Professional Documents
Culture Documents
Define-XML
CDISC Educational Materials
Intellectual Property Notice
Sharing, copying, recording or reselling any
CDISC educational material is strictly prohibited.
© CDISC 2012
Define-XML
© CDISC 2012
Topics
Introduction
Define-XML
XML essentials
Metadata representation
Dataset-XML
© CDISC 2012
Course Objectives
• Learn how Define-XML fits with other CDISC models.
• Learn how to represent SDTM and AdaM metadata for
datasets, variables, controlled terminology using
Define-XML.
• Become familiar with Define-XML 2.0 capabilities for
representing value or parameter level metadata.
• Learn how to provide traceability for SDTM and ADaM
data
• Learn about Analysis Results Metadata in XML
• Learn how Dataset-XML can be used in place of SAS-
XPT datasets
• Learn about strategies for resolving common Define-
XML implementation issues.
© CDISC 2012
Notes
© CDISC 2012
About the Trainer
• Primary CDISC affliation
• Current position/organization
• How your work relates to CDISC XML Technology
Standards
• Other qualification (years experience in related field)
© CDISC 2012
Participant Introductions
• Organization
• Job function/role
• XML background
9
© CDISC 2012
Topics
Introduction
Define-XML
XML essentials
Metadata representation
Dataset-XML
11
Protocol CDASH SDTM ADaM
Define-XML
SDM-XML
SDM-XML ODM
ODM Define-XML
Dataset-XML
Glossary
Therapeutic Areas
BRIDG
SHARE Questionnaires
© CDISC 2012
SEND SDTM ADaM SEND SDTM ADaM
Metadata Metadata Metadata Data Data Data
Define- Dataset-
XML XML
ODM
13
© CDISC 2012
Machine Readable Metadata
• What makes Define-XML ‘machine readable’?
XML can be easily read by a wide variety of application programs
Standard mapping of CDISC submission metadata into a
standard XML format enables content sharing
• How does machine readable metadata help data review?
Enables automated loading and validation of submissions
metadata by any application.
• Eliminates the need to retype dataset and variable names
Enables consistency checking between metadata and data
Encourages development of sophisticated submissions data
review tools
Enables reuse of submission data for research purposes
© CDISC 2012
Purpose of Define-XML
• US Food and Drug Administration (FDA) requires clinical
data for all clinical trials used as part the regulatory
approval process for new drugs and biologics.
• Reviewers need metadata that describes the content of
each submission.
• For SDTM submissions • For ADaM submissions
Datasets included • Dataset definitions
• Variable definitions
Dataset variables
• Derivations
Controlled Terminology • Value level metadata
Value level metadata • Controlled terminology
Traceability to the CRF • Traceability to SDTM
15
© CDISC 2012
Other Applications of Define-XML
• Sponsors use Define-XML to specify requirements
for datasets to be delivered by CROs or other 3rd
party data providers
© CDISC 2012
Define-XML features
• Platform independent
• Vendor neutral
• Open source
• Machine readable
• Easily transformed
using XSL and other
tools
Regulatory Requirements
Japan:
(PMDA Technical
Conformance Guide
(English version))
© CDISC 2012
Regulatory
Requirements
PMDA
DataCatalog_20150731ENG.xls
Regulatory
Requirements
FDA Guidance
Improvements in Define-XML 2.0
• Machine readable value level metadata
• Enhanced support for CDISC controlled terminology
• Improved support for documenting links to annotated
CRFs, ADaM Traceability metadata
• Better support for comments
• New elements, attributes, and data types
• Reorganized specification document
• More extensive examples
• Reduced platform dependencies in XSL style sheet
• Self documenting schema
• No longer need to create Define.pdf
© CDISC 2012
Status of Define-XML 2.0
• Final version posted February 2013
• Can be downloaded from the CDISC website
(http://www.cdisc.org/define-xml)
• Not backward compatible w Define-XML 1.0
• Validation requirements and Implementation
Guide development in progress
© CDISC 2012
Define-XML 2.0 - TOC
Download from the CDISC website http://www.cdisc.org/define-xml)
• Specification
• XML Schema
• Examples
ADaM
SDTM
Includes XSL stylesheet
© CDISC 2012
Topics
Introduction to Define-XML
Define-XML
XML essentials
Metadata representation
Dataset-XML
25
© CDISC 2012
Example
26
© CDISC 2012
<PlayList name=“MyPlaylist”
CreationDate=“2013-09-19”> Root Element
<Song PlayTime=”227">
<Title>Shattered</Title>
<Artist>Rolling Stones</Artist> Attributes
<Album Genre=“Rock”>
<Title>Some Girls (Deluxe Edition)</Title>
</Album>
</Song>
Elements
<Song PlayTime=”231">
<Title>Eet</Title>
<Artist>Regina Spektor</Artist>
<Album Genre=“Alternative”>
<Title>Far</Title>
</Album>
</Song>
…
</PlayList>
27
© CDISC 2012
XML Schemas
• Definition:
XML documents that describe structure and content for
a specific type of XML document
• Purpose:
Standardize contents of XML documents for a specific
use case.
