The Definitive Operations Playbook: PHP Dual Diagnosis Scribe on Sigmund Aura

Table of Contents

• Section 01 — Forensic Logic Group Documentation Reality

• Section 02 — Sigmund Aura Configuration Technical DOM Mapping

• Section 03 — Clinical Logic Masterclass The IOP Evening Group

• Section 04 — Part 2 Inference Isolation 2026 Segmentation Gap

• Section 05 — Audit Defense Provenance & Re-Disclosure

• Section 06 — ROI & Deployment Operational Economics

Forensic Logic: The Group Documentation Reality

CLINICAL UPDATE JUNE 2026: Revised for new CMS standards and Group Diarization accuracy.

Every PHP dual diagnosis group generates a documentation paradox that legacy scribes cannot resolve. One clinical event produces both a shared group artifact and ten distinct medico-legal records. Sigmund Aura's discrete-field architecture demands both be written cleanly.

Most vendors treat group notes as a single, unsegmented block of text. This collapses ten patients into one liability surface and violates 42 CFR Part 2 by cross-referencing identities. The result is silent re-disclosure risk buried in every progress note.

Your Clinical Directors face three simultaneous pressures inside PHP settings. CMS demands minute-level attendance justification, payers demand individualized medical necessity, and Part 2 demands cross-patient PHI isolation. No manual workflow survives all three at scale.

• Double-entry burden crushes throughput when clinicians rewrite the same group content into ten separate Aura records manually.

• Cross-patient PHI leakage occurs when a note mentions another participant's disclosure or safety-planning comment.

• Attendance minute discrepancies trigger CMS clawbacks when group duration is not reconciled per participant.

Sigmund Aura Configuration: Technical DOM Mapping

Sigmund Aura enforces restrictive discrete fields that reject bulk paste operations. Our Chrome extension bypasses this by mapping directly to the DOM selectors of the group template and individual progress note surfaces. This eliminates the copy-paste failure mode entirely.

The configuration logic below demonstrates the write-back pattern for Aura's dual-surface structure. The same principle governs our Behave Health Integration and Kipu AI Workflow deployments.

// scribing.io :: Sigmund Aura DOM Injection Map
const AURA_SELECTORS = {
  groupSummary: "#aura-group-note[data-template='php_group']",
  patientNote:  "div.progress-note[data-patient-id='{PID}']",
  riskFlag:     "select.risk-level[data-encounter='{ENC}']",
  attendMin:    "input.attendance-minutes[data-patient-id='{PID}']"
};

function injectDualOutput(session) {
  writeField(AURA_SELECTORS.groupSummary, session.masterSummary);
  session.patients.forEach(p => {
    if (p.part2) segmentExport(p);          // Inference Isolation
    writeField(resolve(AURA_SELECTORS.patientNote, p.id), p.note);
    setSelect(resolve(AURA_SELECTORS.riskFlag, p.encId), p.risk);
    writeField(resolve(AURA_SELECTORS.attendMin, p.id), p.minutes);
  });
  emitProvenanceLedger(session.consentLedger);
}

// scribing.io :: Sigmund Aura DOM Injection Map
const AURA_SELECTORS = {
  groupSummary: "#aura-group-note[data-template='php_group']",
  patientNote:  "div.progress-note[data-patient-id='{PID}']",
  riskFlag:     "select.risk-level[data-encounter='{ENC}']",
  attendMin:    "input.attendance-minutes[data-patient-id='{PID}']"
};

function injectDualOutput(session) {
  writeField(AURA_SELECTORS.groupSummary, session.masterSummary);
  session.patients.forEach(p => {
    if (p.part2) segmentExport(p);          // Inference Isolation
    writeField(resolve(AURA_SELECTORS.patientNote, p.id), p.note);
    setSelect(resolve(AURA_SELECTORS.riskFlag, p.encId), p.risk);
    writeField(resolve(AURA_SELECTORS.attendMin, p.id), p.minutes);
  });
  emitProvenanceLedger(session.consentLedger);
}

