This page is part of the MII Implementation Guide PRO (v2027.0.0-ballot.1: 2027 Ballot 5) based on FHIR (HL7® FHIR® Standard) R4. This is the current published version in its permanent home (it will always be available at this URL). For a full list of available versions, see the Directory of published versions

Changelog

This page documents the changes between versions of the MII PRO module.

2027.0.0-ballot.1 — manifest hotfix

Content-identical to 2027.0.0-ballot. Fixes the package manifest: the npm-style dependencies of the distributed package still declared the pre-upgrade pins (meta 2026.0.x, SDC 3.0.0, ISiK 5.1.1), so fhir install resolved outdated transitive packages. The manifest now mirrors sushi-config.yaml (meta 2027.0.0-ballot final, SDC 4.0.0, ISiK 6.0.0), the meta pin was lifted from ballot.rc3 to the ballot final, and manifest/config drift is now a hard release check. The formally published package.tgz on this site was always correct — the defect was limited to the package registry channel.

2027.0.0-ballot

The ballot version of the 2027 line. Relative to rc5, five QA error classes were closed (294 → ~205: QuestionnaireResponse texts aligned with their questionnaire definitions, CapabilityStatement search-parameter canonicals corrected to the R4 conformance-* parameters, catalogue displays synchronised, ConceptMap linkId maps moved to logical URIs, unresolvable web references converted to display-only) and the PROMIS-29 example item order was fixed.

Known Issues

The remaining QA findings are documented environment and upstream classes, not defects of the module:

• ~171 LOINC answer displays — the module deliberately renders the instrument-validated German wordings (PHQ-D, Löwe et al.; PROMIS German, PHO/CPCOR), which differ from LOINC's en-US displays. The shipped LOINC supplement documents them machine-readably; measured on both relevant publishers (2.2.11, 2.3.2), display validation does not yet consult local supplements. The class disappears once the terminology server loads the supplement.
• 23 supplement concept checks — the validator cannot confirm the LA codes against LOINC's content and flags every supplement concept. The codes verifiably exist ($lookup against tx.fhir.org, 2026-09-06).
• 4 inherited SDC links — a dead link inside element documentation inherited from the SDC parent profile (2025Jan/rendering.html); an upstream issue of the SDC package.
• ~10 preview-only download links — package.tgz and the *.zip downloads resolve in the formal publication (measured HTTP 200) but not in branch previews.
• 1 dependency FHIR-version notice — subscriptions-backport.r4 declares FHIR 4.0.0 and arrives transitively; not pinnable from this module.

Open ballot question: a hard obligation for item.text and answer.display in QuestionnaireResponse (1..1 plus German-availability invariants) was implemented and measured, but conflicts with licence-restricted (metadata-only) instruments and paper-based capture. The profile currently ships the must-support translation pattern without the obligation — commenters are invited to weigh in.

2027.0.0-ballot.rc5

QuestionnaireResponse: German via translation extension (2026-09-10): item.text and item.answer.valueCoding.display are now must-support with an explicit translation-extension slice, and all 20 example responses demonstrate the pattern end-to-end (350 item texts, 252 answer displays; German-primary instruments carry German directly, English-primary ones the canonical wording plus German translation extensions). The EORTC QLQ-C30 questionnaire gained the official German item wordings along the way. A hard obligation (1..1 plus German-availability invariants) was implemented and measured, but is deferred to ballot commentary: it conflicts with licence-restricted (metadata-only) instruments and paper-based capture workflows.

LOINC supplement (2026-09-08): mii-cs-pro-loinc-supplement ships the validated German answer wordings (PHQ-D, PROMIS/PHO) as designations on 24 LOINC codes — content for consumers and terminology servers; measured on both relevant publishers (2.2.11, 2.3.2), display validation does not consume local supplements yet, so the ~178 display findings remain a documented environment class.

The 2027 ballot version brings the move onto the MII KDS module template (v0.11.3) — the same toolchain as the other KDS modules. For implementers nothing changes in the FHIR artifacts: all 214 resources, their ids, canonicals and versions are unchanged (measured, 0 regressions).

