This publication is for STU1: Since the ballot publication, this version refines the scope and terminology and clarifies how AI involvement is identified and recorded in FHIR resources. It strengthens guidance for resource labeling and Provenance, updates the representation of AI systems and their model cards and prompts, adds a way to convey inference confidence, and expands implementation guidance and future considerations.
The jira specification file appears to be out of date with the versions, artifacts and pages currently defined in the IG. A proposed revised file to be reviewed and, if appropriate, submitted as a pull request against the XML folder in https://github.com/HL7/JIRA-Spec-Artifacts. To see the differences, perform a file compare on 'template/jira-current.xml' (a normalized view of what is in Github) and 'template/jira-new.xml' (reflects current IG content)
Hyperlink '#hcAI-full-lorem-ipsum/Input-Prompt-lorem-ipsum' at 'div/table/tr/td/a' for 'DocumentReference: status = current; type = AI Input Prompt; description = Generate a lorem ipsum text to serve as placeholder copy for use in design, development, and publishing. 1. Specify the exact amount of text or the number of paragraphs required (e.g., 1 paragraph, 3 paragraphs, etc.). 2. Create the lorem ipsum text using a classic style or introduce slight variations while keeping the nonsensical nature to suit the requested length. Ensure that the text maintains a good balance between readability and the traditional lorem ipsum style, giving a realistic impression of how the text will impact the overall layout and design. # Output Format - Provide a continuous block of lorem ipsum text corresponding to the specified amount needed.' does not resolveType_Specific_Checks_DT_XHTML_Resolve
WARNING: DocumentReference/ModelCard-sample-huggingface-attached: DocumentReference.identifier[0].system: No definition could be found for URL value 'https://github.com/huggingface/huggingface_hub/tree/main/tests/fixtures/cards' (1 uses)
03 some examples refer to the CHAI github
WARNING: DocumentReference/ModelCard-sample-CHAI-attached: DocumentReference.identifier[0].system: No definition could be found for URL value 'https://github.com/coalition-for-health-ai/mc-schema/blob/main/v0.1/examples/' (1 uses)
WARNING: DocumentReference/ModelCard-sample-CHAI-binary: DocumentReference.identifier[0].system: No definition could be found for URL value 'https://github.com/coalition-for-health-ai/mc-schema/blob/main/v0.1/examples/' (1 uses)
WARNING: DocumentReference/ModelCard-sample-CHAI-web: DocumentReference.identifier[0].system: No definition could be found for URL value 'https://github.com/coalition-for-health-ai/mc-schema/blob/main/v0.1/examples/' (1 uses)
06 Example that uses import to binary ising ig-loader that has no way to provide a definition
WARNING: DocumentReference/ModelCard-sample-CHAI-binary: DocumentReference.content[0].attachment.url: No definition could be found for URL value 'Binary/ModelCard-sample-CHAI-binary-xml' (1 uses)
WARNING: DocumentReference/ModelCard-sample-CHAI-binary: DocumentReference.content[1].attachment.url: No definition could be found for URL value 'Binary/ModelCard-sample-CHAI-binary-pdf' (1 uses)
08 These are made up codesystems for examples
WARNING: Observation/f206: Observation.code.coding[0].system: A definition for CodeSystem 'http://example.org/codes/foobar' could not be found, so the code cannot be validated (1 uses)
WARNING: Observation/glasgow: Observation.component[0].value.ofType(CodeableConcept).coding[0].system: A definition for CodeSystem 'http://example.org/codes' could not be found, so the code cannot be validated (1 uses)
WARNING: Observation/glasgow: Observation.component[1].value.ofType(CodeableConcept).coding[0].system: A definition for CodeSystem 'http://example.org/codes' could not be found, so the code cannot be validated (1 uses)
WARNING: Observation/glasgow: Observation.component[2].value.ofType(CodeableConcept).coding[0].system: A definition for CodeSystem 'http://example.org/codes' could not be found, so the code cannot be validated (1 uses)
09 known issue in an examples, where the examples are simply illustrative and only used as references
