0.1.0 - ci-build
TrustAITransparencyIG, published by Selina Adlberger. This guide is not an authorized publication; it is the continuous build for version 0.1.0 built by the FHIR (HL7® FHIR® Standard) CI Build. This version is based on the current content of https://github.com/HL7Austria/AIST-TrustworthyAI-R5/ and changes regularly. See the Directory of published versions
The IG defines custom extensions where the FHIR core specification does not sufficiently represent AI transparency, legal context, or regulatory metadata. The Requirement column links to the logical attribute on the Requirements Traceability page that the extension implements.
| Extension | Used in | Value | Requirement |
|---|---|---|---|
| Model Card Reference | Trust_AIDevice | Reference to the Trust_AIModelCard | – |
| EU Conformity Declaration Reference | Trust_AIDevice | Reference to the EU Declaration of Conformity (DocumentReference) | SYS-03a |
| Third-Country Data Transfer | Trust_AIDevice | transferFlag (boolean), destinationCountry (ISO 3166 code) |
LAW-04.1, LAW-04.2 |
| Trust AI DPIA Reference | Trust_AIOrganization | Reference to the DPIA document (DocumentReference) | SYS-09 |
| AI Performance Metrics | Trust_AIModelCard | metric (type, value), biasDisclosure |
QUAL-01.1, QUAL-01.2, QUAL-04 |
| AI Training Data Metadata | Trust_AIModelCard | provenance, Category, SecondaryUsePurpose, Permit, dataQuality |
QUAL-02a, QUAL-02b, QUAL-02c, QUAL-03 |
| AI Retention Information | Trust_AIModelCard | retention (Duration) |
LAW-06 |
| AI Clinical Validation Status | Trust_AIModelCard | CodeableConcept | – |
| Extension | Used in | Value | Requirement |
|---|---|---|---|
| Usage Category | Trust_AIProvenance | CodeableConcept (primary or secondary use) | LAW-03.1 |
| Secondary Use Purpose | Trust_AIProvenance | CodeableConcept | LAW-03.2 |
| Data Permit | Trust_AIProvenance | Identifier | LAW-03.2 |
| Case-Specific Indication | Trust_AIProvenance | CodeableConcept | USE-04 |
| Automated Decision-Making Flag | Trust_AIProvenance | boolean | LAW-02 |
| Trust AI Log Integrity Signature | Trust_AIAuditEvent | Signature | LAW-08 |
| Extension | Used in | Value | Requirement |
|---|---|---|---|
| AI System-Specific Training Status | Trust_AIPractitionerRole | boolean | HL-02.2 |
| Patient AI Info Provided Flag | Trust_AIPatientExplanation | boolean | LAW-01c |