What does change:

• One page set instead of two. The IG-Publisher website and the Simplifier guide are produced from the same sources; the drift between the two documentation trees (e.g. the scoring page: 2025-08 vs. 2026-08 state) is structurally over.
• Bilingual with an English default. English is the default language, German the complete translation — 61 pages per language, switchable on every page. Twelve English pages were newly created from the German originals for this (review pending).
• TF-KDS menu structure with artifact intro notes: the descriptions of the abstract profiles and extensions now sit directly on the generated artifact pages.
• Formal publication prepared: the template's release workflows (tag → build → draft release → Zulip), a pinned toolchain instead of :latest, GitHub Pages deployment via Actions.
• Jurisdiction switched from m49#276 to urn:iso:std:iso:3166#DE (same meaning; the m49 encoding rendered "Unknown region code" in every page header).
• Two small data fixes: the WHODAS example QuestionnaireResponse declares language = #de; the PROMIS-29 bundle no longer references the person module's patient profile with the PRO module version.

EORTC QLQ-C30 restructured to standalone terminology (2026-09-08): the main questionnaire no longer contains its CodeSystem/ValueSets — it binds the standalone mii-vs-pro-eortc-qlq-c30-scale-4pt/-7pt, whose CodeSystem now carries the answer weights as itemWeight properties (ported 1:1 from the retired contained resources; EORTC scales are 1-based). The two modelling-experiment variants (…-variant-a, …-variant-b) are removed — they were referenced nowhere. Resolves the largest remaining QA class (82 contained-resource profile violations).

Dependency upgrade for the 2027 line (2026-09-08): SDC 3.0.0 → 4.0.0 (final), ISiK 5.1.1 → 6.0.0, kerndatensatz.meta 2026.0.0 → 2027.0.0-ballot.rc3, basisprofil 1.5.4 → 1.6.0. The SDC step is the breaking one and was migrated in full: the ordinalValue extension becomes itemWeight (395 answer options), CodeSystem weight properties become concept-properties#itemWeight, and all 90 FHIRPath scoring expressions switch from .ordinal() to .weight(). Existing QuestionnaireResponses are unaffected — weights live on the questionnaire and terminology side, not in the responses.

Details and open decisions: PR #129.

2026.7.0 — 2026-09-02 (released, tag v2026.7.0, package + GitHub release)

Minor release — GAD-7 as a new instrument, a shared PHQ four-point scale and a ConceptMap for item numbering. No breaking changes to existing resources.

Added: GAD-7 (Generalized Anxiety Disorder Scale-7)

Seven items screening for generalized anxiety disorder, sum score 0–21, severity bands 5/10/15 (Spitzer et al. 2006). English as the primary language with the validated German translation (PHQ-D; Löwe et al. 2008, n=5030). LOINC panel 69737-5, score 70274-6. Freely available, no permission required.

linkIds in the PHQ-D block namespace (phq-phq5a…phq-phq5g) rather than an instrument-specific numbering. The reason is the interlocking within the PHQ family: the first two GAD-7 items form the GAD-2, and together with two PHQ-9 items they make up the PHQ-4. The same question should carry the same linkId across instruments — otherwise a migration ConceptMap would become necessary later, as it had to be created for the PHQ-9 in 2026.5.0.

Not carried over: a PROMIS anxiety T-score conversion that existed as a FHIRPath expression in an earlier development state. Score conversions between instruments will be modelled as a CQL library.

Added: shared PHQ four-point scale

The RuleSet Phq4PointFrequencyAnswerOptions bundles the two-week frequency scale (0–3) with the LOINC answer codes of list LL358-3 (LA6568-5…LA6571-9) and the German PHQ-D wordings. GAD-7 uses it; PHQ-9 and PHQ-4 can be switched to it.

Added: ConceptMap for GAD-7 item numbering

mii-cm-pro-gad-7-linkids maps the conventional GAD-7 numbering (items 1–7, as published and common in third-party systems) onto the canonical linkIds. Explicitly not a migration of a published version — a reading aid for taking over data from third-party systems.

2026.6.0 — 2026-09-01 (released, tag v2026.6.0, package + GitHub release)

Minor release — six new instruments from the PCOR-MII PSS set (Persistent Somatic Syndrome), a new shared answer scale and a sharpened display-language convention. No breaking changes to existing resources.

New instruments

All item texts come from the PCOR-MII Item Level Dictionary and were verified character-exactly against that source. All example scores are recalculated from the ordinalValue weights.

Added: shared yes/no answer scale

New RuleSet YesNoAnswerOptions (input/fsh/rulesets/answer-scales.fsh) for generic yes/no answers. It uses SNOMED CT 373066001 (Yes) / 373067005 (No) per the recommendation of the TC Terminologien (HL7 Germany / Interop Council) — instead of a self-defined MII CodeSystem duplicating standardised terminology.