INFORMATION: DiagnosticReport/f202: DiagnosticReport.code: None of the codings provided are in the value set 'LOINC Diagnostic Report Codes' (http://hl7.org/fhir/ValueSet/report-codes|4.0.1), and a coding is recommended to come from this value set (codes = http://snomed.info/sct#104177005) (1 uses)
INFORMATION: Observation/glasgow: Observation.component[0].value.ofType(CodeableConcept).coding[1].extension[0]: The extension http://hl7.org/fhir/StructureDefinition/ordinalValue|5.3.0 is deprecated with the note: 'Replaced by [[[http://hl7.org/fhir/StructureDefinition/itemWeight]]]' (1 uses)
WARNING: Bundle/b3c1f2d4-5c8e-4b0a-9f6d-7c8e1f2d4b5c: Bundle.entry[1].resource/*Observation/608c4de1-1ab6-4bfe-b2e4-dca6b19223f0*/: Best Practice Recommendation: In general, all observations should have a performer (1 uses)
WARNING: Observation/glasgow: Observation.referenceRange[0].high: UCUM Codes that contain human readable annotations like {score} can be misleading (e.g. they are ignored when comparing units). Best Practice is not to depend on annotations in the UCUM code, so this usage should be checked, and the Quantity.unit SHOULD contain the annotation (1 uses)
WARNING: Observation/glasgow: Observation.referenceRange[1].high: UCUM Codes that contain human readable annotations like {score} can be misleading (e.g. they are ignored when comparing units). Best Practice is not to depend on annotations in the UCUM code, so this usage should be checked, and the Quantity.unit SHOULD contain the annotation (1 uses)
WARNING: Observation/glasgow: Observation.referenceRange[1].low: UCUM Codes that contain human readable annotations like {score} can be misleading (e.g. they are ignored when comparing units). Best Practice is not to depend on annotations in the UCUM code, so this usage should be checked, and the Quantity.unit SHOULD contain the annotation (1 uses)
WARNING: Observation/glasgow: Observation.referenceRange[2].low: UCUM Codes that contain human readable annotations like {score} can be misleading (e.g. they are ignored when comparing units). Best Practice is not to depend on annotations in the UCUM code, so this usage should be checked, and the Quantity.unit SHOULD contain the annotation (1 uses)
WARNING: Observation/glasgow: Observation.value.ofType(Quantity): UCUM Codes that contain human readable annotations like {score} can be misleading (e.g. they are ignored when comparing units). Best Practice is not to depend on annotations in the UCUM code, so this usage should be checked, and the Quantity.unit SHOULD contain the annotation (1 uses)
WARNING: Observation/other-model-result: Observation: Best Practice Recommendation: In general, all observations should have a performer (1 uses)
10 inherited from FHIR core
The extension http://hl7.org/fhir/StructureDefinition/elementdefinition-maxValueSet|5.3.0 is deprecated with the note: 'Use additionalBinding extension or element instead' (1 uses)
11 Pinning versions when there are multiples
PIN_VERSION (6 uses)
13 Examples with extra elements beyond minimal slices
INFORMATION: DocumentReference/ModelCard-sample-CHAI-attached: DocumentReference.content[1]: This element does not match any known slice defined in the profile http://hl7.org/fhir/uv/aitransparency/StructureDefinition/AI-ModelCard% (1 uses)
INFORMATION: DocumentReference/ModelCard-sample-CHAI-binary: DocumentReference.content[1]: This element does not match any known slice defined in the profile http://hl7.org/fhir/uv/aitransparency/StructureDefinition/AI-ModelCard% (1 uses)
INFORMATION: DocumentReference/ModelCard-sample-CHAI-web: DocumentReference.content[1]: This element does not match any known slice defined in the profile http://hl7.org/fhir/uv/aitransparency/StructureDefinition/AI-ModelCard% (1 uses)
INFORMATION: Provenance/AI-full-lorem-ipsum: Provenance.reason[0]: This element does not match any known slice defined in the profile http://hl7.org/fhir/uv/aitransparency/StructureDefinition/AI-Provenance% (1 uses)
INFORMATION: Provenance/AI-generated-patient-resource: Provenance.reason[0]: This element does not match any known slice defined in the profile http://hl7.org/fhir/uv/aitransparency/StructureDefinition/AI-Provenance% (1 uses)
14 We should do this, but after public comments. Some of the codes will go into existing codeSystems in THO and thus not new codeSystems.
