Care Bridge AI

• 

• 

• 

Slide Deck

https://gamma.app/docs/CareBridge-fhphqurekdi1y84

Inspiration

CareBridge AI was inspired by the real consequences of broken care coordination in rural communities like Clearwater Ridge. The winter storm that cut off access to the regional hospital exposed a deeper, ongoing problem: missed referrals, fragmented communication, and lack of systematic follow-up can quietly escalate into preventable emergencies.

The cardiology incident described in the case—where a missed appointment was never tracked or rescheduled—highlighted how fragile manual systems are when travel becomes impossible or communication breaks down. We wanted to build a solution that doesn’t just digitize healthcare interactions, but actively closes the loop on care: ensuring that missed appointments, worsening symptoms, and risk signals are seen, tracked, and acted on before they turn into crises.

What it does

CareBridge AI is a healthcare coordination platform that connects patients and providers through intelligent check-ins, risk-aware timelines, and centralized patient dashboards.

For patients, it provides:

• An AI-powered care companion for health check-ins and symptom reporting

• Conversational interactions that generate structured health summaries

• A low-friction, private way to report concerns without needing to travel or call

• Ongoing engagement that supports follow-up rather than one-off interactions

For providers, it delivers:

• A unified dashboard showing all patients and their current status

• A chronological patient timeline that tracks symptoms, alerts, and interactions

• Risk signals surfaced from patient conversations to flag potential escalation

• Clear visibility into patient history to support continuity of care

• Summaries on missed calls,k through the AI-powered care companion..

By turning conversations into actionable data and organizing patient history into a single timeline, CareBridge AI addresses one specific breakdown in Clearwater Ridge’s system: missed follow-ups and lack of continuity when care happens outside the community or remotely.

How we built it

We built CareBridge AI as a full-stack, codable prototype within the hackathon timeframe.

• Frontend: React 19 with TypeScript and Vite for a fast, responsive UI

  • Tailwind CSS for rapid layout iteration and accessibility
  • Zustand for lightweight state management
  • React Router for role-based navigation (patient vs provider)
  • Backend: FastAPI serving REST endpoints
  • Supabase as the database with Row Level Security and role-based access control
  • Secure authentication and session management
  • Timeline event storage for structured patient history

• AI Layer:

  • Cohere for conversational analysis and health summary generation
  • ElevenLabs for voice-based check-in capabilities
  • Keyword-based and AI-assisted risk assessment to surface alerts

Challenges we ran into

Designing for trust and privacy: In small communities, privacy concerns can prevent people from engaging with mental health or symptom reporting tools. We had to think carefully about how information is surfaced and who sees what.

Data modeling timelines: Turning unstructured conversations into clean, chronological patient timelines required careful schema design.

Balancing scope: Clearwater Ridge has many breakdowns—transportation, referrals, mental health, staffing—but we had to focus on one core coordination failure to build something meaningful within the time limit.

Async data and UI layout issues: Handling dynamic timelines without breaking layout (especially horizontal overflow) required careful UI constraints and iteration.

Accomplishments that we're proud of

Building a fully functional end-to-end prototype within 38 hours

Creating a provider dashboard that clearly visualizes patient history and risk

Successfully integrating AI-generated summaries into a usable clinical workflow

Designing a system that works for both patients and providers, rather than optimizing for only one side

Aligning the solution tightly with real-world rural healthcare constraints instead of idealized urban systems

What we learned

Many healthcare failures aren’t caused by lack of treatment options, but by lack of follow-through

Rural healthcare solutions must be resilient to disruptions like weather, travel, and staffing gaps

AI is most useful when it supports continuity and triage—not when it replaces human care

Good care coordination tools need to reduce cognitive load for providers, not add to it

Small UX details (timelines, alerts, visibility) can have outsized impact on patient outcomes

What's next for Care Bridge AI

Next steps for CareBridge AI would include:

Closed-loop referral tracking with automated reminders and escalation

Transportation coordination integration for ride scheduling and cancellations

More advanced risk stratification based on longitudinal patient data

Expanded mental health pathways with privacy-first self-screening

Offline and low-bandwidth support for remote or disconnected environments

Pilot testing with real rural health workers to validate workflows and impact

Our long-term goal is to help communities like Clearwater Ridge prevent avoidable emergencies by ensuring no patient, referral, or follow-up quietly falls through the cracks.

Built With

• cohere
• elevenlabs
• fastapi
• react
• supabase
• vite