MICROSIVE
← All work
MobileHealthcare2025

Midori Health

Built an offline-first care tracker used by 800 clinicians.

Midori Health provides community healthcare programs in semi-urban Maharashtra. Their clinicians — mostly ASHA workers and field nurses — were tracking patient visits, vitals, and medication adherence in paper registers. The paper registers were being digitised weeks later by a data entry team. By then, the data was stale.

↑ 3.4×

daily active use

800

clinicians on day one

< 2s

sync time on reconnect


01 / Problem

No internet. No data. No insight.

The core constraint: many field sites have intermittent or no connectivity. Any solution that required a live connection would fail in the field. The previous app (a third-party mHealth tool) had been tried and abandoned — clinicians stopped using it after the first month because it failed silently when offline. Trust in the tool was zero. We were rebuilding that trust, not just the software.


02 / Solution

Works with no signal. Syncs when it gets one.

Offline-first wasn't a feature — it was the architecture. We used WatermelonDB for local SQLite storage with a custom sync engine that resolves conflicts intelligently (last-write-wins for most fields, merge for append-only fields like notes). The UI was designed with Midori's field team: large tap targets, minimal text input, and a patient-lookup flow that works in under 4 taps from home screen. Training was built into the onboarding — 15 minutes, no manual.

  • WatermelonDB offline-first architecture — works with zero connectivity
  • Custom sync engine with conflict resolution for concurrent edits
  • Patient lookup, vitals entry, medication adherence tracking
  • Visit scheduling and reminder system (local notifications)
  • Supervisor dashboard (web) for programme managers
  • Onboarding flow: 15-minute self-guided setup, no trainer required
  • React Native + Expo — single codebase for Android (primary) and iOS

03 / Result

From paper registers to live data in 11 weeks.

The first cohort of 200 clinicians went live in week 8 of build. Nisha told us the adoption curve was unlike anything they'd seen with previous tools — clinicians were showing the app to each other without being asked. By month 2, daily active use was 3.4× higher than their best previous tool had achieved in its entire lifetime. The paper registers were formally discontinued in month 3.

  • 3.4× increase in daily active use vs. previous tool
  • 800 clinicians onboarded across two cohorts
  • Paper registers discontinued in month 3 of deployment
  • Data latency: weeks → under 2 seconds on reconnect
  • Sync conflict rate: < 0.1% of records in first 90 days

04 / Deliverables

What they walked away with.

  • React Native mobile app (Android + iOS)
  • Offline-first sync engine (WatermelonDB)
  • Supervisor web dashboard (Next.js)
  • Push notification system
  • Onboarding + training flow
  • GitHub repo (theirs), full documentation
  • Notion handover + sync architecture guide
  • 30-day post-launch support

05 / Stack

The tools we used.

React NativeExpoWatermelonDBSQLiteNode.jsPostgreSQLPrismaAWSPostHog

Weekly Friday demos changed how our team thinks about shipping. We've kept the ritual. Oh, and the app is great.

NK

Nisha K.

CPO, Midori Health

Next case study

↓ 96% close time

Coast FinTech

Rebuilt their reconciliation engine. Monthly close: 4 days → 4 hours.

Want results like these?

Tell us what you're building.

Start a project →