INFORMATION: CodeSystem/AIdeviceTypeCS: CodeSystem: Most code systems defined in HL7 IGs will need to move to THO later during the process. Consider giving this code system a THO URL now (See https://confluence.hl7.org/display/TSMG/Terminology+Play+Book, and/or talk to TSMG) (1 uses)
INFORMATION: CodeSystem/AIinputsCS: CodeSystem: Most code systems defined in HL7 IGs will need to move to THO later during the process. Consider giving this code system a THO URL now (See https://confluence.hl7.org/display/TSMG/Terminology+Play+Book, and/or talk to TSMG) (1 uses)
15 We are not using the C-80 codes for DocumentReference.type
Terminology_TX_NoValid_3_CC (7 uses)
16 The AIconfidence extension is intended to be used anywhere
WARNING: StructureDefinition/AIconfidence: StructureDefinition.context[1]: Review the extension type for AIconfidence: extensions should not have a context of Element unless it's really intended that they can be used anywhere (1 uses)
17 the use of requirements formally will happen in future releases, but is not intended for now
The html source contains the word 'SHALL' but it is not in a text phrase marked as a conformance clause:% (1 uses)
18 examples need to have Provenance.entities that are representing example ANY other stuff
INFORMATION: Provenance/AI-Authored-Element: Provenance.agent[0]: This element does not match any known slice defined in the profile http://hl7.org/fhir/uv/aitransparency/StructureDefinition/AI-Provenance|1.0.0 (this may not be a problem, but you should check that it's not intended to match a slice) (1 uses)
INFORMATION: Provenance/AI-Contributed: Provenance.agent[0]: This element does not match any known slice defined in the profile http://hl7.org/fhir/uv/aitransparency/StructureDefinition/AI-Provenance|1.0.0 (this may not be a problem, but you should check that it's not intended to match a slice) (1 uses)
INFORMATION: Provenance/AI-Generated-Lab-Results: Provenance.entity[0]: This element does not match any known slice defined in the profile http://hl7.org/fhir/uv/aitransparency/StructureDefinition/AI-Provenance|1.0.0 (this may not be a problem, but you should check that it's not intended to match a slice) (1 uses)
INFORMATION: Provenance/AI-PriorAuth-Provenance: Provenance.agent[0]: This element does not match any known slice defined in the profile http://hl7.org/fhir/uv/aitransparency/StructureDefinition/AI-Provenance|1.0.0 (this may not be a problem, but you should check that it's not intended to match a slice) (1 uses)
INFORMATION: Provenance/AI-PriorAuth-Provenance: Provenance.entity[0]: This element does not match any known slice defined in the profile http://hl7.org/fhir/uv/aitransparency/StructureDefinition/AI-Provenance|1.0.0 (this may not be a problem, but you should check that it's not intended to match a slice) (1 uses)
INFORMATION: Provenance/AI-full-lorem-ipsum: Provenance.agent[0]: This element does not match any known slice defined in the profile http://hl7.org/fhir/uv/aitransparency/StructureDefinition/AI-Provenance|1.0.0 (this may not be a problem, but you should check that it's not intended to match a slice) (1 uses)
INFORMATION: Provenance/AI-full-lorem-ipsum: Provenance.entity[0]: This element does not match any known slice defined in the profile http://hl7.org/fhir/uv/aitransparency/StructureDefinition/AI-Provenance|1.0.0 (this may not be a problem, but you should check that it's not intended to match a slice) (1 uses)
INFORMATION: Provenance/AI-full-lorem-ipsum: Provenance.entity[1]: This element does not match any known slice defined in the profile http://hl7.org/fhir/uv/aitransparency/StructureDefinition/AI-Provenance|1.0.0 (this may not be a problem, but you should check that it's not intended to match a slice) (1 uses)
