Proof Lab — Deterministic Artifacts

Every claim is traceable. Every decision is auditable.

Synthetic clinical scenarios processed by the deterministic extraction engine. Each artifact shows the full decision chain: entity matching, negation scoping, section priority, specificity resolution, conflict detection, and tier classification. Nothing is inferred. Everything is rule-executed.

All scenarios use synthetic notes. No PHI. No upload required.

4Scenarios
90%+Deterministic rule execution
0Critical false positives
SyntheticAll data

Complex MA Encounter

71yo male · CHF, DM, CKD, COPD, AFib · 9 conditions

Full coder workflow scenario. 6-page clinical note processed through deterministic extraction. Demonstrates conflict detection, tier classification, HCC mapping, and suppression logic.

9Codes surfaced
6HCC categories
1Conflicts
3Suppressions

COPD Conflict Scenario

Contradictory evidence across note sections

HPI reports cough and dyspnea suggesting exacerbation. Assessment says 'stable, no acute exacerbation.' Without extraction, this conflict auto-submits and triggers a denial.

1Conflicts detected
YesAuto-resolved
1Denial prevented

HCC Capture Scenario

Specificity drift and cross-family linkage

Demonstrates how implicit causal linkages, buried lab values, and PMH reassertion recover HCC categories that manual coding routinely misses.

2.8 avgHCC categories
4 typesRecovery mechanism
Evidence-backedRAF preserved

PHI Egress Gate

Three-layer fail-closed architecture

Four payload scenarios demonstrating the egress gate: clean pass, PHI in phrase field, unknown field injection, and structural deviation. Default is block.

3Gate layers
BlockDefault posture
4Test payloads

All scenarios use synthetic clinical notes. No PHI is required, accepted, or stored. The Proof Lab demonstrates extraction capability on representative clinical complexity — not your data. Your data stays in the pilot.