Updates on CDISC Standards
Validation
NJ CDISC User Group
September 19, 2013
Topics
 CDISC standards validation initiative
 FDA update on SEND checks
 OpenCDISC v1.4.1 release
 OpenCDISC plans
2
CDISC validation projects
 Past
 ADaM validation checks
 Define.xml tool (schema only)
 CAB project
 Current plan (CDISC Intrachange 2013)
 Each CDISC development team will create a
comprehensive list of business rules as a supplement to
their standards
 Rationale
 Standards developers are the best experts in defining
standards compliance requirements
3
CDISC ADaM Validation Checks
 First v1.0 was introduced in 2010-10
 Current v1.2 was released on 2012-07
 v1.2 includes 244 checks based on ADaM IG text
 ADSL and BDS data only
 Today the team is planning 2 iterations
 update existing checks and create new checks for ADAE,
ADTTE data
 update checks after a new release of ADaM IG
4
Define.xml
 XML Schema Validation for Define.xml v1.0, 2009-11
 Validator for Define.xml schema, 2009-12
 Define.xml v2.0 includes many business rules
embedded in standard specifications
 Today the team is working on detailed list of validation
rules
 Executable specification will be published as
Schematron
5
SDTM and SEND
 CAB/VP project, 2010
 Evaluated compliance rules for SDTM/define.xml
 Reviewed checks across different vendors
 Current plan:
 Create comprehensive list of business rules as a
supplement to new standards. E.g., SDTM v3.1.5, TA
standards, etc.
 Challenges:



Extra resources are needed
Synchronization across teams
Exact project scope and format are still under discussion
6
FDA’s validation rules for SEND data
http://www.fda.gov/downloads/ForIndustry/DataStandards/StudyDataStandards/UCM367129.xlsx
FDA Rule
number
MESSAGE
DESCRIPTION
DOMAINS
SEVERITY
Invalid variable name. Variable
Only variable names listed in the SENDIG
NONCLIN- names in the domain datasets
Version 3.0 should be used in a domain
0006
should be compliant with the SEND
dataset.
standard
ALL
Error
Invalid format for TEDUR. The
The variable values for all timing variables
NONCLIN- variable value values must conform
must conform to the ISO 8601
0007
to the ISO 8601 international
international standard
standard
TE
Error
Invalid format for RFSTDTC. The
The variable values for all timing variables
NONCLIN- variable value values must conform
must conform to the ISO 8601
0008
to the ISO 8601 international
international standard
standard
DM
Error
Invalid format for RFENDTC. The
The variable values for all timing variables
NONCLIN- variable value values must conform
must conform to the ISO 8601
0009
to the ISO 8601 international
international standard
standard
7
DM
Error
Off-Topic: FDA’s Position Statement
Study Data Standards for Regulatory Submissions
FDA recognizes the investment made by sponsors over the past decade to develop the expertise and
infrastructure to utilize Clinical Data Interchange Standards Consortium (CDISC)[1] standards for study
data. The submission of standardized study data enhances a reviewer’s ability to more fully understand
and characterize the efficacy and safety of a medical product.
The Prescription Drug User Fee Act (PDUFA V)[2] Performance Goals state that FDA will develop
guidance for industry on the use of CDISC data standards for the electronic submission of study data in
applications. In the near future, FDA will publish guidance that will require study data in
conformance to CDISC standards.[3]
FDA envisions a semantically interoperable and sustainable submission environment that
serves both regulated clinical research and health care. To this end, FDA will continue to
research and evaluate, with its stakeholders, potential new approaches to current and emerging
data standards. FDA does not foresee the replacement of CDISC standards for study data and will
not implement new approaches without public input on the cost and utility of those
approaches.
September 13, 2013
http://www.fda.gov/ForIndustry/DataStandards/StudyDataStandards/ucm368613.htm
8
OpenCDISC v1.4.1 release, 2013-09-12
 Bug fixes, enhancements, but no global changes
 Support for QS Terminology
 CDISC CT checks for extensible codelists are
Informational and independent from user’s codelists
in Define.xml
 SDTM 3.1.1 with all applicable checks
 Simplified, consistent and concise error messages and
rule descriptions
 Excel report includes versions of Validator, CT, and
MedDRA
 Enhancement to check definition
syntax
9
SDTM Terminology
 New CDISC Control Terminology checks based on CT
rather than on outdated SDTM IG assignments
 PKUNIT instead of UNIT in PC and PP domains
 PKPARM, SCTEST, etc.
 QS Terminology (CT0100-CT0255 checks)
 CT 1:1 consistency checks for paired variables
 TSPARM/TSPARMCD, EGTEST/EGTESTCD, etc.
 Terminology checks on non-extensible codelists were
changed to Errors