Since no ordinalValue property can be attached to SNOMED concepts, these instruments use inline answerOption rather than answerValueSet. Answers in QuestionnaireResponses thereby carry a terminologically unambiguous coding with system and code instead of systemless integers.

Sharpened: display language in answer CodeSystems

The display value of an answer concept is in the language of the questionnaire; further languages are carried as designation. This is not a matter of style: in a resource with language = #de the FHIR validator expects the German display in valueCoding. With the English one there, the entire answerValueSet check fails — with the misleading follow-up message that the value is not in the options value set, suggesting a terminology problem that does not exist.

Modelling notes

• ISR-Z forms a mean, not a sum. The ISR defines scale values as means of the items (Tritt et al. 2008); the FHIRPath calculation accordingly uses .avg() instead of .sum(), range 0-4.
• Language per instrument. language follows the instrument's original language: SCOFF, WI-7 and EURONET-SOMA are EN-primary with a German translation, PC-PTSD, SSD-12 and ISR-Z DE-primary. Where no validated original could be evidenced, no translation was invented.
• Cut-offs stay documentation. Published cut-offs (SCOFF ≥ 2, PC-PTSD ≥ 3) are documented as qualifiedInterval in the ObservationDefinitions but not shipped as executable interpretation logic.
• Terminology. For none of the six instruments do LOINC or SNOMED CT carry a code for the instrument itself (checked against LOINC 2.83 and SNOMED International 2026-05-01); coding is via the MII questionnaire catalogue.

2026.5.2 — 2026-07-27 (released, tag v2026.5.2, package)

Patch release (fix-only) — no changes to instruments, profiles or terminology. Recommended for everyone running 2026.4.0–2026.5.1.

Bugfix: import fails in HAPI FHIR (server does not start)

The packages 2026.4.0–2026.5.1 could not be loaded into HAPI FHIR — the server aborted at startup:

HAPI-0838: ConceptMap[url='…/ConceptMap/mii-cm-pro-bdi-ii-to-promis-depression-observation']
           contains at least one group without a value in ConceptMap.group.source
→ HAPI-1286: Error installing IG … → Application run failed → Unable to start web server

• Cause: the ConceptMap mii-cm-pro-bdi-ii-to-promis-depression-observation introduced in 2026.4.0 was an incomplete stub (only 2 of 64 PROsetta Stone anchor points, group without source/target CodeSystem). HAPI rejects ConceptMaps without group.source on package install and aborts the server start. 2026.3.0 contained no ConceptMap and therefore loaded cleanly.
• Fix: the ConceptMap stub was removed from source and package. The BDI-II→PROMIS depression T-score conversion remains as a complete FHIRPath crosswalk in the BDI-II questionnaire; a CQL library solution is planned for 2027.
• Verified: 2026.5.2 was tested in HAPI FHIR (v7.6.0) and starts cleanly; the remaining ConceptMap mii-cm-pro-phq-9-linkid-migration installs correctly (it carries sourceScope/targetScope).

Package hygiene (fixed alongside)

• Removed: three accidentally packaged IG-Publisher lifecycle artifacts (onLoad-ig-working.json, onLoad-ig-updated.json, onGenerate-ig-working.json — a triple copy of the ImplementationGuide with an identical canonical URL) are now filtered out in the Firely Bake build (onLoad-*/onGenerate-*/onCheck-*), as are macOS files (._*, .DS_Store) defensively. Not a HAPI startup blocker, but unclean packaging.
• Fixed: uniform versioning — all conformance resources consistently carry 2026.5.2. Previously shipped packages contained mixed version states (down to 2026.0.0) because fsh-generated was not regenerated on every version bump.

2026.5.1 — 2026-07-13 (released, tag v2026.5.1, GitHub release + package)

Patch release on 2026.5.0.

Migration & breaking changes

• BREAKING (carried over from 2026.5.0): the PHQ-9 item linkIds were switched in 2026.5.0 from the phq-phq9-q01…q10 scheme to the shared PHQ-D block namespace (phq-phq2a…i + phq-phq9-difficulty). Existing PHQ-9 QuestionnaireResponses with old linkIds no longer match the questionnaire from 2026.5.0 on.
• Added: ConceptMap mii-cm-pro-phq-9-linkid-migration — maps the old PHQ-9 linkIds 1:1 onto the new ones (10 items: phq-phq9-q01…q09 → phq-phq2a…i, phq-phq9-q10 → phq-phq9-difficulty) for migrating existing responses. The calculated items (phq-phq9-score-total, phq-phq9-promis-tscore) were not renamed.

