The recommended learning path
This tutorial sequences every page on USDM Play into one path, in the order we recommend actually reading them. If you only bookmark one page on this site, make it this one.
Step 1 — Concepts: what USDM is and why it exists
- What is USDM? — the one-paragraph definition, the backbone, and what a study actually contains.
- Why was USDM created? — the DDF history, the ICH M11 connection, and the "write once, read many" motivation.
- USDM vs SDTM and USDM vs ODM — how USDM relates to standards you may already know.
- CDISC DDF Explained and CDISC 360 Explained — the two programs USDM sits inside.
Step 2 — The model: classes, UML, and relationships
- USDM Object Model — the ten model areas and standard attributes shared by every class.
- USDM UML Tutorial — how to actually read the official class diagram (cardinalities, inheritance, naming).
- USDM Relationships — Arm×Epoch=Cell, Cell→Element, Encounter↔Epoch, Timing.
- USDM Version 4.0 Explained — what's new, and what's stayed stable since v1.0.
Step 3 — Build with JSON and the API
- USDM Examples — three fully worked designs: parallel arms, a PK sub-timeline, and conditional branching.
- USDM JSON Examples — annotated payloads for every class you'll touch most.
- USDM API Tutorial — serialization rules, cross-references, and the extension mechanism.
Step 4 — Validate and apply
- USDM Validation — the CORE conformance rules, with real rule text.
- USDM Implementation Guide Explained — the official IG, section by section.
- USDM Use Cases — what real automation built on USDM looks like in production.
Prefer a curated course structure?
See USDM Training for how CDISC's own official curriculum sequences these same concepts across ten chapters, if you want to follow that structure instead.
How to know you're actually ready to build
Rather than a time estimate, use this checklist to know when you're ready to move from reading to building your own USDM JSON:
- You can draw the backbone (Study → StudyVersion → StudyDesign → Arms/Epochs/Timeline) from memory.
- You can explain the difference between a Study Cell and a Study Element without looking it up.
- You can read a cardinality bracket like
[0..*]or[1..*]correctly on sight. - You understand why a JSON field ending in
Idis a cross-reference, not embedded content. - You know where to check whether your content is CORE-conformant before assuming it's correct.
If any of those feel shaky, revisit the corresponding step above before moving forward — the model rewards a solid backbone understanding far more than it punishes a slow start.
Frequently asked questions
How long does it take to learn USDM?
Most learners with a clinical data background (SDTM, ODM, or protocol authoring experience) can understand the core backbone and Schedule of Activities modeling within a few hours using this tutorial path, and become comfortable reading/writing conformant JSON within a few days of hands-on practice.
Do I need to know UML to learn USDM?
No prior UML knowledge is required — this tutorial's UML lesson teaches exactly the notation USDM uses (cardinalities and inheritance arrows) from scratch, since that's all you need to read the official diagrams.
What should I learn first: JSON or UML?
Learn the UML backbone first. It's the normative model everything else derives from, so understanding it makes the JSON's nesting and cross-reference patterns immediately intuitive instead of arbitrary.
Is there an official CDISC training course?
Yes — see USDM Training for a walkthrough of CDISC's own on-demand curriculum structure, which this tutorial path is designed to complement, not replace.