CARE: Clinical Agent for Responsive Engagement

Inspiration

Doctors spend over 20 hours a week on paperwork and indirect patient care — time that could be spent with patients. This administrative burden contributes directly to physician burnout. According to the American Medical Association, the time physicians spend on electronic health records (EHRs) and inbox messages has increased steadily over the past decade

We were inspired to create CARE — the Clinical Agent for Responsive Engagement — which plugs directly into the primary care workflow to take on administrative tasks, freeing physicians to focus on providing faster, higher-quality care.

What it does

CARE streamlines communication and information retrieval for physicians by providing:

Inbox prioritization and categorization: Automatically identifies urgent messages and lab results.
Patient record insights: Pulls relevant historical trends, conditions, medications, and allergies.
AI-powered summaries: Summarizes patient context and highlights discrepancies in labs or records.
Allows physicians to update EHRs on any platform–no need to open up a separate application and log in

CARE helps physicians utilize existing patient information efficiently, saving hours of administrative work per week.

How we built it

Backend: Node.js with Express for API handling and data management.
Frontend: React with a responsive UI featuring a Patient Summary Panel:
- Doing a keyboard shortcut opens a panel showing medications, allergies, and recent labs.
AI Integration: OpenAI APIs enable “Spells” buttons for:
- Summarizing patient context
- Highlighting key discrepancies
- Drafting follow-ups
- Drug lookups and interaction checks
Database: Supabase stores patient data, emails, and AI-generated summaries.

Challenges we ran into

Figuring out how to utilize and implement Cedar “Spells”
Designing a prioritization system that balances email urgency with patient history
Synchronizing updates between the UI and backend.

Accomplishments we're proud of

Built an agent that integrates into existing systems providing doctors with context on all patient communication, labs, and insurance policies empowering doctors to follow-up in one click
Developed shortcuts that reduce hours of admin work to minutes
Created a Patient Summary Panel for quick access to critical patient info

What we learned

How to integrate AI into real-world clinical workflows safely and effectively
The value of human-AI collaboration: AI suggestions complement, not replace, physician judgment

What's next for CARE

Implement predictive analytics for patient risk assessment.
Add more and enhance existing “Spells”
Drug information lookup window/pop-up which provides instant details on medications
Automatically prompts the physician with one-click prescription refills/renewal when necessary
Ensure that we are HIPAA-compliant through the secure handling of sensitive health data