2026.5.0 — 2026-07-13 (released, tag v2026.5.0, GitHub release + package)

Minor release: two new instruments (WHODAS 2.0 12-item and PHQ-15), the shared PHQ-D item bank (PHQ-9/PHQ-15), uniform MII score coding across all score ObsDefs and a CI improvement (ValueSet expansion) that permanently fixes answerValueSet answer validation.

New instruments

• Added: WHODAS 2.0 12-item (self-report) — WHO Disability Assessment Schedule 2.0, 12 items across 6 ICF domains, 30-day recall, 5-point scale (0–4). English primary with German translations (validated PCOR-MII wordings). Answers via answerValueSet (mii-vs-pro-whodas-12-answer-list) with ordinalValue weights on the CodeSystem concepts. Limitation score mii-obsdef-pro-score-whodas12-simple-sum (0–48, SNOMED 715823002, MII catalogue whodas12-simple-sum, higher = more impairment). Incl. example QuestionnaireResponse + score Observation and IG page.
  • Licence: WHODAS 2.0 © WHO 2010. Terms are machine-readable as copyright on Questionnaire and CodeSystem: free clinician self-use; electronic/data-capture use requires a WHO usage agreement (free for non-commercial use); translations require WHO approval; MII FHIR content CC0, WHODAS item text © WHO.
• Added: PHQ-15 (somatic symptom burden, PHQ-D) — 15 items, 4-week recall, 3-point scale (0–2). English primary with German translations (PHQ-D, Löwe et al. 2002). Answers via answerValueSet with ordinalValue weights. Score mii-obsdef-pro-score-phq-15 (0–30, LOINC 70273-8) incl. severity categories (Kroenke et al. 2002: 0–4 / 5–9 / 10–14 / 15–30) as qualifiedInterval reference ranges. Incl. IG page. Licence: freely available (public domain), PHQ-D.

Architecture: PHQ-D item bank

• Changed: PHQ-9 and PHQ-15 use a shared PHQ-D block linkId scheme (phq-phq…). Items contained in both instruments (sleep phq-phq2c, fatigue phq-phq2d) share the same linkId — the basis for item-based scoring across instruments.
• Changed: PHQ-15 switched to answerValueSet + CodeSystem ordinalValue (EN-first); copyright/licence status set explicitly on PHQ resources.

Scoring & terminology

• Changed: MII catalogue code (code.coding[mii]) added to all score ObservationDefinitions (PROMIS-29 ×8, PROMIS Cognitive Function SF4a, depression T-score, BDI-II, PHQ-9, PHQ-15) — alongside LOINC/SNOMED, uniformly queryable via the MII score catalogue (EQ-5D/EORTC/DASS-21/PRO-CTCAE already had it).

Technical improvements

• Changed: CI — expand-valuesets.js now runs after SUSHI in the IG build and writes vs.expansion from local CodeSystems. The validator can thereby resolve the MII-controlled answerValueSet answer scales; the previously reported "value not in the specified options" findings on example QuestionnaireResponses disappear (PHQ-15, WHODAS and future answerValueSet instruments).

2026.4.1 — 2026-06-15 (released, tag v2026.4.1)

Patch release on v2026.4.0: fixes missing version metadata on the 3 EQ-5D-5L ObservationDefinitions, discovered while baking v2026.4.0.

FHIR resources:

• Fixed: mii-obsdef-pro-score-eq5d5l-index, …-profile, …-vas — * insert ObsDefVersion and * insert MetaProfile(...) added (previously both RuleSets were missing, so no artifact-version extension in the package and no meta.profile entry with a versioned canonical). All other ObsDefs already had both.

Verification:

• Before: 36 ObsDefs with artifact-version=2026.4.0, 3 without (EQ-5D)
• After: 39 ObsDefs with artifact-version=2026.4.1

2026.4.0 — 2026-06-14 (released, tag v2026.4.0)

Two thematic additions in one release: PROMIS consolidation (PROMIS-16 PROPr, wording migration, copyright model) and symptom screening for oncological and palliative care (MIDOS2 + PRO-CTCAE oncology base screening).

