01
HIPAA beyond the database
Problem
Compliance had to hold for every vendor, request and log line, not just stored records.
How we solved it
Vendor BAAs, encryption and an audit log designed for review.
Case study · 16Women's Health
OBGYNRx
A women's health app that fills the weeks between OBGYN visits, sending patient-reported symptoms straight into the clinic's records.
- Client
- OBGYNRx
- Industry
- Women's Health · Healthcare
- Platform
- iOS · Android · Web
- Timeline
- 20 weeks
- Outcome
- Patient-reported symptoms reaching clinic records between visits
- 20 wksFrom kickoff to launch
01
Overview
OBGYNRx gives patients a way to stay in contact with their OBGYN team between appointments: they record symptoms in their own words, follow their care plan, message the practice and receive lab results. We delivered the patient app, the clinician portal and the HIPAA-compliant backend that connects them.
02
The challenge
Long gaps separate one appointment from the next, and almost nothing a patient experienced in between was written down. Conditions like endometriosis and PCOS, and high-risk pregnancies, can't wait for the next visit to be discussed. Clinicians opened each appointment by reconstructing the last two months from memory. The hard parts were HIPAA, sync across devices and EMR integration.
03
Our approach
We spent our effort where it counted, compliance and symptom understanding, and chose proven, boring infrastructure everywhere else.
04
How we built it
4 steps, from architecture to launch.
- 01
Compliance as the first deliverable
Access rules, encryption and an audit trail were designed before features, with BAAs in place for infrastructure vendors.
- 02
Patient app and clinician portal
A Flutter app for iOS and Android and a Next.js portal, both backed by Supabase for the database, sign-in and file storage.
- 03
Understanding everyday descriptions
A medical language model, tuned on annotated patient descriptions, converts plain wording into structured clinical entries.
- 04
Writing back to the EMR
Records flow to the practice through HL7 FHIR, with a manual export for systems that can't connect.
05
Problem → How we solved itTechnical challenges
02
Patient language to clinical data
Problem
Patients write "it hurts on my right side when I sit," not clinical terms.
How we solved it
A tuned medical NLP model maps everyday descriptions to structured symptoms a clinician can scan.
03
FHIR first, export as fallback
Problem
Practices run many different EMR systems, which had defeated earlier attempts.
How we solved it
A FHIR-first connector covers the common systems, and a simple export handles the rest.
06
6 screensScreens
07
6 featuresWhat we shipped
- 01
Symptom journal
Log symptoms in plain words and build a timeline to share before or during a visit.
- 02
Care plan
Appointments, tests and medications from the clinician, with reminders.
- 03
Messages and results
Secure messaging with attachments inside the clinic's hours, and lab results with plain-language next steps.
- 04
Urgent flag
Marking an entry urgent pages the on-call clinician outside office hours.
- 05
Clinician portal
Each patient's activity since the last visit, written back to the EMR over FHIR.
- 06
More than one practice
Patients seeing both a high-risk pregnancy specialist and a local midwife link both practices to one account.
08
Tech stack
- Flutter
- iOS and Android patient app
- Next.js + TypeScript
- Clinician web portal
- Supabase
- Database, sign-in and file storage
- HL7 FHIR
- EMR integration
Tell us what you are shipping.
Start a Project →