Save on courses or Exam now
About Instructors
-
frankieblues- Instructor Rating - Reviews - Students - Courses
No biography available, please complete
📘 About This Course
Health data is only useful when someone can defend it. This
course walks the whole path — from patients who never existed to a running,
audited platform — and every claim on screen is backed by a command you watch
execute.
You build one system across ten modules. Synthetic patients are generated and
their provenance recorded. They become validated HL7 FHIR R4 resources on a
local HAPI FHIR server you install and health-check yourself. Imaging arrives
through a local Orthanc archive and a DICOM viewer, queried over DICOMweb.
A dbt and DuckDB warehouse turns all of it into a star schema, and nothing
reaches a dashboard until it passes a data contract. A WebRTC telehealth
sandbox, a constrained clinical NLP step, an honest model evaluation, a
hash-chained audit trail and a release-readiness review complete the picture.
Everything runs on a Linux-first local lab. Every component is
downloaded from its official source and installed on screen — no jumping to an
already-configured interface. Every lesson contains one successful end-to-end
workflow and one safe, controlled failure that is diagnosed from the evidence,
corrected minimally and retested in front of you.
The data is synthetic throughout. No real patients, no production
EHR, no PHI. Diagnostic AI is framed only as constrained decision support under
qualified clinician oversight — never autonomous diagnosis. Privacy, security
and regulatory work is engineered and evidenced, not asserted.
The capstone ties it together: a nine-stage pipeline you can run with one
command, scored against thirteen written acceptance criteria, rehearsed through
an incident drill, and torn down cleanly at the end.
This course is educational. It is not medical advice, legal advice, a
diagnosis, clinical validation, a HIPAA or GDPR compliance certification, nor
regulatory clearance to deploy in patient care.
What You'll Learn
Install and health-check a complete Linux-first health data lab — HAPI FHIR, Orthanc, PostgreSQL, DuckDB, dbt — from official sources, on screen
Generate legally safe synthetic patients with Synthea and record their provenance
Model FHIR R4 resources, bundles, profiles and validation, and work the REST API with search, paging and conditional update
Run DICOM and DICOMweb workflows (QIDO, WADO, STOW) and link imaging to clinical context without duplicating sensitive data
Build an idempotent ingestion pipeline and a dbt/DuckDB star schema with tests that run on every build
Enforce data contracts and quality gates so nothing is published until a human has reviewed what changed
Construct reproducible cohorts and responsible dashboards that suppress small cells and state their limits
Deploy a local WebRTC telehealth sandbox, remote monitoring pipelines, and accessible, equitable access paths
Frame diagnostic AI as constrained decision support, build a schema-constrained clinical NLP step, and evaluate it honestly on a locked test set
Apply HIPAA and GDPR controls in the lab: layered access, hash-chained audit, pseudonymisation, retention and a breach tabletop
Map medical-device, FDA, EU MDR and AI Act questions without promising clearance, and run a release-readiness review with named owners
Prove where every analytical result came from, and leave no unnecessary resources running
🛠️ Requirements
- Comfort on the Linux command line and with Git
- Working knowledge of Python and SQL
- A machine able to run Docker Engine with Compose (16 GB RAM recommended)
- No clinical background required — the healthcare concepts are taught from scratch
- No real patient data of any kind; the entire course runs on synthetic fixtures
👤 Who This Course is For
- Data and analytics engineers moving into healthcare
- Health-tech and digital-health engineers who need interoperability depth
- Product, security and compliance professionals who must evaluate health data systems
- Clinical informaticians who want a reproducible, evidence-led lab
- Anyone responsible for defending a health data platform to a reviewer
53618 Students applied
100 Comments
-
-
John DoeMorbi nec nisi ante. Quisque lacus ligula, iaculis in elit et, interdum semper quam. Fusce in interdum tortor. Ut sollicitudin lectus dolor eget imperdiet libero pulvinar sit amet
-
John Doe
Morbi nec nisi ante. Quisque lacus ligula, iaculis in elit et, interdum semper quam. Fusce in interdum tortor. Ut sollicitudin lectus dolor eget imperdiet libero pulvinar sit amet
-
-
John Doe
Morbi nec nisi ante. Quisque lacus ligula, iaculis in elit et, interdum semper quam. Fusce in interdum tortor. Ut sollicitudin lectus dolor eget imperdiet libero pulvinar sit amet
-
John Doe
Morbi nec nisi ante. Quisque lacus ligula, iaculis in elit et, interdum semper quam. Fusce in interdum tortor. Ut sollicitudin lectus dolor eget imperdiet libero pulvinar sit amet
-
John Doe
Morbi nec nisi ante. Quisque lacus ligula, iaculis in elit et, interdum semper quam. Fusce in interdum tortor. Ut sollicitudin lectus dolor eget imperdiet libero pulvinar sit amet
Morbi nec nisi ante. Quisque lacus ligula, iaculis in elit et, interdum semper quam. Fusce in interdum tortor. Ut sollicitudin lectus dolor eget imperdiet libero pulvinar sit amet