Symptom screening (MIDOS2 + PRO-CTCAE oncology base screening)

PRO instruments:

• Added: MIDOS2 (Minimal Documentation System for Patients in Palliative Care, DGP) — palliative symptom screening with 13 items (11 symptoms on the DGP 4-point severity scale, well-being on a 4-point scale, free text)
• Added: PRO-CTCAE oncology base screening (DKG) — subset of the NCI PRO-CTCAE with a MIDOS2-equivalent symptom selection, 10 AEs / 23 items, depressed mood via PRO-CTCAE-conform multi-mapping (#55 Discouraged + #56 Sad), anxiety+tension via #54 Anxiety

Architecture:

• Added: PRO-CTCAE Complete as master item bank, derivatives via derivedFrom
• Added: catalogue entries midos-midos2 and proctcae-onkologisches-basisscreening

Terminology & mapping:

• Added: MIDOS2 CodeSystem with the DGP 4-point severity scale; separate well-being scale
• Added: SNOMED properties on MIDOS2 item codes (where unambiguous; Snowstorm verification pending)
• Added: MIDOS2 ↔ PRO-CTCAE item-level mapping (11 AEs, 23 items)
• Verified: PRO-CTCAE item wordings against the NCI original (EN/DE); discrepancies #27 Hair Loss + #59 Vaginal Discharge (int → amt) documented

IG documentation & examples:

• Added: midos2.md and proms-onkologisches-basisscreening.md pages, registered under the PRO Library
• Added: example responses for both instruments (MIDOS2 sum 19/33; oncology patient under chemotherapy)
• Status: both instruments draft + experimental; MIDOS2 source verification against Stiel et al. 2010/2012 pending (bead 5jd)

PROMIS consolidation

PRO instruments:

• Added: PROMIS-16 Profile v2.1 (PROPr) — ultra-short multi-domain instrument, 16 items across 8 domains (incl. Cognitive Function); all items with the official German wordings from the PHO PDF "PROMIS-16 Profile v2.1 (PROPr), German, 20 September 2024". Score calculation (PROPr utility + 8 domain T-scores) deliberately deferred (CQL library mii-lib-promis-16 planned)
• Fixed: PROMIS-29 wording migration — 25 item texts aligned with the official German wordings (previously AI-generated/LOINC-fragment translations). Source hierarchy: the PHO PDF "PROMIS-29 Profile v2.1 German, 06 April 2020" is normative; the PCOR-MII Master Item-Level Dictionary is the secondary cross-check
• Fixed: PROMIS-29 pain-interference verb tense — 4 items switched to past tense per the PHO PDF; answer scales — 5 items (HI7, AN3, Sleep116, Sleep20, Sleep44) moved from the frequency to the intensity scale with the corresponding LOINC code swap; SRPPER23-CaPS — wrong LOINC 62041-9 replaced by 76709-5 with matching German text
• Changed: PROMIS Cognitive Function SF 4a — copyright RuleSet added (wordings were already PHO-conform)

Architecture & licensing:

• Added: central RuleSet PROMIS_Copyright_DE with layered attribution on all PROMIS questionnaires (MII FHIR structure / PROMIS items © PHO / German translations via PCOR-MII, curated by CPCOR Charité / LOINC © Regenstrief)
• Added: publisher, copyright, useContext and CPCOR contact on all PROMIS questionnaire instances; new ValueSet MII_VS_PRO_PROMIS_Intensity_Response_Scale (LOINC LL1024-0)
• Clarified: mii-qst-pro-promis-29-de is a presentation variant only (DE-primary for renderers that ignore the translation extension); the EN-primary main version is authoritative

Tooling & verification:

• Added: skill extract-qst-translations — extracts item texts and translation extensions from generated questionnaires and diffs them against the PCOR-MII master (MATCH/DIFF counters)
• Verified: PROMIS-29 against the PHO PDF word-identical after the fixes; PROMIS-16 MATCH=15, DIFF=0; Cognitive Function SF 4a MATCH=4, DIFF=0

2026.3.0 — 2026-05-08 (released, tag v2026.3.0)

PRO instruments:

• Added: GAD-7 — first implementation with questionnaire, terminology and scoring (#80); superseded by the 2026.7.0 re-implementation in the PHQ-D namespace

FHIR resources:

