Eir-Space/health-md-parser

Health.md Parser Skill

Parse and analyze Health.md files - the open standard for LLM-optimized healthcare data.

What This Skill Does

• Parse Health.md files into structured data
• Extract key information (medications, lab results, conditions)
• Generate LLM-optimized summaries for AI analysis
• Validate file format and clinical data
• Anonymize sensitive information based on privacy levels
• Create patient timelines and clinical insights

Setup

Install Dependencies

pip install health-md pyyaml beautifulsoup4

Usage Examples

# Parse a health.md file
python scripts/parse_health.py patient-001.health.md

# Generate LLM summary
python scripts/parse_health.py patient-001.health.md --summary

# Extract specific information
python scripts/parse_health.py patient-001.health.md --medications --labs

# Validate file format
python scripts/parse_health.py patient-001.health.md --validate

# Anonymize a record
python scripts/parse_health.py patient-001.health.md --anonymize

OpenClaw Integration

This skill integrates seamlessly with OpenClaw agents:

# In your OpenClaw agent
from skills.health_md_parser.scripts.parse_health import HealthMdParser

# Parse a health record
parser = HealthMdParser("patient.health.md")
record = parser.parse()

# Generate AI-optimized context
context = record.to_llm_context()

# Use in agent conversation
response = agent.process(f"""
Patient context: {context}

Question: What are the key health concerns for this patient?
""")

Clinical Use Cases

1. Clinical Decision Support

# Extract current medications and conditions
medications = record.get_current_medications()
conditions = record.get_conditions()

# Check for drug interactions or contraindications
analysis = analyze_clinical_context(medications, conditions)

2. Patient Education

# Generate patient-friendly summaries
summary = record.to_patient_summary()

# Create educational content based on conditions
education = generate_patient_education(record.get_conditions())

3. Research & Analytics

# Extract de-identified data for research
if record.get_privacy_level() == 'anonymous':
    research_data = record.to_research_dataset()

Privacy & Security

This skill respects Health.md privacy levels:

• Anonymous: No identifiable information processed
• Pseudonymized: Consistent fake identifiers maintained
• Identified: Full access (requires appropriate permissions)

All processing follows HIPAA/GDPR principles and the Health.md privacy specification.

File Format Support

Supports all Health.md specification features:

• ✅ Demographics with privacy-aware handling
• ✅ Current Medications with clinical context
• ✅ Lab Results with trend analysis
• ✅ Medical History with ICD coding
• ✅ Clinical Timeline with provider context
• ✅ Vital Signs with temporal tracking
• ✅ Allergies & Intolerances
• ✅ Care Team information

Examples

The examples/ directory contains sample Health.md files:

• anonymous-diabetes-patient.health.md - Comprehensive diabetes case
• cardiac-patient.health.md - Cardiovascular conditions
• mental-health-patient.health.md - Mental health considerations
• pediatric-patient.health.md - Pediatric healthcare data

Integration with EIR Space

This skill is designed to work with EIR Space health literacy platform:

# Convert Health.md to EIR-compatible format
eir_data = record.to_eir_format()

# Generate health literacy explanations
explanations = generate_health_explanations(record)

Error Handling

Robust error handling for clinical data:

try:
    record = HealthRecord.from_file('patient.health.md')
except HealthMdValidationError as e:
    # Handle validation errors (missing sections, invalid dates, etc.)
    print(f"Validation error: {e}")
except HealthMdPrivacyError as e:
    # Handle privacy violations
    print(f"Privacy error: {e}")

Advanced Features

Clinical Reasoning

# Generate clinical insights
insights = record.generate_clinical_insights()
# Returns: medication adherence, lab trends, care gaps, etc.

Timeline Analysis

# Analyze patient journey over time
timeline = record.get_clinical_timeline(days=365)
events = analyze_clinical_progression(timeline)

Risk Assessment

# Calculate health risks based on available data
risks = calculate_health_risks(record)
# Returns: diabetes complications, cardiovascular risk, etc.

Contributing

This skill is part of the open-source Health.md standard. Contributions welcome:

• Clinical validation - Review medical accuracy
• Parser improvements - Handle edge cases better
• New features - Additional analysis capabilities
• Privacy enhancements - Stronger anonymization

See CONTRIBUTING.md in the main repository.

---

Healthcare data deserves better standards. This skill helps make it reality. 🏥💙