// scribing.io :: Sigmund Aura DOM Injection Map
const AURA_SELECTORS = {
  groupSummary: "#aura-group-note[data-template='php_group']",
  patientNote:  "div.progress-note[data-patient-id='{PID}']",
  riskFlag:     "select.risk-level[data-encounter='{ENC}']",
  attendMin:    "input.attendance-minutes[data-patient-id='{PID}']"
};

function injectDualOutput(session) {
  writeField(AURA_SELECTORS.groupSummary, session.masterSummary);
  session.patients.forEach(p => {
    if (p.part2) segmentExport(p);          // Inference Isolation
    writeField(resolve(AURA_SELECTORS.patientNote, p.id), p.note);
    setSelect(resolve(AURA_SELECTORS.riskFlag, p.encId), p.risk);
    writeField(resolve(AURA_SELECTORS.attendMin, p.id), p.minutes);
  });
  emitProvenanceLedger(session.consentLedger);
}

This DOM-mapped write-back achieves zero double entry across the entire roster. One click populates the Master Group Summary plus every individual note, risk flag, and attendance field. The clinician never leaves the Speaker Review Dashboard.

Clinical Logic Masterclass: The IOP Evening Group

Consider an IOP evening group of ten patients, co-facilitated, running sixty minutes with mixed MDD and OUD presentations. This is the exact stress-test our V6 architecture was engineered to survive. Follow the lifecycle below step by step.

Pre-Session Roster Preparation

The clinician selects Group Session and pre-loads the roster directly from the Chrome extension. Two late adds are appended without breaking the encounter structure. The roster now anchors every downstream diarization assignment.

Processing and Diarization

Post-session audio uploads to the Aura-linked processing pipeline. Diarization separates voices and drafts a note skeleton per identified speaker. Each utterance is timestamped and bound to a roster position, not a raw voice cluster.

Interactive Speaker Review Dashboard

The clinician opens the Interactive Speaker Review Dashboard to correct attribution errors. A misattributed safety-planning comment is reassigned to the correct patient in two clicks. High-risk utterances are confirmed with their original timestamps intact.

The Dual-Output View

The Dual-Output View then splits the corrected transcript into two distinct artifacts. The Master Group Summary captures topic, modality, and interventions with no patient identifiers whatsoever. Each Per-Patient Progress Note personalizes the record independently.

Part 2 Inference Isolation in Action

Inference Isolation auto-masks cross-patient references in every individual note simultaneously. A consent ledger is recorded for each participant at the moment of segmentation. Patients covered by Part 2 have their exports segmented and tagged with full provenance.

One-Click EHR Injection

The clinician clicks inject once and Sigmund Aura's discrete fields populate instantly. The group template, ten individual progress notes, and risk flags are DOM-mapped and written back. The result is compliant, audit-ready documentation with no cross-patient PHI leakage.

Part 2 Inference Isolation: The 2026 Segmentation Gap

The critical 2026 technical gap is Part 2-compliant Inference Isolation for group documentation. Most vendors still export a group note as one unsegmented file, ignoring per-patient consent boundaries entirely. This exposes every operator to re-disclosure findings.

Our architecture applies per-patient consent segmentation to both the Master Group Summary and each individual note. Re-disclosure accounting travels with every artifact as structured metadata. Diarization and transcription provenance is preserved through every AI-assisted edit.

• Consent segmentation isolates each patient so a Part 2 participant's disclosures never surface in another's record.

• Re-disclosure accounting logs every export destination, giving Clinical Directors a defensible chain of custody.

• Provenance preservation keeps AI edits exportable and auditable at the individual participant level, not just the group level.

This is the definitive information gain over legacy scribes. Where competitors deliver a single flat artifact, we deliver a segmented, provenance-tagged, participant-auditable record set. That difference is the audit boundary.

Audit Defense: Provenance and Re-Disclosure

When a Part 2 audit arrives your defense is the consent ledger, not clinician memory. Each Per-Patient Progress Note carries its own provenance chain from raw diarization to final injection. The Master Summary proves no identifiers ever crossed.

Timestamped high-risk utterances form the backbone of medical necessity defense. Every confirmed safety-planning comment links to its origin second in the session audio. CMS reviewers see minute-level attendance reconciled per participant.

ROI and Deployment: Operational Economics

The economic case is direct and measurable across your PHP census. Eliminating double entry across a ten-patient group recovers roughly forty minutes of clinician time per session. Multiply that across daily evening groups for the annualized recovery.

Model your specific program numbers using the AI Scribe ROI Calculator before deployment. Clinical Directors typically see payback inside the first quarter. The compliance risk reduction is the uncosted upside.

Deploy the Chrome extension first against a single evening IOP group as a controlled pilot. Validate the diarization accuracy in the Speaker Review Dashboard for one week. Then scale the DOM injection across every PHP and IOP track in Sigmund Aura.