Clinical Gap Audit
Generated 2026-08-17T08:07:07Z. This page tracks the largest remaining clinical-coverage gaps. It is a maintainer audit, not clinical guidance.
| Gap | Current | Target | Status |
|---|---|---|---|
| Clinical sign-off | 15/2614 signoff-eligible entities reviewed (0.6%) | >=85% reviewed before public guideline-grade claims | blocked_on_reviewers |
| Solid tumor 2L+ coverage | 23/67 solid diseases have a 2L+ algorithm; 26/67 have a 2L+ indication | Every modeled solid disease has at least one advanced/relapsed-line algorithm and indication. | coverage_gap |
| Surgery/radiation detail | 21 structured procedure entities; 9 structured radiation-course entities; 816 indications mention surgery/radiation in text | Dedicated modality entities for surgery and radiation with dose/fraction/intent/timing fields. | coverage_gap |
| Supportive-care depth | 136/404 regimens have mandatory supportive care (33.7%); 43 have monitoring; 358 have dose adjustments | Every active regimen has supportive care, monitoring, dose-adjustment, and patient-watchpoint coverage. | coverage_gap |
| Drug indication and off-label tracking | 1/862 inferred drug-disease-indication pairs have a first-class record; 1 carry a source-backed assessed status | Every drug-use pair has explicit regulatory-label status, NCCN/ESMO category, and source provenance. | coverage_gap |
Clinical sign-off
Current: 15/2614 signoff-eligible entities reviewed (0.6%)
Blocker: Cannot be fixed by code; requires qualified Clinical Co-Lead review.
Next: Batch the largest STUB queues by entity type and assign reviewer owners.
Solid tumor 2L+ coverage
Current: 23/67 solid diseases have a 2L+ algorithm; 26/67 have a 2L+ indication
Blocker: Missing disease-by-disease 2L+ algorithm/indication authoring queue.
Next: Prioritize missing high-volume solid diseases, then rare solid diseases.
Surgery/radiation detail
Current: 21 structured procedure entities; 9 structured radiation-course entities; 816 indications mention surgery/radiation in text
Blocker: The modality schemas and initial entities exist, but many indication-level references remain prose-only.
Next: Prioritize high-volume prose mentions; attach an existing or newly verified procedure/radiation-course entity to each reviewed indication phase.
Supportive-care depth
Current: 136/404 regimens have mandatory supportive care (33.7%); 43 have monitoring; 358 have dose adjustments
Blocker: Supportive-care records exist, but regimen attachment is incomplete.
Next: Audit high-toxicity regimens first, then fill missing regimen attachments.
Drug indication and off-label tracking
Current: 1/862 inferred drug-disease-indication pairs have a first-class record; 1 carry a source-backed assessed status
Blocker: The first-class model is present, but most regimen-derived drug-use pairs still need source-backed assessment.
Next: Backfill high-volume and high-risk pairs from current regulatory labels first; keep uncertain pairs explicitly not_assessed until verified.
/audits/clinical_gap_audit.json
and /audits/clinical_gap_audit.md.