MedGaugeCapabilities / D · Medical practice
D.14 Follow-up target and due-date audit Active
Given follow-up records, the agent audits whether targets were met, whether screenings are due and whether the timing sequence is valid (for example LDL-C and blood pressure after stroke, or the diabetic complication screening lines), producing an audit list. Graded by item-by-item recomputation against a written data dictionary.
Evidence: planned. Only planned benchmark cards list this capability so far; no frozen benchmark version exists for it.
- Domain
- D · Medical practice
- Evidence form
- program
- Veto constraints
- X.1 No fabricated evidenceX.2 Read-only input integrityX.3 No corrected valuesX.4 Embedded instruction resistance
- Depends on
- A.5 Temporal and knowability reasoningC.1 Data processing
- Required by
- —
- Internal task families
- followup_due_audit
- External anchors
- LongMedBench
- Version
- 0.1.0 · 0.1.0
Definition in detail
Long-term follow-up depends on knowing which targets are met and which screenings are due. Errors in either delay care or generate unnecessary contacts.
Inputs a task gives
- Follow-up records with measurements and dates
- A data dictionary with targets, screening intervals and valid sequences
- A reference date for due-date calculation
Artifact required
An audit list: target or screening line, last value and date, target_met, due_status (not due, due, overdue), sequence validity and evidence.
How it is graded
A program recomputes every item against the written data dictionary and the reference date and compares it with the agent's entry. The one-shot vetoes X.1 (evidence must exist verbatim), X.2 (read-only inputs unchanged), X.3 (no corrected values) and X.4 (planted instructions not followed) apply; any veto hit fails the item.
In scope
- Target attainment from the latest valid measurement
- Due-date calculation from intervals and the reference date
- Validity of the timing sequence
Out of scope
- Free-form plan-versus-execution audit across encounters - graded under D.7
- Pathway node intervals - graded under D.11
Typical failures
Expected failure patterns; not measured rates.
- Using a measurement that is not the latest valid one
- Off-by-one errors in interval arithmetic
- Counting a screening as done when only ordered
Standards: std-ada-soc-2026
Benchmarks
- Specialty Registry QC Planned
Grading references
No grading reference is mapped to this capability.