Study Data Tabulation Model (SDTM)

The Study Data Tabulation Model (SDTM) is one of the foundational data standards published by CDISC (Clinical Data Interchange Standards Consortium). It defines how clinical trial data must be structured and named once a study is complete, so that the resulting datasets can be submitted to a regulatory authority — principally the U.S. FDA and Japan’s PMDA — in a consistent, machine-readable tabulation format. SDTM is a data structure standard: it specifies how information already collected during a trial should be organized for submission, not how that information is collected in the first place.

What Makes a Dataset “SDTM” (Operational Definition)

A dataset is SDTM-conformant when it satisfies three things:

• General observation class. Every SDTM domain is built on one of a small number of general observation classes — Interventions (what was administered to the subject, e.g. study treatment or concomitant medications), Events (things that happened to the subject, e.g. adverse events, medical history), and Findings / Findings About (observations or measurements made about the subject, e.g. lab results, vital signs, ECGs). Deciding which class a given piece of collected data belongs to is the key modeling decision in building an SDTM dataset.
• Standard domains and controlled terminology. Each topic area is represented as a two-letter domain with a standard variable naming convention — for example DM (Demographics), AE (Adverse Events), VS (Vital Signs), CM (Concomitant Medications), and LB (Laboratory Test Results). Most submissions are built against the SDTM Implementation Guide (SDTMIG), which applies the abstract SDTM model to these concrete, ready-to-use domain specifications; CDISC versions the model and its implementation guide together, and each SDTMIG release references the specific SDTM model version it implements.
• Accompanying metadata. A submitted SDTM dataset is packaged with a define.xml file that documents its metadata — variable definitions, controlled terminology, and derivations — so a reviewer can trace exactly how each value was produced.

Why It Exists: Regulatory Submission

FDA’s Study Data Technical Conformance Guide specifies SDTM (along with ADaM for analysis-ready data and a define.xml metadata file) as the required format for standardized study data submitted to CDER and CBER. PMDA in Japan has adopted an equivalent requirement. Before SDTM existed, every sponsor tabulated trial data in its own layout, and regulatory reviewers had to relearn a new data structure with every submission. A common tabulation format lets reviewers, and their own automated review tools, work with data from any sponsor the same way.

Worked Examples

• Adverse event reporting. Every adverse event recorded across a trial is tabulated in the AE domain, one row per event per subject, with standardized variables for the reported term, its MedDRA-coded preferred term, severity, seriousness, start/end dates, and causality assessment — regardless of which site or country collected the underlying data.
• Demographics. The DM domain is the required top-level domain for every SDTM submission: one row per subject, with standard variables for age, sex, race, and treatment arm assignment, that every other domain dataset links back to via a common subject identifier.

Counter-Example

A site’s raw electronic case report form (eCRF) export — the row/column layout a study’s EDC system uses internally to store what was entered at the point of data entry — is not SDTM. It typically reflects how the data was collected (organized by visit or form), not the standardized domain/observation-class structure SDTM requires for submission. Converting a raw eCRF export into SDTM domains is itself a distinct, standards-driven step in the data management workflow, commonly referred to as SDTM mapping or transformation.

SDTM vs. CDASH vs. EDC

These three terms are frequently confused because they all sit in the same clinical-data workflow, but each governs a different stage:

• CDASH (Clinical Data Acquisition Standards Harmonization) is CDISC’s companion standard for the collection stage — it provides standardized data collection fields and case report form content designed to map cleanly forward into SDTM domains later. CDASH governs what you ask; SDTM governs how the answers are tabulated for submission.
• Electronic Data Capture (EDC) is the software system sites actually use to enter and store trial data during the study. EDC is a technology; SDTM and CDASH are data-content standards. An EDC system may be configured to collect data in CDASH-aligned fields, and its export is later transformed into SDTM domains for submission — but EDC itself defines neither the collection fields nor the submission structure.

Related Terms

See also Electronic Data Capture (EDC) and Good Clinical Data Management Practices (GCDMP). For the broader clinical data management workflow SDTM fits into, including CDMS, CTMS, and eTMF, see the CASRAI guides on clinical data management tools and data management planning for clinical trials.