CareBridge Navigator

Offline-first, non-diagnostic care navigation for community health workers that turns household intake into prioritized, SDG 3-aligned next steps and resource connections.

Comment

Inspiration

Frontline community health workers in low-resource environments are the backbone of primary care. However, they face a double challenge: they have no offline tools to prioritize household triage, and they spend more time writing paper reports than helping people. We built CareBridge Navigator to bridge this gap. Guided by UN SDG 3 (Good Health and Well-being), we wanted to build a practical, safe, and privacy-respecting care-navigation assistant that works anywhere—even without internet access.

What it does

CareBridge Navigator is an offline-first tool that converts simple household intakes into priority-coded navigation plans and local service connections:

  • Triage & Safety: Instantly assesses priority levels (emergency, same-day, follow-up, or routine) without making clinical diagnoses.
  • Frontline Action Cards: Generates copyable or printable handoff instructions with translation notes for local languages.
  • Supervisor Dashboard: Aggregates community outreach tasks, highlights missing service tags in the local directory (Resource Gaps), and identifies key access barriers.
  • Offline Directory Editor: Allows clinic coordinators to edit and export local resources (clinics, transport, food banks) via JSON without needing databases or backends.

How we built it

To make the codebase 100% readable by AI documentation engines, we focused on modular domain architecture:

  • State & Local-First: Built with RxDB and Dexie.js for reactive local storage, keeping patient data 100% client-side for maximum privacy compliance.
  • PWA Capabilities: Custom Service Worker caching for seamless offline load.
  • Verification Gates: Standardized scripts checking accessibility (Aria-live, focus rings), code safety (zero external APIs, zero hardcoded secrets), and Vitest suite (39 automated unit tests passed).

Challenges we faced

Integrating a complex local-first database (RxDB) with reactive hooks in Vite React under strict offline-first requirements without downloading external web fonts was tough. Resolving peer dependencies for deep PWA routing in a sandboxed, zero-internet environment tested our limits.

What we learned

We learned that the best way to support SDG 3 is not by replacing clinicians with black-box AI chatbots, but by building robust, auditable, offline-friendly software that empowers existing community health workforces.

Presentation

You can view our interactive project presentation generated via PresentMeApp here: Interactive Presentation

Built With

Share this project:

Updates