MedGaugeCapabilities / D · Medical practice
D.3 Disease registry quality control Active
Given disease registry records, the agent recomputes scales and checks missing items, thresholds and time-point relationships, producing a QC report. Graded by pairing report entries against the reference recomputation.
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.4 Multi-source cross-analysisA.5 Temporal and knowability reasoningC.1 Data processingC.4 Data validation
- Required by
- E.1E.4
- Internal task families
- careline, care_journey
- External anchors
- LongMedBench, CHI-Bench
- Version
- 0.1.0 · 0.1.0
Definition in detail
Disease registries feed quality reporting and research. Scale totals that do not match their items, missing required fields and out-of-window time points distort every analysis built on the registry.
Inputs a task gives
- Registry records across several time points
- Raw scale item values alongside recorded totals
- A data dictionary with thresholds, required fields and time windows
Artifact required
A QC report with one entry per issue: record_id, field, check_type (scale recomputation, missing item, threshold, time-point relationship), recorded_value, recomputed_value where applicable, and the dictionary rule applied.
How it is graded
A program pairs report entries with a reference recomputation. Recomputed values must match the reference; recorded values are logged, not changed. 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
- Recomputing scale totals from items
- Checking required fields, thresholds and time windows
- Checking relationships between time points within a record
Out of scope
- Structural defects in the dictionary or mapping table itself - graded under D.16
- Follow-up target attainment and due dates - graded under D.14
- Gold standard versus raw data checks - graded under C.4
Typical failures
Expected failure patterns; not measured rates.
- Ignoring reverse-scored items when recomputing a total
- Imputing a missing item instead of reporting it
- Applying a time window from a different version of the dictionary
Standards: std-omop-cdm-5.4
Benchmarks
- Specialty Registry QC Planned
Grading references
No grading reference is mapped to this capability.