Synthetic teaching case
Case 4: decision-critical source missing
Synthetic example showing why missing tomography is not interpreted as normal.
Not real patient dataThis worked example was created to teach the current CER-AI v2.0 rules. It is not a clinical decision.
Source inputs
| Field | Value |
|---|---|
| Procedure | PRK |
| B.Ele.Th | Unreadable / unavailable |
| Other measurements | Available but insufficient for complete NICE and PS3 |
Pathway evaluation
| System | Calculation / finding | Result |
|---|---|---|
| ERSS | Complete | PASS |
| BAD-D | Complete | PASS |
| NICE | B.Ele.Th missing | ASSESSMENT INCOMPLETE |
| PS3 | Elevation factor not evaluated | ASSESSMENT INCOMPLETE |
| Report gate | Decision-critical dependency unresolved | No complete report token |
Final CER-AI result
ASSESSMENT INCOMPLETE
Report summary
CER-AI requests the exact missing BAD Display field or surgeon-confirmed value. It does not infer zero elevation, reuse a similarly named field, or issue a reassuring completed report.
Learning points
- Missingness is a workflow state, not a low-risk observation.
- Source-specific rereading protects against field interchange.
- The report becomes available only after completeness and integrity gates are satisfied.