Digital Data Flow, or DDF, is an initiative for machine-readable reuse of protocol information. Its associated Unified Study Definitions Model, USDM, is a reference architecture for the structured representation of a study definition. Instead of repeatedly transferring content from a document into operational systems, objectives, endpoints, activities, and other protocol elements should be passed on from a digital source.
The architecture thus makes relationships between protocol components systematically available for downstream technical processes.
From Document to Digital Study Definition
In a document-based workflow, information from the clinical protocol is often reinterpreted and re-entered for EDC, CTMS, randomization, laboratory orders, or other systems. DDF aims to reduce these multiple transfers through a common, structured data flow. The central idea is to capture a study definition once and reuse its content in various downstream applications.
USDM describes the concepts relevant to a study in a model. Structured elements include, for example, study objectives, endpoints, planned activities, and biomedical concepts. The model can also accommodate structured and unstructured protocol content. It thus provides a technical reference framework for systems that exchange study definitions or derive study-specific configurations from them.
Components of the USDM Reference Architecture
The CDISC and TransCelerate work includes a logical model, controlled terminology, an API specification, implementation guidance, and examples. The logical model provides the classes and relationships of the study definition. Controlled terminology specifies the meaning of attributes and permissible value ranges. Via the API specification, a single study can be exchanged between machines in JSON format.
A reference architecture is not a finished product solution. It provides common vocabulary, reusable designs, and guardrails against which concrete system architectures can be aligned. The goal is interoperability between multiple systems within a study, not the specification of a particular EDC or authoring product. Alignment with the ICH M11 template connects the technical representation with a harmonized protocol structure.
The benefit of a structured study definition depends heavily on the same terms being understood identically in receiving systems. USDM therefore combines the logical model not only with a graphical class structure, but also with controlled terminology and an interface description. A receiving system can thus not merely read text, but process specific attributes and relationships of study information.
The reference architecture also serves the development of conformant solutions rather than the direct implementation of a single overall system. A sponsor or service provider must still decide which interfaces, checks, and approval steps are required for their landscape. DDF does not shift responsibility for the substantive correctness of the protocol to the data model, but makes this responsibility more interoperable across system boundaries.
Distinction: Machine-Readable Version, Not the Protocol Itself
USDM is not the clinical protocol as an approval-ready document. The protocol remains the binding professional description of study objectives, methods, and procedures for investigators, ethics committees, and regulatory authorities. The model represents selected content in a form that software can read, check, and reuse.
DDF also does not replace the medical and regulatory authorship of a protocol. Changes to inclusion criteria, safety procedures, or the Schedule of Activities still require professional assessment and formal governance. However, the digital structure can improve consistency between the approved version and configurations in operational systems if a controlled change process ensures the update of all affected outputs.
Digitization of the protocol therefore begins with professionally unambiguous information, not with the mere conversion of a PDF file into a technical container.
Relevance for clinical trials
The practical benefit arises primarily at interfaces: A clearly modeled schedule can link EDC forms, appointment scheduling at investigational sites, laboratory processes, and evaluation timepoints with the same study. The prerequisite is that the study definition is sufficiently precise before configuration and that responsibilities for approval, versioning, and mapping are established. Without regulated change management, a machine-readable model can even propagate inconsistent system states more rapidly.
Full-service CROs such as Mediconomics support DDF-oriented projects by reviewing protocol structure, translating activities and endpoints into system requirements, coordinating between Clinical Operations and Data Management, and reconciling EDC configuration with the approved study definition. For amendments, they can systematically track the impact on CRF, monitoring plan, randomization, and data transfers.
Frequently Asked Questions (FAQ)
Is USDM a file format for the complete protocol?
USDM is a model with a reference architecture for study definitions. It supports structured and unstructured content, but is not equivalent to a single document format.
Does DDF replace the EDC system?
No. DDF is intended to make information from a study definition usable for various systems. The EDC system remains the operational environment for capturing and processing study data.
Why is ICH M11 mentioned in connection with USDM?
ICH M11 provides a harmonized template for clinical protocols. The USDM work aligns the technical representation of protocol information with this structure.
Regulatory References
- CDISC Digital Data Flow – describes the initiative for reuse of digital protocol information.
- CDISC Unified Study Definitions Model – provides the reference architecture for study definitions.
- ICH M11 Clinical electronic Structured Harmonised Protocol – defines the harmonized protocol template as a professional reference point.
- HL7 FHIR and Vulcan UDP – support the technical exchange of structured protocol content.
