Skip to content
Browse docs

USDM/DDF

Learn how the Digital Data Flow initiative uses a shared model to make protocol information available across applications.

The protocol now has a common structure. Other applications also need to understand the decisions inside it: which activities are planned, when they happen and how they relate to the trial’s objectives.

Digital Data Flow (DDF) is the industry initiative. The Unified Study Definitions Model (USDM) is the shared model for representing trial information. TransCelerate and CDISC developed this foundation to support reuse between authoring tools and consuming systems.

Why a shared model matters

A schedule can be presented as a table in a protocol. A person can read its rows, columns and footnotes to work out what happens at each visit. Software needs those relationships represented explicitly.

USDM gives applications a common way to express the design. An activity has an identity. Its scheduled occurrences have timing and context. A trial version holds a particular definition of that design.

This is the practical meaning of a structured trial definition: trial decisions stored as identifiable information with relationships that another application can follow.

What USDM contains

The model covers objectives, endpoints, populations, interventions and trial design. This selected part of the model shows how a trial version contains a design with arms, epochs, activities and schedule timelines.

flowchart TD
    Study["Study"] -->|versions| Version["StudyVersion"]
    Version -->|studyDesigns| Design["InterventionalStudyDesign"]
    Design -->|arms| Arm["StudyArm"]
    Design -->|epochs| Epoch["StudyEpoch"]
    Design -->|activities| Activity["Activity"]
    Design -->|scheduleTimelines| Timeline["ScheduleTimeline"]

The names in the diagram are USDM model names. StudyVersion holds a version of the trial information. StudyArm describes an arm. StudyEpoch describes a period in the design. Activity describes planned work, while ScheduleTimeline organizes scheduled occurrences. USDM also supports observational designs.

Follow the assessment

Our endpoint is change from baseline in an assessment score at Week 12. The definition needs to identify the endpoint, the assessment and its planned timing.

The assessment can occur at baseline and again at Week 12. These are two scheduled occurrences of the same assessment. Keeping the assessment definition separate from its placement makes reuse possible without losing the timing of each occurrence.

The model and its exchange contract

USDM includes a logical model, supporting terminology and an API specification. An implementation guide explains its use. Conformance rules provide checks against the standard.

JSON is a machine-readable representation used for exchange. The useful part is the agreed meaning of the objects and relationships inside it. Applications can keep their own interfaces and internal databases while exchanging information through that contract.

M11 supplies the protocol structure introduced on the previous page. USDM supplies a connected representation that can support that content and other uses of the trial design. The next question is where applications obtain the same version of the definition. That is the role explored in SDR/MDR.