Clinical Notes — Cowork Plugin

v1.1 - Turn rough clinical notes into structured SOAP notes, treatment plans, discharge summaries, progress notes, and referral letters — with regulatory compliance and clinical intelligence — included with your subscription per provider.

---

Important: All documents generated by this plugin are drafts intended as starting points. Clinical and healthcare documents must be reviewed by a qualified healthcare professional before use. This plugin does not provide medical advice and is not a substitute for professional clinical judgment. Ensure all output complies with your organization's HIPAA policies and applicable regulations.

Installation

1. Download the pf-clinical-notes.plugin file
2. Open Claude Desktop and navigate to Settings > Plugins
3. Click Install Plugin and select the downloaded .plugin file
4. The plugin will be installed and available immediately

Note: All data stays local on your machine. No external API calls or cloud storage required.

Why This Exists

Clinicians pay $3,588–$9,000 annually for AI-powered note formatting (Suki AI, Freed AI, Nuance DAX). Clinical Notes offers the same structured output locally, with zero per-provider cost, full HIPAA alignment, and no vendor lock-in.

Hospital systems and practices spend thousands every year training staff on documentation standards. This plugin automates the tedious reformatting work—turning scribbled encounter notes into professional, auditable SOAP documents.

---

Quick Start

1. Setup (One-Time)

/clinical-notes-setup /path/to/workspace

Creates folder structure, checks Python/dependencies, generates practice config.

2. Add Notes

Drop raw clinical notes (.txt or .md) into inbound/ folder.

3. Run

/clinical-notes-batch-format

Processes all files, generates DOCX output.

4. Review & Use

Open generated DOCX files in outbound/, provider reviews, signs, integrates into EHR.

---

Commands

Command	Purpose
/clinical-notes-setup	Initialize workspace, check dependencies, configure practice
/clinical-notes-soap-note	Generate SOAP note with E/M level estimation and ICD-10 validation
/clinical-notes-treatment-plan	Create treatment plan with SMART goals and escalation pathways
/clinical-notes-discharge-summary	Generate hospital discharge with readmission prevention (AHRQ RED)
/clinical-notes-batch-format	Process multiple notes with CDI query triggers and quality scoring
/clinical-notes-full-pipeline	Verify setup + run batch with dry-run and checkpoints
/clinical-notes-status	View workspace status, config, recent outputs, statistics
/clinical-notes-progress-note	Generate hospital progress note or H&P follow-up (NEW)
/clinical-notes-referral-letter	Generate specialist referral or consultation letter (NEW)

---

How It Works

Three-Stage Document Automation with Regulatory Compliance:

1. Input: Raw encounter notes (unstructured text, handwritten transcription, voice-to-text)
2. Processing: AI synthesis → classification → regulatory validation (CMS E/M, ICD-10-CM FY2026, HIPAA, Joint Commission) → DOCX generation with python-docx library
3. Output: Professional DOCX with [AI-Assisted] markers, medical disclaimer, regulatory compliance statement, provider review checkpoint with clinical intelligence (E/M level, CDI triggers, medication safety)

Folder Architecture:

inbound/          → Drop raw notes here
processing/       → Batch reports, file tracking, logs
outbound/         → Generated DOCX files
archive/          → Historical records (optional)
.clinical-notes/  → Config, README, setup logs

---

What It's an Alternative To

Tool	Cost	Deployment	Lock-In
Suki AI	$3,588–$4,788/provider/year	Cloud-hosted	Proprietary format
Freed AI	$2,000+/user/year	Cloud-hosted	Vendor-locked templates
Nuance DAX	Enterprise pricing (000s)	Enterprise servers	Tight EHR integration
Clinical Notes	included with your subscription/provider	Local, self-hosted	Open DOCX

Clinical Notes advantage: No per-user licensing, no data transmission to external servers, HIPAA alignment with your infrastructure, standard DOCX output compatible with any EHR.

---

Key Differentiators

1. Regulatory Compliance Built-In

Enforces CMS E/M Documentation Guidelines (1995 & 1997), ICD-10-CM FY2026 specificity/sequencing/laterality rules, Joint Commission RC.02.01.01, CMS Conditions of Participation §482.24(c) & §482.43, and HIPAA Privacy Rule 45 CFR §164.502–514. Every document includes compliance statements and audit trails.

