SDTM Studio

From raw clinical data to submission-ready SDTM.

Upload your EDC exports and external data. SDTM Studio plans the domains, drafts the mapping and transformation workflows, generates the datasets and validates them — while your team reviews and approves every step.

Request a Demo Sign in →
  • SDTMIG
  • CDISC CT
  • Define-XML 2.1
  • CDISC CORE
  • Dataset-JSON
SDTM Studio transformation workflow for the DM domain with the SDTM output preview
How it works

One platform for the whole SDTM process.

Run each step on its own, or all of them at once with a dry run: one click from uploaded sources to a full set of deliverables, in a sandbox copy of the study.

  1. 01

    Upload sources

    EDC exports, external vendor data, EDC metadata and the blank CRF.

  2. 02

    AI plans & drafts

    AI proposes the domain inventory and drafts each domain's workflow and mapping rules.

  3. 03

    Review

    Mappers and reviewers check each proposal with live previews on the study data.

  4. 04

    Generate & validate

    Produce the SDTM datasets and check them with CDISC CORE conformance rules.

  5. 05

    Deliver

    Define-XML, annotated CRF, reviewer's guide, validation report and submission package.

Auto Mapping

AI drafts. People decide.

The AI reads the source metadata, the SDTM Implementation Guide and the controlled terminology to propose which domains a study needs and how each one is built. Every proposal is test-run on the study data before review, so reviewers see records and warnings, not just rules.

  • Domain inventory planned from sources and protocol
  • Controlled terminology value mapping
  • Choice of AI provider, including a local model
Auto Mapping: AI workflow proposals with dry-run results
Transformation workflows

Visual, reviewable, reproducible.

Each domain is a workflow of clear steps — filter, join, transpose, derive — ending in the SDTM variable mapping. Preview the output of any step, and export the program in Python, R or SAS for independent review.

  • Standard derivations: USUBJID, --SEQ, --DY, EPOCH, ISO 8601 dates
  • Supplemental qualifiers (SUPP--) and RELREC
  • SAS transport (XPT v5) or Dataset-JSON output
Transformation workflow with the SDTM DM output preview
Mapping rules

Every SDTM variable, one reviewable rule.

The mapper lists every variable of the domain in SDTMIG order with its label, type, core status and origin. Each rule is a single formula; reviewers accept AI drafts, confirm rows and see open issues as they work, and the specification round-trips to Excel.

  • Conditional rules and SUPP-- destinations
  • Accept AI drafts, confirm or reject row by row
  • Excel export and import of the specification
SDTM variable rules grid for the AE domain
Validation

Find issues before the agency does.

Validate the generated datasets with the CDISC CORE rules engine or Pinnacle 21-style rule sets, record an explanation for each issue and export the validation report.

  • Issues traced to datasets, rules and records
  • Rules that could not run are reported, not hidden
Validation results by severity, rule and message
Study overview

Know where every domain stands.

Track review progress, open issues and warnings per domain. Study-level roles separate administrators, mappers, reviewers and viewers.

Study overview with domain progress
Capabilities

Everything a submission needs.

Sources

SAS7BDAT, XPT, CSV, Excel, Parquet and ZIP; CDISC ODM 1.3 / 2.0; EDC metadata such as Rave ALS; blank CRF and protocol.

Standards

SDTM and SDTMIG, CDISC Controlled Terminology and the CDISC Library, versioned and locked per study.

Deliverables

SDTM datasets, Define-XML 2.1, annotated CRF, reviewer's guide (cSDRG), mapping specification, validation report and submission package.

Dry run

One click runs the whole process in a sandbox copy of the study, fixes common issues automatically and reports what still needs a decision.

Programs

The approved specification can be exported as Python, R or SAS programs for independent programming review and QC.

Audit trail

Every change to specifications, runs, approvals and AI requests is recorded with who, when and why.

Data Protection

Patient data stays with you.

The AI works from metadata, not from patient records. Transformations run on the platform as deterministic, reviewable programs.

  • No row-level data to AIOnly variable names, labels, formats and de-identified coded values are sent; identifiers and free text are withheld.
  • Every AI request checked and loggedA leak scan blocks requests that contain identifying values, and each request is kept for audit.
  • Your choice of modelUse a cloud AI provider or run a local model on your own infrastructure.
  • Role-based accessStudy-level administrators, mappers, reviewers and viewers; new accounts require approval.
Get started

See SDTM Studio on your own study.

Request a demo, or request access and an administrator will approve your account and set up a study workspace.

Request a Demo Request access →

AI proposals are drafts that qualified people review. Validation of the system for regulated use remains the responsibility of each organization.