A DocumentReference profile representing technical documentation about an AI system, such as intended use, limitations, risk-related information, performance-related information, and model documentation.

A Consent profile documenting patient-facing information, permission status, and opt-out preferences for AI-supported processing where applicable.

An AuditEvent profile documenting execution-related metadata of an AI-supported processing event to support retrospective reconstruction and auditability.

An Observation profile representing a clinical output generated by an AI system, including AI-related transparency metadata and links to the relevant patient, encounter, and AI system.

An ArtifactAssessment profile documenting professional review of an AI-generated output, including whether the result was accepted, corrected, modified, or overridden.

A Communication profile documenting patient-facing information about the AI-supported workflow, including the role of the AI system and the related clinical review where applicable.

A PractitionerRole profile representing the role, qualification context, specialty, and AI-related training information of the human reviewer involved in oversight of an AI-supported workflow.

A Provenance profile linking an AI-generated output to the contributing AI system, source data, and relevant processing or governance context.

An Organization profile representing an organization involved in manufacturing, providing, deploying, or operating an AI system, including relevant accountability and contact information.

A Device profile representing an AI system or software component, including system identification, versioning, intended purpose, and selected regulatory documentation metadata.

Documents whether the AI system is clinically validated, not clinically validated, under validation, or only technically validated.

Captures quantitative metrics and bias disclosures.

GDPR and AI Act privacy parameters. Third country transfer flags and data retention policies at the model level.

Mandatory flag indicating whether the human actor has received specific training for the utilized AI tool.

Details regarding provenance, EHDS categories, and data quality.

Indicates whether the AI-generated output was used as part of a solely automated decision-making process within the meaning of GDPR Article 22.

The clinical reason why the AI was used for this specific patient.

The unique ID of the Health Data Access Body permit (required if secondary use).

Documents the permitted purpose for secondary use of electronic health data under the EHDS.

Categorizes the data processing as Primary Care or Secondary Use according to the EHDS.

Flag indicating if the patient (data subject) explicitly requested a clear and meaningful explanation of the AI's role and the clinical decision.

Cryptographic signature or verification hash to ensure the integrity, accountability, and non-repudiation of the AI execution audit log.

A reference to the DocumentReference resource that acts as the Model Card, containing detailed documentation, intended purpose, and risk assessments.

Confirmation that the patient was informed about the use of AI systems according to AI Act transparency rules.

Captures if patient data is transferred outside the EU by this device.

Codes describing the clinical validation status of an AI system.

Codes describing data-quality characteristics relevant for training-data documentation (Art. 10 AI Act).

Codes for technical quality and performance metrics (QUAL-01).

Categories of electronic health data relevant for secondary-use documentation under the EHDS (EHDS Art. 51).

Permitted purposes for secondary use of electronic health data under the EHDS.

Codes defining whether data usage is documented as primary use or secondary use (LAW-03.1).

Clinical and administrative reasons for applying an AI system in a specific case (USE-04).

Value set including GDPR Article 6 legal bases relevant for documenting the lawful processing of personal data.

Value set including GDPR Article 9 exceptions relevant for documenting the processing of special categories of personal data, including health data.

Codes representing the type of human oversight or intervention (HL-03.1).

Code system defining permitted EHDS secondary-use purposes relevant for documenting secondary use of electronic health data.

Custom codes used in this IG to represent AI transparency, data-use, provenance, technical documentation, data-quality, and human-oversight documentation concepts.

Code system defining GDPR Article 6 legal bases relevant for documenting legal-basis metadata in AI-supported processing contexts.

Code system defining GDPR Article 9 exceptions relevant for documenting the processing context of special categories of personal data, including health data.

AI-generated early warning risk assessment output.

Human oversight assessment documenting validation of the AI output.

Audit event documenting the simulated AI execution.

Synthetic patient-facing explanation about AI-supported processing.

Consent and patient information context for AI-related processing.

Example AI system used for the NEWS2-inspired PoC scenario.

Synthetic encounter for suspected infection and early-warning risk assessment.

Synthetic NEWS2-inspired input parameter.

Synthetic NEWS2-inspired input parameter.

Synthetic NEWS2-inspired input parameter.

Synthetic NEWS2-inspired input parameter.

Synthetic NEWS2-inspired input parameter.

Synthetic NEWS2-inspired input parameter.

Manufacturer organization for the synthetic AI system.

Synthetic model card for the RiskAssist AI PoC system.

Healthcare organization operating the synthetic AI system.

A fictional female patient used in the NEWS2-inspired PoC scenario.

The fictional clinician responsible for reviewing the AI-generated output.

Human reviewer role used for AI output validation.

Provenance linking the AI-generated output to the AI system and input observations.