2. Clinical Intelligence (Domain Knowledge)

E/M level estimation (99202–99215), CDI query triggers, medication safety (ISMP High-Alert, AGS 2023 Beers Criteria), treatment escalation pathways, SMART goals, specialty-specific documentation profiles, and quality scoring rubric.

3. included with your subscription

No subscription, no per-provider seat licensing, no hidden fees. DOCX generation via proprietary python-docx library.

4. Local Data Processing

All notes remain on your system. No transmission to cloud AI services. Complies with strict hospital data residency policies and HIPAA minimum necessary principle.

5. Customizable

Edit .clinical-notes/config.json to set practice name, specialty, provider defaults. Generate output matches your practice standards, not vendor template.

6. No Vendor Lock-In

Output is standard DOCX files. Works in Word, Google Docs, LibreOffice. Move between systems freely. Not trapped in proprietary formats.

7. Transparent AI

[AI-Assisted] and [AI-Suggested] markers show exactly what was automated vs. provider-documented. Medical disclaimer + regulatory compliance statement + provider signature ensure accountability.

---

AI-Powered Features (14 Total)

Core Document Generation:

1. SOAP Structuring: Automatically converts unstructured clinical notes into proper Subjective-Objective-Assessment-Plan format
2. Clinical Synthesis: Generates assessment and clinical reasoning based on presented findings with differential diagnosis suggestions
3. Treatment Planning: Creates actionable medication and intervention plans with SMART goals and escalation pathways
4. Discharge Summaries: Synthesizes hospital course narratives with readmission prevention (AHRQ Project RED)
5. Progress Notes: Generates hospital follow-up documentation with interval history and plan updates
6. Referral Letters: Creates specialist consultation requests with clear clinical context

Regulatory & Clinical Intelligence: 7. E/M Level Estimation: Recommends evaluation and management codes (99202–99215) based on CMS MDM complexity matrix 8. ICD-10 Code Validation: Enforces FY2026 specificity, sequencing, laterality, and 7th character rules (CMS guidelines) 9. CDI Query Triggers: Detects clinical documentation improvement opportunities per ACDIS standards 10. Medication Safety: ISMP High-Alert Medications, AGS 2023 Beers Criteria, abbreviation safety (ISMP/TJC "Do Not Use" list) 11. Treatment Escalation: Pathway recommendations for T2DM, Hypertension, Depression, COPD, Heart Failure (ADA/AHA/GOLD/APA/KDIGO) 12. Document Quality Scoring: Rubric-based assessment (HPI completeness, ICD-10 specificity, ISMP compliance, E/M sufficiency)

Operational: 13. Abbreviation Expansion: Detects and expands medical abbreviations (HTN→Hypertension, DM2→Diabetes Mellitus Type 2) 14. Batch Processing: Processes multiple clinical notes simultaneously with CDI triggers and quality scoring while following HIPAA guidelines

Feature Comparison

Feature	Clinical Notes	Suki	Freed	Nuance DAX
SOAP note generation	✅	✅	✅	✅
Treatment plan creation	✅	✅	⚠ Limited	✅
Discharge summary	✅	✅	⚠ Limited	✅
Progress notes	✅	⚠ Limited	❌	⚠ Limited
Referral letters	✅	⚠ Limited	❌	⚠ Limited
Batch processing	✅	❌	❌	✅
E/M level estimation	✅	⚠ Limited	❌	⚠ Limited
CDI query triggers	✅	❌	❌	❌
Abbreviation expansion	✅	✅	✅	✅
ICD-10 code validation (FY2026)	✅	⚠ Limited	⚠ Limited	⚠ Limited
Medication safety checks	✅	⚠ Limited	❌	⚠ Limited
Document quality scoring	✅	❌	❌	❌
Medical disclaimer	✅	✅	✅	✅
Local data processing	✅	❌	❌	❌
Standard DOCX output	✅	⚠ Proprietary	⚠ Proprietary	⚠ Proprietary
Zero subscription cost	✅	❌	❌	❌
Provider review checkpoints	✅	✅	✅	✅
Customizable templates	✅	⚠ Limited	⚠ Limited	⚠ Limited
Resume capability (batch)	✅	⚠ Limited	❌	⚠ Limited

---

Estimated Cost per Use

Disclaimer: Token estimates are approximate and based on typical usage patterns measured from skill prompt sizes. Actual costs vary with input data size, conversation length, and complexity. Estimates use Claude Sonnet 4.6 pricing ($3/1M input, $15/1M output). Cowork and Claude Desktop subscription users (Pro/Max/Team) are not charged per-token — these estimates apply only to direct Anthropic API usage. Running stages individually in fresh sessions uses fewer input tokens than running the full pipeline sequentially, because pipeline mode accumulates conversation history across stages.

