• 

  The window will look like this for the application displaying all the details.

About the Project

Inspiration Preventive care is one of the most effective ways to improve patient outcomes and reduce healthcare costs, yet care gaps are common. As a healthcare provider, it's challenging to manually track every recommended preventive measure for each patient. The Care Gap Predictor was designed to address this challenge by automatically identifying potential care gaps based on patient demographics and medical history, helping clinicians provide more comprehensive care. What We Learned This project deepened our understanding of healthcare interoperability standards, particularly HL7 FHIR and SMART on FHIR. We gained valuable experience implementing CDS Hooks for clinical decision support and learned how to create an app that integrates seamlessly with electronic health record systems. The project also demonstrated how AI algorithms can analyze patient data to generate meaningful clinical recommendations. How We Built It The Care Gap Predictor was built using a modern tech stack:

Backend:

Express.js server for the CDS Hooks service AI prediction algorithms to identify care gaps based on patient demographics ngrok for secure tunneling to our local development environment

Frontend:

React for the user interface Tailwind CSS for styling SMART on FHIR integration for EHR launch capabilities

Integration:

MeldRx platform for FHIR data access CDS Hooks for delivering recommendations within clinical workflow SMART on FHIR for launching the detailed analysis dashboard

The application follows a two-step workflow:

The CDS Hooks service analyzes patient data and returns care gap cards with recommendations Clinicians can click on "View Detailed Analysis" to launch the dashboard for more comprehensive information

Challenges Faced We encountered several challenges during development:

FHIR Integration: Understanding the FHIR data model and properly extracting patient information required careful study of the specification. Authentication Flow: Implementing the SMART on FHIR authentication flow was complex, requiring precise configuration and error handling. Cross-Origin Requests: Setting up proper CORS handling between our service and MeldRx required several iterations. Clinical Decision Logic: Developing algorithms that provide meaningful care gap recommendations based on limited patient data required balancing simplicity with clinical relevance.

Despite these challenges, we were able to build a functional application that demonstrates the potential of AI-powered preventive care assistance. Built With

JavaScript/Node.js Express.js React Tailwind CSS HL7 FHIR SMART on FHIR CDS Hooks MeldRx Platform

Try It Out To experience the Care Gap Predictor: - CDS Hooks Service: https://adjusted-bluejay-living.ngrok-free.app/cds-services To set up and run your own instance: 1. Clone the repository 2. Install dependencies with npm install 3. Start the CDS Hooks service with node server.js 4. Start the SMART app with `npm

Built With

• cds
• css
• express.js
• fhir
• hl7
• hooks
• javascript/node.js
• meldrx
• on
• react
• smart
• tailwind