CT0001 (ACN), CT0002 (AESEV), CT0030 (RELTYPE),
CT0059 (NY), etc.
10
Examples of SDTM checks “tuning”
 Rule SD0006 (“No baseline results”) was modified to




exclude NOTASSGN subjects
Rules SD0026 and SD0029 (“units are required”) were
modified to exclude PH, SPGRAV and all Antibody lab tests
Rule SD0037 (data vs. user’s codelist) was modified to
exclude EGORRES variable
Rule SD0069 (“No Disposition record for subject”) was
modified to exclude SCRNFAIL and NOTASSGH subjects
Rule SD1037 (“Missing value for --TOX, when --TOXGR is
populated”) was modified to ignore records with TOXGR=0
11
SEND specific
 Fixed incorrect variable labels and order of
variables in domains
 Several Check ID replacement for consistency with
SDTM checks
 CT1003-> CT0088, CT1004->CT0089, etc.
12
ADaM
 Checks compliant with ADaM Validation Checks 1.2
 Change all consistency rules to only run when primary
variables are populated
 Remove AD1002 - Expected variable is not present
within dataset. This is not a real rule. It was copied
from SDTM logic
 Changes to improve label validation (AD0018)
 Addition of other rules interpreted from IG, which
were not released by CDISC team
13
OpenCDISC v2.0
 Date of release is planned for 2013Q4
 Support for Define.xml v2.0
 Metadata driven
 Specifications based on Define.xml v2.0 format
 Validation organized by functional area. E.g.,
- SDTM Compliance
- Control Terminology
- Metadata
- Regulatory Specific
- Data Quality
14
Example of metadata driven checks
 SD1117 “Duplicate records” in Findings domain
 The “generic” check uses 10 Key Variables
USUBJID
ESTCD
--CAT
--SCAT
--METHOD
--SPEC
--DRVFL
--EVAL
VISITNUM
--TPTNUM
15
Key variables as defined in SDTM IG
IE
STUYID, USUBJID, IETESTCD
LB
QS
STUDYID, LBTESTCD, LBSPEC, VISITNUM, LBTPTREF, LBTPTNUM
STUDYID, USUBJID, QSCAT, QSTESTCD, VISITNUM, QSTPTREF,
QSTPTNUM
DA
STUDYID, USUBJID, DATESTCD, DADTC
FA
STUDYID, USUBJID, FAOBJ, VISITNUM, DATPTREF, FATPTNUM
TU
STUDYID, USUBJID, TULINKID, TUTESTCD, TUSTRESC, (TULOC,
TULAT, TUDIR, TUMETHOD), TUEVAL, VISITNUM
SDTM IG: “Note that the key variables shown in this table are examples only.
A sponsor‘s actual key structure may16be different.”
Combined Key variables
STUDYID
USUBJID
POOLID
--LNKID
--SPID
--TESTCD
--OBJ
--ORRES
--CAT
--SPEC
--LOC
--DIR
--LAT
--METHOD*
--DRVFL*
--ANTREG
VISITNUM
--DTC
--ENDTC
--TPTREF
--TPTNUM
17 data
* added by OpenCDISC as common cases in actual study
The check implementation options
 Generic check with 20 key variables?
 Separate check for each domain?
 v.2.0 approach
 Take domain key variables from Standard metadata
 If define.xml is available, then utilize key variables
provided in Study metadata
 Compare Standard and Study key variables
18
OpenCDISC Network
 More direct involvement in OpenCDISC project
development process
 Beta testing
 Requirements input
 Additional resources for new projects
 Better support
 Face-to-face meetings with users
 Webinars
 Training and sharing users experience
 First meeting is before CDISC Interchange 2013
19
References
 CDISC ADaM resources
http://www.cdisc.org/adam
 CDISC define.xml resources
http://www.cdisc.org/define-xml
 FDA Study Data Standards Resources
http://www.fda.gov/ForIndustry/DataStandards/StudyDataStandards/default.htm
 OpenCDISC
http://www.opencdisc.org
20
Questions/Feedback
Sergiy Sirichenko
Max Kanevsky
sergiy@opencdisc.org
Mike DiGiantomasso
mike@opencdisc.org
max@opencdisc.org
http://www.opencdisc.org/
21