iris-medmatch 🏥🤖

An AI-powered patient matching engine built on InterSystems IRIS, utilizing Vector Search and Natural Language Processing (NLP) to find clinically similar patients across FHIR repositories.

🚀 Overview

iris-medmatch bridges the gap between traditional healthcare data and modern AI. While standard searches look for exact words, this engine understands clinical intent.

Example: It can match a patient with “Hypertension” to a search for “High Blood Pressure” using mathematical vector similarity.

✨ Key Features

• Semantic Search: Uses all-MiniLM-L6-v2 embeddings to vectorize clinical conditions.
• Vector Database: Leverages the native VECTOR data type in InterSystems IRIS 2024.1+.
• FHIR R4 Ready: Fully compatible with standard Patient and Condition resources.
• Embedded Python: Runs AI models directly inside the database for zero-latency inference.
• Modern UI: Angular-based dashboard to visualize similarity scores and patient data.

🛠️ Tech Stack

• Core: InterSystems IRIS , Embedded Python, InterSystems FHIR Server, Vector search
• AI: Python, ONNX Runtime, HuggingFace Transformers
• Frontend: Angular 18+
• DevOps: Docker & Docker Compose

📦 Getting Started

Prerequisites

• Docker Desktop
• InterSystems IRIS for Health 2024.1+ (Community Edition works great)

⚙️ Installation

Clone the Repository

git clone https://github.com/AshokThangavel/iris-medmatch.git
cd iris-medmatch

Running the Application with Docker

Build and start the app using Docker Compose:

docker-compose up --build

Stopping the Application

To stop and remove the running containers:

docker-compose down

In your README.md, the “Usage” section explains how a developer actually interacts with the project once the Docker containers are running.

Here is a clear breakdown you can use to explain these two endpoints:

🖥️ Usage & Access Points

Once the project is started, you can access the different layers of the application via these URLs:

1. Frontend Interface (Angular)

• URL: http://localhost:8080
• What it is: This is the user-facing dashboard.
• What you can do here: * Search for patients using natural language.
• Interact with the FHIR data visualized in a clean, modern UI.

2. Backend Management (InterSystems IRIS)

• URL: http://localhost:52773/csp/sys/UtilHome.csp
• What it is: The Management Portal for the InterSystems IRIS database.

🚀 Application Walkthrough

1. Authentication

• Demo Credentials: User: _SYSTEM | Pass: SYS

2. Semantic Similarity Search (The “Wow” Factor)

This module uses Vector Search to understand medical synonyms and clinical intent.

• How it works: A search for “Cardiac Issues” will mathematically find “Myocardial Infarction” by comparing their vector positions in IRIS.
• Tech Highlight: Uses Embedded Python to vectorize the search query and Native IRIS SQL to calculate similarity scores in sub-seconds.

3. Patient Directory & Condition Enrichment

This module manages existing FHIR resources. Users can add new diagnoses through a high-performance modal.

• The Workflow:

1. Create Condition for the particular patient and stored into the FHIR Server. Tech Highlight: Demonstrates real-time synchronization between standard FHIR data and AI-ready Vector data.

2. Add condition and save into InterSystems iris fhir server

4. New Patient Registration

A streamlined entry point for creating new Patient resources within the InterSystems ecosystem.

• Tech Highlight: Direct interaction with the FHIR R4 Repository via standard RESTful POST requests, ensuring data is indexed and searchable immediately.