• Fixed: PROMIS-29 English variant repaired — 33 defective item texts corrected
• Removed: incomplete PROMIS-29 minimal variant
• Removed: incomplete ConceptMap stub mii-cm-pro-bdi-ii-to-promis-depression-observation (only 2 of 64 PROsetta Stone entries). Architectural decision: score conversions will be modelled as CQL libraries (planned for the 2027 release)
• Fixed: capabilities extension completed on mii-qst-pro-promis-cognitive-function-sf4a; added to mii-qst-pro-pro-ctcae-breast-de (displayable+collectable=true; calculatable/extractable/domainAligned=false — deliberate, pending composite-grade scoring decisions)

2026.2.0 — 2026-03-30 (released, tag v2026.2.0, Simplifier package published)

PRO instruments:

• Added: DASS-21 (Depression Anxiety Stress Scales) — full implementation with questionnaire, terminology (CodeSystem, ValueSets), scoring and examples; IG documentation page

Dependencies and tooling:

• Changed: de.gematik.isik 5.0.0 → 5.1.1; hl7.fhir.uv.extensions.r4 5.1.0 → 5.2.0 (consistent with the Imaging and Laboratory modules)

ObservationDefinition:

• Fixed: depression T-score ObsDef — population information moved from qualifiedInterval.context.coding to qualifiedInterval.appliesTo

EORTC QLQ-C30 ValueSet architecture:

• Changed: ValueSets converted from contained to standalone external ValueSets (variant A); post-SUSHI expansion script for LHC Forms rendering; display texts added

Quality assurance:

• Fixed: ObsDef titles deduplicated; broken {{render:infobox}} syntax replaced with Markdown blockquotes; CodeSystem/ValueSet switched to short ids

Note: the DASS-21 and QA items were originally planned as a standalone v2026.1.0 release but were never published as their own package; they first shipped as part of v2026.2.0.

2026.0.1 — 2026-01-26

Bugfix release with PROMIS-29 additions and canonical URL correction.

FHIR resources:

• Added: PROMIS-29 German variant and comprehensive examples
• Fixed: PROMIS-29 QuestionnaireResponse validation errors; canonical URLs standardised to https; packageId corrected (kerndatensatz.pro → kerndatensatz.pros)

IG documentation:

• Changed: PROMIS IG pages reorganised, tree rendering fixed; individual pages for the abstract profiles; SDC STU3 spec links added

2026.0.0 — 2026-01-12

First stable release of the MII PRO module for productive use.

FHIR resources and profiles:

• Added: SearchParameters for Questionnaire, QuestionnaireResponse, Observation; CapabilityStatement documenting the supported FHIR operations
• Fixed: meta.profile added to all Observation example instances

Version management and metadata:

• Added: MetaProfile RuleSet for consistent profile conformance with versioned canonicals ({canonical}|{version}); ObsDefVersion RuleSet (R5 backport of artifact-version for R4); version RuleSets centralised in input/fsh/rulesets/version.fsh

Dependencies and tooling:

• Changed: de.medizininformatikinitiative.kerndatensatz.meta updated to 2026.0.0; package.json name corrected to de.medizininformatikinitiative.kerndatensatz.pros; package.bake.yaml added for the Firely Bake packaging workflow

Quality assurance and developer experience:

• Added: EORTC QLQ-C30 validation suppressions and QA documentation; questionnaire and score catalogue maintenance; CI with inline Java validation and Gradle caching; Claude Code build/release/IG-export skills

2026.0.0-ballot

Initial ballot version focused on SDC-based PRO capture and score calculation.

PRO instruments: PHQ-9 (raw + T-score), EQ-5D-5L (index, VAS, profile scores), PROMIS-29 Profile v2.1 (7 domains with automatic scoring), PROMIS Cognitive Function SF4a, BDI-II (reference implementation), EORTC QLQ-C30 (15 subscales).

FHIR profiles and extensions: MII_PR_PRO_Questionnaire (SDC-based, calculated expressions, conditional display), MII_PR_PRO_QuestionnaireResponse (SDC extraction), MII_PR_PRO_Score_Blueprint and MII_PR_PRO_Score_Instance, the questionnaire capabilities extension.

Terminology strategy: MII-controlled CodeSystems and ValueSets with ordinalValue extensions for reliable score calculation; German translations via translation extensions; FHIRPath .ordinal() for weight-based calculations.

Technical basis: FHIR R4 (4.0.1) with SDC 3.0.0; de.medizininformatikinitiative.kerndatensatz.meta 2026.0.0; de.gematik.isik 5.0.0.