Menu Back to Poster-Presentations-Details

P121: Adapting Real-World Data (RWD) into CDISC Submission Standards: Updates from a Case Study

Poster Presenter

James Browning

Director of Biostatistical Programming, Center for Observational Research
Amgen
United States

Objectives

To present challenges and solutions encountered when adapting Real-World Data (RWD) into CDISC standards for a planned drug submission of a retrospective observational study serving as an external control for a single-arm clinical trial (CT).

Method

This poster shares updates from a case study highlighting challenges and potential solutions when mapping RWD into CDISC standards to serve as an external control arm (ECA) for a CT.

Results

Standardizing study data derived from RWD sources into CDISC standards, notably the Study Data Tabulation Model (SDTM) and Analysis Data Model (ADaM), presents many challenges as they were designed for CTs. Implementing FDA’s expectations from the Final Guidance on Data Standards and Biological Product Submissions Containing Real-World Evidence (December 2023), Amgen documented instances where RWD did not align with CDISC standards and outlined potential methods of conforming these RWD to CDISC. The following key concepts highlight challenges and potential solutions taken during transformation and analysis: Inclusion/Exclusion (IE) Criteria Fulfillment: IE criteria fulfillment was not prespecified in the RWD, so IE fulfillment was evaluated in ADaM based on information collected in SDTM domains. This hindered the team from indicating in SDTM datasets which patients would serve in the ECA. The team reported all RWD in the appropriate SDTM domains but were unable to use some expected domains and variables. Line of Therapy (LOT) Implementation: A LOT algorithm used treatment and disease response information from ADaM datasets to assign lines of therapy and finalize IE fulfillment for the ECA. Explanations of the LOT algorithm implementation and the unique order of ADaM domain creation were described in the Supplemental Data Definition. ADaM ECA Integration: Data for the ECA came from ADaM datasets in the retrospective observational study. Those data were integrated into the CT ADaM datasets, and the integration details were documented in the Define-XML file and ADaM Reviewer’s Guide (ADRG) of the clinical study data package. Supplemental XPORT Files: XPORT files were proposed to communicate the intermediary rationale of how the large variety of unstructured RWD data values were mapped and transformed in SDTM and ADaM domains. These files were referenced in the Define-XML, Reviewer’s Guide, and in SAS programs to clarify when the transformations were applied.

Conclusion

RWD does not always fit perfectly within existing standards, but there are concepts and strategies that can be utilized to support data transformation, consistent with FDA’s final guidance (2023). Integral to this process is early and frequent communication with FDA to ensure alignment on solutions, in addition to providing transparency and documentation of methods applied during data transformation. The retrospective nature of observational studies can have many implications on the data standards including domain utilization, variable utilization, and the order of domain creation. Supplemental XPORT files may be useful to provide the appropriate level of traceability of the conformance methods used to transform the original RWD values that exist from unstructured documentation to the study data standards. By sharing our experience mapping RWD into CDISC standards, we hope to contribute to the existing knowledge base as sponsors address similar challenges, and industry groups and regulators develop best practices for conforming RWD to CDISC standards to support the enhancement of data submission standards for RWD.