Back to packs

FDA CDISC SDTM Submission

FDA_CDISCfree

10 data-quality rules for FDA clinical-trial data submitted as CDISC SDTM datasets (within an eCTD submission). Covers required identifier variables, USUBJID cross-domain referential integrity, CDISC Controlled Terminology conformance, ISO 8601 dates, DOMAIN consistency, sequence uniqueness, and visit referential checks. Use it to catch SDTM conformance issues before a define.xml / Pinnacle 21 validation run.

20 rules 0 downloadsNo ratings yet
cdiscsdtmfdaectdclinical-trialspharma
0.0(0 ratings)

Sign in to rate this pack

Test this pack with your data

Download the template, fill in your data, and see quality results instantly.

Test This Pack

Download & Install

Choose your tool — get a ready-to-run file

Run this on your data? Upload your CSV — we'll auto-map the columns, validate, and report the bad rows.Test my dataWant this certified on your data? Book a 30-min demo.Book a demo
Or use the CLI
$ npx dqhub install fda-cdisc-sdtm --format soda --table YOUR_TABLE

Reference data required

This contract's checks join against 3 reference tables. Load them into your warehouse — DQHub ships the public ones, the rest link to their source. All reference data →

Customer-loaded — source

About this pack

10 data-quality rules for FDA clinical-trial data submitted as CDISC SDTM datasets (within an eCTD submission). Covers required identifier variables, USUBJID cross-domain referential integrity, CDISC Controlled Terminology conformance, ISO 8601 dates, DOMAIN consistency, sequence uniqueness, and visit referential checks. Use it to catch SDTM conformance issues before a define.xml / Pinnacle 21 validation run.

What's included

7referential integrity rules
6format rules
3consistency rules
2completeness rules
2uniqueness rules

Checks included (20)

USUBJID Exists in DM

Every subject referenced in a domain (e.g. AE, EX, LB) must exist in the Demographics (DM) domain.

Coded Values Conform to CDISC Controlled Terminology

Variables governed by CDISC Controlled Terminology (e.g. AESEV, AEOUT, units) must contain values from the applicable CT codelist.

VISITNUM Exists in Trial Visits (TV)

Each VISITNUM used in a domain must be defined in the Trial Visits (TV) dataset.

AE Coded in MedDRA

Adverse-event dictionary-derived terms (AE.AEDECOD / AEBODSYS) must be valid MedDRA terms for the study's MedDRA version.

ARMCD Exists in Trial Arms (TA)

DM.ARMCD must be defined in the Trial Arms (TA) dataset.

Result Units Conform to CT

Standardized result units (--STRESU) must be values from the CDISC Unit codelist.

SUPP-- Records Link to a Parent

Each Supplemental Qualifier (SUPP--) record's RDOMAIN + USUBJID + IDVAR/IDVARVAL must reference an existing parent-domain record.

--DTC Dates Are ISO 8601(aestdtc)

SDTM date/time variables (--DTC) must be in ISO 8601 format (YYYY-MM-DD or with a time component).

DM.SEX Conforms to Sex Codelist(sex)

DM.SEX must be a value from the CDISC Sex codelist (M, F, U, UNDIFFERENTIATED).

AE Severity Conforms to CT(aesev)

AE.AESEV must be MILD, MODERATE, or SEVERE per CDISC Controlled Terminology.

AE Outcome Conforms to CT(aeout)

AE.AEOUT must be a value from the outcome CT (RECOVERED/RESOLVED, RECOVERING/RESOLVING, NOT RECOVERED/NOT RESOLVED, RECOVERED/RESOLVED WITH SEQUELAE, FATAL, UNKNOWN).

Y/N Variables Are Valid(aepresp)

Variables governed by the (N)Y codelist (e.g. --PRESP, --OCCUR) must be Y or N.

Baseline Flag Is Valid(blfl)

The baseline flag (--BLFL) must be 'Y' or null (never 'N').

DOMAIN Value Matches the Dataset

The DOMAIN variable must equal the two-character domain code of the dataset it appears in.

STUDYID Is Consistent

All records in a submission should carry the same STUDYID.

Start Date Not After End Date

For interval events/interventions, --STDTC must not be after --ENDTC.

Required Identifier Variables Present

Every SDTM record must carry STUDYID, DOMAIN, and USUBJID — the core identifier variables required by the SDTM IG.

USUBJID Present(usubjid)

USUBJID (unique subject identifier) must be populated on every record.

--SEQ Is Unique Within Subject(usubjid_seq)

The sequence number (--SEQ) must be unique within each subject in a domain.

No Duplicate Records (Natural Key)(natural_key)

Each domain record must be unique on its natural key (STUDYID + USUBJID + domain-specific identifying variables).