Health Monitoring Dashboard

Challenge

A digital health startup had built an MVP in Flutter with a small team of contractors. The codebase had grown brittle — shared state was scattered, the GraphQL layer was undocumented, and adding new device integrations required touching code in five or six unrelated places. The clinical team was fielding manual exports to spreadsheets because the dashboard lacked the filtering and annotation tools clinicians actually needed. The client came to us needing a rebuild that their own team could maintain and extend.

Solution

We rebuilt the mobile application in React Native with TypeScript, using a well-structured GraphQL client backed by Apollo. The architecture separated device integration logic from presentation entirely — new wearable integrations could now be added by implementing a single typed interface, with the rest of the app unchanged.

We worked directly with two clinical leads throughout the project to understand the workflow. The result was a clinician mode and a patient mode in a single app, with role-based data visibility and a notification system that distinguished between advisory alerts and urgent escalations.

Technical Approach

The frontend was React Native with TypeScript, targeting iOS and Android from a single codebase. Apollo Client managed all data fetching and caching, with subscriptions used for the real-time vitals stream. We used Zustand for local UI state, keeping the global store minimal and predictable.

Data visualisation was built on Victory Native — interactive sparklines for trend views, and larger detail charts on drill-down screens. Threshold bands and annotation overlays were custom components built on top of Victory's SVG layer.

The GraphQL API was built by the client's backend team; we worked with them to co-design several schema additions needed for the clinician dashboard features, including annotation mutations and filtered subscription topics.

CI/CD ran on Expo EAS for builds, with automated tests using Jest and React Native Testing Library. We maintained a component library in Storybook so the client's team could extend the UI without breaking established patterns.

Results

• Clinical review time per patient reduced by 28% based on time-motion study conducted by the client post-launch
• Patient app retention at 14 days improved from 41% to 67% following the new onboarding flow and alert tuning
• Codebase handover completed in two sessions — the client's engineering lead described it as the cleanest React Native project they had inherited
• Four new wearable integrations shipped by the client's team in the three months after handover, with no Procverse involvement
• App store rating improved from 3.1 to 4.6 on iOS and 4.4 on Android within 60 days of launch