Question 14 of 20 Intermediate Concept
What does Define-XML do in a regulatory submission?
Short answer
Define-XML is the metadata file that describes every submitted dataset: its variables, types, controlled terms, value level rules and where each value came from. FDA and Japan's PMDA require it for each study in an electronic submission, so reviewers can understand and trace SDTM, SEND and ADaM data.
Why
A reviewer opening an SDTM or ADaM dataset sees columns and codes. Without metadata they cannot tell how a variable was derived or which codelist applies. CDISC describes Define-XML as the standard that transmits this metadata, and states that FDA and PMDA require it for every study in each electronic submission.
It records dataset structure, variable labels and lengths, controlled terminology, value level metadata such as which test codes appear in a laboratory dataset, and origins: whether a value was collected on a case report form, derived, assigned or came from another source. For derived variables it carries the derivation method. FDA's Study Data Technical Conformance Guide, updated in October 2026, explains how Define-XML fits alongside the reviewer's guides.
A common failure is a define file generated at the end that no longer matches the datasets, for example a codelist missing values that appear in the data. Because Define-XML is produced from specifications, keeping specifications current during the study is what makes it accurate.
A strong answer covers
- Metadata for SDTM, SEND and ADaM datasets
- Required by FDA and PMDA for each study
- Origins and derivation methods
- Controlled terminology and value level metadata
- Must match the datasets actually submitted
Common mistakes
- Treating Define-XML as optional documentation.
- Generating it at the end without checking it against the final data.
- Confusing Define-XML with the reviewer's guide, which explains rather than describes.
The interviewer might ask next
How would you check that a Define-XML file matches the submitted datasets?