Summary
Add a reusable scenario/sensitivity report pattern for worked examples so tutorials can show how assumptions, classification provenance, LOS, ICU, and overlay choices affect results without implying authoritative real-world conclusions.
Linked Conductor track
conductor/tracks/public_clinical_dataset_worked_example_20260704
Why this is separate
The MIMIC-IV Demo worked example can include a bounded scenario appendix using synthetic overlays, but reusable scenario generation and reporting should be shared infrastructure for future examples.
Acceptance criteria
- Define a scenario manifest shape for baseline, synthetic overlay, local precomputed AR-DRG, and blocked/no-provenance cases.
- Generate comparison outputs that show changed NWAU values, changed assumptions, and support-state caveats.
- Include explicit warnings that synthetic-overlay scenarios are educational only.
- Add tests using tiny committed fixtures and existing acute weights.
Dependencies
- Public Clinical Dataset Worked Example.
- Provenance/data-quality report contract if implemented.
Summary
Add a reusable scenario/sensitivity report pattern for worked examples so tutorials can show how assumptions, classification provenance, LOS, ICU, and overlay choices affect results without implying authoritative real-world conclusions.
Linked Conductor track
conductor/tracks/public_clinical_dataset_worked_example_20260704Why this is separate
The MIMIC-IV Demo worked example can include a bounded scenario appendix using synthetic overlays, but reusable scenario generation and reporting should be shared infrastructure for future examples.
Acceptance criteria
Dependencies