Per skill (run individually in a fresh session):

Stage	Skill Prompt	User Input	Total Input	Output	Est. Cost
soap-note	~17.4K	~800	~21.9K	~6.0K	~$0.16
treatment-plan	~25.2K	~800	~29.7K	~6.0K	~$0.18
progress-note	~8.6K	~800	~13.1K	~1.0K	~$0.05
referral-letter	~10.7K	~800	~15.5K	~6.0K	~$0.14
discharge-summary	~17.5K	~800	~22.2K	~6.0K	~$0.16
batch-format	~10.7K	~800	~15.4K	~6.0K	~$0.14
Standalone total			~117.7K	~31.0K	~$0.82

Full pipeline (all stages in one session — context accumulates):

Stage	Base Input	+ History	Total Input	Output	Est. Cost
soap-note	~22.6K	0	~22.6K	~6.0K	~$0.16
treatment-plan	~30.4K	~6.8K	~37.2K	~6.0K	~$0.20
progress-note	~13.7K	~13.6K	~27.3K	~1.0K	~$0.10
referral-letter	~15.9K	~15.4K	~31.3K	~6.0K	~$0.18
discharge-summary	~22.7K	~22.2K	~44.9K	~6.0K	~$0.22
batch-format	~15.9K	~29.0K	~44.9K	~6.0K	~$0.22
Pipeline total			~208.3K	~31.0K	~$1.09

Running the full pipeline once typically costs $0.76–$1.42 in API tokens (Claude Sonnet 4.6).

Known Limitations & Workarounds

Limitation	Impact	Workaround
Requires Python 3.8+	System dependency	Pre-installed on most systems; install from python.org if needed
Input must be text (.txt/.md)	Cannot process PDFs directly	Convert PDF to text manually or use OCR tool first
AI synthesis requires context	Vague input → generic output	Provide complete encounter notes (vitals, exam, assessment) for quality output
ICD-10 codes suggested (not verified)	Provider must verify accuracy	Use official coding guidelines; codes are suggestions only
Batch limited to 20 files	Large practices need multiple runs	Run batch repeatedly; resume logic skips already-processed files

---

Context & Performance Guide

Session Management

• Cowork mode: 100,000-token context per session (sufficient for 200–500 notes per batch)
• File size limits: Up to 50 MB total in inbound/ folder (practical limit ~500 files)
• Processing time: 30–90 seconds per note depending on length and complexity

Performance Tips

1. Batch in groups of 10–20: Balances processing time and session context
2. Use plain text input: Markdown also supported; binary formats require conversion
3. Archive old files: Move processed input to archive/ periodically to keep inbound/ clean
4. Review spot-check: Check 3–5 generated DOCX files per batch for quality
5. Organize notes chronologically: Process oldest first for better narrative flow

---

Requirements

• Claude Desktop: Latest version (Cowork plugin support)
• Cowork Mode: Enabled for this workspace
• Python 3.8+: For document generation
• python-docx: Auto-installed during setup
• Disk space: ~100 MB for typical 50-note batch
• Internet: Only needed for initial setup (plugin download); processing is local

---

Important Disclaimers

• AI-Generated Content: This plugin uses AI (LLM) technology which can produce inaccurate or incomplete outputs. All content should be treated as a starting point and reviewed for accuracy before use.
• Not Professional Advice: Outputs do not constitute legal, financial, tax, medical, or other professional advice. Consult qualified professionals before making decisions based on generated content.
• No Compliance Guarantee: References to industry standards, regulations, or guidelines are for informational purposes only. This plugin does not guarantee compliance with any law or regulation. Users are responsible for verifying all outputs meet their specific regulatory requirements.
• No Endorsement or Affiliation: Mention of third-party products, standards, or organizations does not imply endorsement, partnership, or certification by those entities.
• Not Medical or Health Advice: Health and medical-related outputs are not a substitute for professional medical advice, diagnosis, or treatment. Always consult a licensed healthcare provider for health-related decisions.