• XML namespace declarations
• Define allowable elements and attributes
• Define allowable datatypes for element and
attributes
• Define relationships between elements and attributes
• Define enumerations
Support validation of XML documents
28
© CDISC 2012
Playlist Schema
Playlist Attributes name
CreationDate
Title
Artist
Title
© CDISC 2012
CDISC ODM Schema
MetaDataVersion
ODM Include
Protocol
Study StudyEventDef
FormDef
AdminData
ItemGroupDef
ReferenceData
ItemDef
ClinicalData CodeList
Presentation
ConditionDef
MethodDef
30
© CDISC 2012
Structure of Define-XML Metadata
MetaDataVersion
ItemGroupDef Datasets
def:CommentDef Comments
Xpath Description
/Playlist/Song/Album Any Album element
© CDISC 2012
XSL – Extensible Stylesheet Language
© CDISC 2012
Define-XML files Transformed using XSL
© CDISC 2012
Topics
Introduction
Define-XML
XML essentials
Metadata representation
Dataset-XML
© CDISC 2012
Datasets
© CDISC 2012
Dataset Metadata
• Every submitted tabulation and analysis dataset should be
described with metadata (i.e., data about the data):
Dataset name
Label - descriptive name
Class – SDTM, ADaM or SEND class information
Structure – organization of observation records
Purpose – Tabulation or Analysis
Keys
Location – folder and filename for dataset file
Documentation – annotated CRF, reviewers guide
© CDISC 2012 39
SDTM Dataset Metadata
© CDISC 2012
ADaM Dataset Metadata
© CDISC 2012
Dataset Metadata ItemGroupDef
@ OID
@ Name
attributes
@ Repeating
@ IsReferenceData
ItemGroupDef
@ SASDatasetName
ItemRef @ Domain
@ Origin
Alias @ Role
@ Purpose
def:leaf @ Comment
@ def:Class
Description
@ def:Structure
@ def:ArchiveLocationID
SDTM Dataset Metadata Example
1 2 3 5 7 8
4 6
1
3 4 5
2
1
2
3 4
© CDISC 2012
ADaM Dataset Metadata - Comments
<def:CommentDef OID="COM.ADQSADAS">
<Description>
<TranslatedText>See referenced dataset creation program and Analysis Data Reviewer's Guide, Section 2.1</TranslatedText>
</Description>
<def:DocumentRef leafID="LF.ADQSADAS.PGM"/>
<def:DocumentRef leafID="LF.ADRG">
<def:PDFPageRef PageRefs="Section2.1" Type="NamedDestination"/>
</def:DocumentRef>
</def:CommentDef>
© CDISC 2012
SDTM Split Datasets
• Required when individual dataset size exceeds
1GB (FDA Common Issues Document)
© CDISC 2012
SDTM Split Domain Datasets Example 1
3
1
1 4
1
ItemGroup Name = SASDatasetName = root file name
2
Domain = parent Domain
3
Description = Dataset description/label
4
Alias (@DomainDescription) = Parent domain label
1
<ItemGroupDef OID="IG.QSCG” 2 Domain="QS” Name="QSCG” Repeating="Yes” IsReferenceData=”N”1 SASDatasetName="QSCG"
Purpose="Tabulation” def:Structure="One record per questionnaire per question per visit per subject"
def:Class="FINDINGS” def:CommentOID="COM.DOMAIN.QSCG"
1 def:ArchiveLocationID="LF.QSCG">
3 <Description>
<TranslatedText xml:lang="en">Questionnaire-QSCG</TranslatedText>
</Description>
…
4 <Alias Context="DomainDescription" Name="Questionnaires"/>
<def:leaf ID="LF.QSCG" 1 xlink:href="qscg.xpt">
<def:title>qscg.xpt</def:title>
</def:leaf>
</ItemGroupDef>
Dataset Summary
• For regulatory submissions, a Domain
specification is required for SDTM (and SEND)
datasets
© CDISC 2012
Variables
© CDISC 2012
Variable-Level Metadata
Note the single asterisk for AEREL. The values for this variable are
subject to sponsor-defined controlled terminology
© CDISC 2012 53
Variable-Level Metadata
© CDISC 2012 54
SDTM Variable Metadata
© CDISC 2012
ADaM Variable Metadata
Variable Metadata ItemRef/ItemDef
ItemGroupDef
