud when it returns. Messages are stored in the database in plaintext today. Carrying no patient data to pagers is the design, and there are no pagers in the pilot.

In the pilot (v0.5.0, synthetic data)Permalink

C0036Does keystroke analytics monitor staff typing?

It is off by default. If a customer enables it where law allows, it is a review flag only and never a sole lockout trigger, with workforce notice and privacy-officer sign-off.

Designed, not yet builtPermalink

C0037Is patient contact information exposed in alerts?

Alert payloads in the pilot do not carry patient identifiers to external channels, and there are no external channels (no pager or SMS). The design keeps pager and SMS text free of patient data.

In the pilot (v0.5.0, synthetic data)Permalink

C0038What does "closed loop" mean for a clinic?

No inbound connections from the internet to the clinic server and controlled outbound connections only. It is a network design term, not a claim that nothing can ever go wrong.

Designed, not yet builtPermalink

C0039How long is data retained?

Retention depends on country, state and record type and is set in a signed profile. The specification leaves the values open pending counsel. Audit defaults are proposals only.

Open decisionPermalink

C0040Can patients see who looked at their record?

Not yet. The design derives a patient-visible access history from the audit log. A patient portal now exists (appointments, released results, medications, allergies, messages and refill requests), but it does not show access history, so the pilot offers a per-patient access report to privacy staff only.

Designed, not yet builtPermalink
See the synthetic-data demo first.Request a demo