Integrating DTA metadata into DDS: Scope, touchpoints, and E2E automation

[chowsanthony]

Core: What metadata must exist to support technology-independent automation from source tranmitted data to SDTM and define.xml?

Objective

Explore how DTA metadata and DDS should work together to support an automated data flow from study metadata to SDTM deliverables while maintaining clear separation of concerns between standards and implementation technologies.

Key questions

1. What metadata should flow from DTA into DDS?

Should:

• A: DDS contain all DTA entities and attributes?
• B: DDS only contain operational metadata required for downstream deliverables and automation?
• C: DTA remain a standalone specification with well-defined touchpoints into DDS?

2. What is the minimum metadata needed to support SDTM generation?

Given that:

• DTA comes with BCs.
• BC metadata is carried from study USDM.
• SDTM specializations are available.

Can we identify the minimum set of metadata required to support:

• Source-to-target mappings
• SDTM dataset generation
• Define.xml generation

3. Where should transformation metadata live?

Given an automated data flow supports:

flowchart LR
    A[Raw Data] --> B[DTA Metadata]
    B --> C[Transformation Specification]
    C --> D[SDTM]
    D --> E[Define.xml]

Loading

Where should transformation rules be represented?

• DTA?
• DDS?
• Separate model?
• Shared model?

4. What should DDS represent beyond data?

A question raised during the meeting was whether DDS should also capture (or, recapture) business and operational context associated with data transfer agreements, not only the technical data structure. Examples:

• Vendor information
• Data source information
• Transfer context
• Other operational metadata

Should DDS support these concepts, or should they remain external references?

[chowsanthony]

A visual representation about the DTA model created by @rchaitanyapradeep is attached, here:
dta-explorer.zip