INFORMATION: Provenance/AI-full-lorem-ipsum: Provenance.entity[2]: This element does not match any known slice defined in the profile http://hl7.org/fhir/uv/aitransparency/StructureDefinition/AI-Provenance|1.0.0 (this may not be a problem, but you should check that it's not intended to match a slice) (1 uses)
INFORMATION: Provenance/AI-generated-patient-resource: Provenance.agent[0]: This element does not match any known slice defined in the profile http://hl7.org/fhir/uv/aitransparency/StructureDefinition/AI-Provenance|1.0.0 (this may not be a problem, but you should check that it's not intended to match a slice) (1 uses)
19 this is a tooling problem with contained resources in Provenance. Solution has been proposed, but tooling team want to solve this the same for all resources https://github.com/hapifhir/org.hl7.fhir.core/pull/2702
Hyperlink '#hcAI-full-lorem-ipsum/Input-Prompt-lorem-ipsum' at 'div/table/tr/td/a' for 'DocumentReference: status = current; type = AI Input Prompt; description = Generate a lorem ipsum text to serve as placeholder copy for use in design, development, and publishing. 1. Specify the exact amount of text or the number of paragraphs required (e.g., 1 paragraph, 3 paragraphs, etc.). 2. Create the lorem ipsum text using a classic style or introduce slight variations while keeping the nonsensical nature to suit the requested length. Ensure that the text maintains a good balance between readability and the traditional lorem ipsum style, giving a realistic impression of how the text will impact the overall layout and design. # Output Format - Provide a continuous block of lorem ipsum text corresponding to the specified amount needed.' does not resolve (0 uses)
The link '#hcAI-full-lorem-ipsum/Input-Prompt-lorem-ipsum' for "DocumentReference: status = current; type = AI Input Prompt; description = Generate a lorem ipsum text to serve as placeholder copy for use in design, development, and publishing. 1. Specify the exact amount of text or the number of paragraphs required (e.g., 1 paragraph, 3 paragraphs, etc.). 2. Create the lorem ipsum text using a classic style or introduce slight variations while keeping the nonsensical nature to suit the requested length. Ensure that the text maintains a good balance between readability and the traditional lorem ipsum style, giving a realistic impression of how the text will impact the overall layout and design. # Output Format - Provide a continuous block of lorem ipsum text corresponding to the specified amount needed." cannot be resolved (valid targets: [hcAI-full-lorem-ipsum, segment-header, example-provenance-provenance-for-ai-with-model-card-input-and-output, publish-box, bottom, hl7-logo, segment-post-footer, hl7-search, hl7-search-lnk, family-logo, family-nav, ig-status, top, AI-full-lorem-ipsum, root, stripe, segment-footer, hl7-nav, segment-navbar, segment-content, segment-breadcrumb]) (1 uses)
Hyperlink '#hcAI-full-lorem-ipsum/Input-Prompt-lorem-ipsum' at 'div/table/tr/td/a' for 'DocumentReference: status = current; type = AI Input Prompt; description = Generate a lorem ipsum text to serve as placeholder copy for use in design, development, and publishing. 1. Specify the exact amount of text or the number of paragraphs required (e.g., 1 paragraph, 3 paragraphs, etc.). 2. Create the lorem ipsum text using a classic style or introduce slight variations while keeping the nonsensical nature to suit the requested length. Ensure that the text maintains a good balance between readability and the traditional lorem ipsum style, giving a realistic impression of how the text will impact the overall layout and design. # Output Format - Provide a continuous block of lorem ipsum text corresponding to the specified amount needed.' does not resolve
hl7.fhir.uv.aitransparency
hl7.terminology.r4
