health-md-standard

EIR Healthcare Viewer MCP Server

Model Context Protocol (MCP) server for viewing and analyzing EIR (Electronic health Information Record) files in AI conversations.

Overview

This MCP server allows users to load .eir healthcare files directly into any MCP-compatible AI application (Claude Desktop, ChatGPT, etc.) and have natural conversations about their health data.

Features

📁 File Loading - Load any .eir file into AI context
🔍 Data Exploration - Query medications, lab results, conditions, timeline
📊 Analysis - Generate insights, trends, and clinical summaries
🔒 Privacy-Aware - Respects privacy levels in EIR files
✅ Validation - Validates EIR format and provides helpful error messages
🌐 Multi-Language - Supports Swedish healthcare terminology

Quick Start

Installation

npm install -g eir-mcp-server
# or
pip install eir-mcp-server

Configuration

Add to your MCP client (Claude Desktop, etc.):

{
  "mcpServers": {
    "eir-healthcare": {
      "command": "eir-mcp-server",
      "args": []
    }
  }
}

Usage

Load EIR file: “Load my healthcare data from patient-2024.eir”
Query data: “What medications am I currently taking?”
Analyze trends: “How has my HbA1c changed over time?”
Get insights: “What are my main health risks?”

MCP Tools

`load_eir_file`

Load an EIR healthcare file into the conversation context.

Parameters:

file_path (string): Path to the .eir file
privacy_check (boolean): Verify privacy level before loading

Returns:

Structured healthcare data ready for AI analysis
Patient summary and key metrics
Validation status and warnings

`query_medications`

Get current medications and treatment details.

Parameters:

active_only (boolean): Show only active medications
include_history (boolean): Include medication history

`query_lab_results`

Retrieve lab results with filtering options.

Parameters:

test_name (string, optional): Filter by specific test
days_back (number): How many days back to include
include_trends (boolean): Calculate trends and changes

`query_conditions`

Get medical conditions and diagnoses.

Parameters:

status (enum): active, resolved, chronic, all
include_icd_codes (boolean): Include ICD-10/SNOMED codes

`analyze_timeline`

Generate clinical timeline and care journey analysis.

Parameters:

days_back (number): Time period to analyze
event_types (array): Filter by visit types
include_outcomes (boolean): Include treatment outcomes

`validate_eir_file`

Validate EIR file format and content.

Parameters:

file_path (string): Path to file to validate
strict_mode (boolean): Enable strict validation rules

`generate_summary`

Create comprehensive health summary for AI analysis.

Parameters:

summary_type (enum): brief, detailed, clinical, patient_friendly
focus_areas (array): Specific areas to emphasize

Example Conversations

Loading and Basic Queries

User: Load my healthcare data from ~/health/birger-2024.eir

AI: I've loaded your healthcare data (privacy level: pseudonymized). I can see:
- 2 active medical conditions (Type 2 Diabetes, Hypertension) 
- 2 current medications (Metformin, Lisinopril)
- Recent lab results showing excellent diabetes control (HbA1c: 6.8%)
- 15 clinical events over the past year

What would you like to know about your health data?

User: How is my diabetes management going?

AI: Your diabetes management is going exceptionally well! Here's what I see:

**Blood Sugar Control:**
- HbA1c improved dramatically: 8.2% → 6.8% (now at target <7%)
- Latest fasting glucose: 118 mg/dL (near normal range)
- Trend: Consistently decreasing since starting metformin

**Treatment Response:**
- Metformin 500mg twice daily - excellent adherence (90%+)
- No reported side effects
- 7kg weight loss since diagnosis (78kg current)

**Key Achievements:**
✅ Met HbA1c target in just 4 weeks
✅ Sustainable weight loss
✅ No hypoglycemic episodes
✅ Good blood pressure control

Your endocrinologist Dr. Smith noted "excellent response to lifestyle and metformin" at your last visit. Keep up the great work!

