Clunker Busters
A connected platform for customers, inspectors, operations, reporting and payments.
Case study · Healthcare scheduling
A FHIR-compliant scheduling experience letting patients book and manage appointments for themselves and their dependents, built across a five-party delivery.
Production system
01The business situation
North East Medical Services serves medically underserved populations across the San Francisco Bay Area. Its board set out to improve the patient experience by making scheduling something a patient could do themselves: for their own appointments and for those of the dependents in their care, against a FHIR repository that had to remain the system of record throughout.
Core challenge
Deliver a self-service scheduling experience on top of a clinical FHIR repository, across five organisations, without ever becoming a second source of truth.
Before
Scheduling mediated by staff, records held apart
After
Patients scheduling directly against the clinical record
NEMS Scheduling Platform
02The system at a glance
03Workflow 01
A guardian holds appointments for dependents as well as their own. The platform treats that as the normal case rather than an edge case, so one login manages a household.
assets/portfolio/nems/workflow-booking.png · 16 / 11
04Workflow 02
Existing patients log in with the medical record number they already have. New patients register and receive a QR code that makes the next visit quicker. Check-in itself is geo-fenced, so it happens when the patient is actually at the clinic.
1Log in with MRN
assets/portfolio/nems/step-login.png · 1 / 2.04
2Verify coverage
assets/portfolio/nems/step-insurance.png · 1 / 2.04
3Check in on arrival
assets/portfolio/nems/step-checkin.png · 1 / 2.04
05Workflow 03
Quantfarm led design and development for NEMS, collaborating with us for engineering and with Apperbuilder for design; Smile CDR hosted the FHIR repository. What held five organisations together was a shared contract at the API boundary and a test-first route to production.
assets/portfolio/nems/architecture.png · 16 / 10
assets/portfolio/nems/registration.png · 1 / 2.04
06Engineering behind the product
The engineering problem here was less about the interface than about integrating cleanly with a clinical record standard owned by another team.
The product was decomposed into services so parts of it could evolve and deploy independently of one another.
Appointments, patients and coverage are read from and written to the FHIR repository, which stays the system of record throughout.
Existing patients authenticate with the medical record number they already hold, avoiding a parallel identity system.
A guardian and their dependents are modelled explicitly, which is what allows one account to manage several patients.
Check-in is gated on the patient being at the clinic, so the record reflects arrival rather than intent.
Coverage is checked through Experian eligibility verification as part of the booking flow.
Production ran in containers on one cloud while staging ran on another, deliberately proving the deployment was not tied to a single provider.
Redis and Celery handle queued and scheduled work, and uvicorn serves the application so it can hold the websocket connections the product needs.
One-time passwords and transactional SMS run through Twilio, with email over SMTP.
07Technical architecture
A microservice product integrating a FHIR repository owned by another team, deployed across two clouds to prove portability.
Backend
Frontend
Async & serving
Cloud
Integrations
Delivery tooling
08Delivery journey
Gather scope with the clinical team
Stand up the repository and resources
Mock up and approve the patient experience
Document and agree expected behaviour
Build services and integrate the APIs
Acceptance testing before release
Release and support Phase 1
Phase 1 scope is described above. Engagement dates, durations and adoption figures are withheld until confirmed. Verify before publishing
09What the system enabled