Interoperability project · FHIR R4
I built a working FHIR pipeline and put 2,325 patients on it.
Most healthcare informatics candidates can talk about interoperability. I can show you mine: a synthetic patient population generated with Synthea, expressed as FHIR R4 resources with US Core profiles, explored live in the viewer below. This is the same standard Epic and Cerner use to exchange data today.
The live viewer
Search the population, open a record, and see it rendered from real FHIR resources. Every section below maps to the FHIR REST query that produced it. The full 2,325-patient population lives behind this 334-patient sample.
Select a patient to open their FHIR record
Try searching a name, or pick anyone from the list.
How I built it
The problem is real. U.S. healthcare runs on disconnected systems, and the 21st Century Cures Act now requires hospitals and health IT developers to exchange a standard set of clinical data, the USCDI, using FHIR APIs, with penalties for information blocking. Yet most informatics candidates have never touched a FHIR resource. I decided to change that for myself.
The pipeline has four stages. First, Synthea generates a realistic synthetic population, complete longitudinal histories, conditions, medications, labs, encounters, immunizations. Second, everything is exported as FHIR R4 transaction bundles with US Core profiles applied, the same profiles EHR vendors certify against. Third, I query those resources the way any FHIR client would: Patient, Condition, MedicationRequest, Observation, Encounter, Immunization. Fourth, this page renders the results.
I also ran a SMART on FHIR app against the public HAPI FHIR test server, walking through both standalone launch and the full OAuth2 clinical launch flow in the SMART sandbox. That is the same launch sequence Epic uses for third-party apps.
What this taught me: reading raw FHIR JSON changed how I think about EHR data. A lab result is not a row in a table, it is an Observation with a LOINC code, a value, a timestamp, and a patient reference. Once you see data this way, mapping workflows, building reports, and troubleshooting interfaces all get clearer. That is the lens I bring to analyst work.
What the build covers
- FHIR R4 resources: Patient, Encounter, Condition, Observation, MedicationRequest, Procedure, Immunization, DiagnosticReport
- US Core profiles and USCDI data classes in practice, not just on paper
- SMART on FHIR launch flows: standalone and EHR launch via public sandbox
- Synthetic data generation at population scale with Synthea
- Static-site delivery of a clinical data viewer with zero hosting cost
Honest scope
This is an educational build on synthetic data. It is not connected to any real EHR, and it does not handle real patient information. What it demonstrates is genuine: I can generate, query, and present interoperable clinical data using the standards U.S. healthcare actually runs on.
Hiring for an EHR, Epic, or interoperability role?
Ask me to walk through this pipeline in an interview. I will show you the data, the queries, and the code behind this page.
Get in touch View the code on GitHub LinkedIn