Trend Analysis

User: Show me how my lab values have changed over time

AI: Here's your lab trend analysis:

**Hemoglobin A1C (Diabetes Control):**
📉 Dec 2023: 8.2% (diagnostic)
📉 Jan 2024: 7.4% (improving) 
📉 Feb 2024: 6.8% (target achieved!)
Trend: -1.4% in 2 months (excellent improvement)

**Blood Pressure:**
📉 Aug 2023: 158/96 (high, pre-treatment)
📉 Dec 2023: 145/88 (improving)
📉 Feb 2024: 126/78 (well controlled)
Trend: 32/18 mmHg reduction (great response to lisinopril)

**Weight Management:**
📉 Oct 2023: 85kg (BMI 31.2)
📉 Feb 2024: 78kg (BMI 28.7)
Trend: -7kg in 4 months (sustainable pace)

**Clinical Insight:** Your multi-faceted approach (medication + lifestyle) is working excellently. All major health metrics trending in the right direction!

Risk Assessment

User: What should I be watching out for health-wise?

AI: Based on your health data, here are key areas to monitor:

**Immediate Priorities:**
1. **Cardiovascular Risk** - You have diabetes + controlled hypertension
   - Current LDL: 128 mg/dL (target <100 for diabetics)
   - Consider statin discussion with your doctor
   - Keep up excellent BP control (current: 126/78)

2. **Diabetes Complications Prevention**
   - Annual diabetic eye exam (scheduled ✅)
   - Foot exams (last exam normal ✅)
   - Kidney function monitoring (current eGFR >60 ✅)

**Positive Factors:**
✅ Excellent diabetes control (HbA1c 6.8%)
✅ No medication side effects
✅ Strong healthcare team (endocrinologist + primary care)
✅ Good medication adherence

**Recommended Actions:**
- Continue current diabetes management
- Discuss cholesterol treatment at next visit
- Maintain weight loss momentum (goal BMI <25)
- Keep up regular monitoring schedule

**Red Flags to Watch For:**
- Frequent hypoglycemia episodes
- Rapid weight gain/loss
- Vision changes
- Foot injuries or numbness

Your proactive approach and excellent clinical response suggest low risk for complications. Keep doing what you're doing!

Privacy and Security

Privacy Levels

Anonymous: No personal identifiers displayed
Pseudonymized: Consistent fake names, full clinical data
Identified: Real names and contact information (use carefully)

Security Features

Local processing: EIR files processed locally, not sent to servers
Validation: Schema validation prevents malformed data processing
Access control: Respects file permissions and privacy settings
Audit trail: Logs all data access for compliance

GDPR/HIPAA Compliance

Data minimization: Only loads requested data sections
Retention limits: Data cleared after conversation ends
Consent tracking: Records user consent for data processing
Right to erasure: Complete data deletion on request

Swedish Healthcare Integration

1177 Vårdguiden Support

Native terminology: Understands Swedish medical terms
Provider mapping: Maps Swedish healthcare providers
ICD-10-SE: Supports Swedish ICD-10 extensions
Referral system: Understands Swedish referral patterns

Regional Integration

Uppsala Region: Enhanced support for local providers
Folktandvården: Dental care integration
Vaccination records: Integration with Swedish vaccination system

Development

Contributing

git clone https://github.com/BirgerMoell/eir-mcp-server
cd eir-mcp-server
npm install
npm run dev

Testing

# Run tests
npm test

# Test with sample EIR file
npm run test:sample

# Validate EIR files
npm run validate examples/*.eir

API Documentation

Full API documentation available at: https://docs.eir.space/mcp-server

Roadmap

Version 1.1 (Q2 2024)

Real-time sync with Swedish healthcare systems
Family sharing for caregiver access
Predictive analytics for health risk assessment
Integration with Apple Health, Google Fit

Version 2.0 (Q3 2024)

Multi-patient support for healthcare providers
Clinical decision support with evidence-based recommendations
Population health analytics for research
Blockchain integration for data provenance

Support

Documentation: https://docs.eir.space/
Issues: https://github.com/BirgerMoell/eir-mcp-server/issues
Discord: https://discord.gg/eir-format
Email: support@eir.space

EIR MCP Server - Healthcare data that actually works with AI. 🏥🤖💙

This site is open source. Improve this page.