ItemRef
ItemDef attributes
Alias
Description @ OID
def:leaf
@ Name
CodeListRef
Description
@ DataType
def:Origin
@ Length
def:ValueList
@ SignificantDigits
@ SASFieldName
@ def:DisplayFormt
@ def:CommentOID
SDTM Variable Metadata
© CDISC 2012
SDTM Variables - ItemRef
1 2 3 4 5
1 3 4
<ItemDef OID="IT.STUDYID" Name="STUDYID" DataType="text" Length="7" SASFieldName="STUDYID">
<Description>
2
<TranslatedText xml:lang="en">Study Identifier</TranslatedText>
</Description> 5
<def:Origin Type="Protocol"/>
</ItemDef>
© CDISC 2012
ADaM Variables - ItemDef
5
1 2 3 4
1
<ItemDef OID="IT.ADSL.DISONSDT" Name="DISONSDT" SASFieldName="DISONSDT"
3 DataType="integer" Length="8" 4 def:DisplayFormat="date9.">
<Description> 2
<TranslatedText xml:lang="en">Date of Onset of Disease</TranslatedText>
</Description>
5<def:Origin Type="Derived"/>
</ItemDef>
Define-XML DataType Submission Length/Format
(example) DataType
text Char Maximum allowable length
(hematocrit)
integer Num Largest allowable integer width. Note
(160) that ADaM date variables are
provided as integers
float Num Use when value has a fractional
(57.5) component.
datetime Char ISO8601
(2013-10-02T10:30:25) YYYY-MM-DDTHH:MM:SS[.SS]
date Char ISO8601 YYYY-MM-DD
(2013-10-02)
time Char ISO8601 HH24:MM:SS[.SS]
(10:30:25)
partialDateTime Char ISO8601 YYYY[-MM[-
(2013-10-02T10:30) DD]]THH24[:MM[:SS[.SS]]]
incompleteDatetime Char [YYYY-]-[MM-]-[DD-
(10-02T10) ]T[HH24:][:MM:][SS.][SS]
durationDatetime Char ISO8610 [-]PHH:MM:SS
(-P02:15) © CDISC 2012
Translating Data Types
• SDTM • Define.xml
Num integer
float
Char text
Char (ISO8601) date
datetime
time
© CDISC 2012 63
Data Type Considerations
• ItemDef Length attribute is required when
DataType is text, integer or float
Note
• It is invalid to use the Length attribute with other
data types.
© CDISC 2012
Exercise 1
© CDISC 2012
Exercise 1 Instructions
1. Locate Exercises Folder in course materials
(USB)
2. Copy folder to desktop
3. Navigate to Exercises Folder and locate
Exercises spreadsheet.
4. Open spreadsheet using Excel
5. Save a copy of the worksheet to desktop
6. Click on TOC tab
7. Click on Datasets and Variables link.
Follow steps.
(optional) Use the Notes column to record any
questions and notes as you complete each step.
© CDISC 2012
Origin/Source
© CDISC 2012
SDTM Origin
© CDISC 2012
SDTM Annotated CRF Links – def:Origin
© CDISC 2012
Annotated CRF <def:AnnotatedCRF>
<def:DocumentRef leafID="LF.acrf"/> 3
1 </def:AnnotatedCRF>
© CDISC 2012
def:Origin Types
Allowable Values Context/Comments
CRF SDTM - Data collected on CRF
Derived SDTM & ADaM - Data value calculated from
other data values
Assigned SDTM & ADaM - Data values set independent
of subject related data values (e.g. Domain)
Protocol SDTM - Data values defined in the study
protocol
eDT SDTM - Data received via electronic data
transfer
Predecesser SDTM & ADaM - Data value is a direct copy of
a variable in another dataset
© CDISC 2012
ADaM Source – def:Origin
© CDISC 2012
ADaM Source - Assigned
<def:CommentDef
OID="COM.ADQSADAS.AVISITN">
<Description>
<TranslatedText>Numeric code for
AVISIT</TranslatedText>
</Description>
</def:CommentDef>
© CDISC 2012
ADaM Source -
Predecessor
© CDISC 2012
<ItemGroupDef OID="IG.ADQSADAS" Name="ADQSADAS" SASDatasetName="ADQSADAS" Repeating="Yes"
IsReferenceData="No" Purpose="Analysis"
def:Structure="One record per subject per parameter per analysis visit per analysis date" def:Class="BASIC DATA
ADaM Source – Derived
STRUCTURE" def:CommentOID="COM.ADQSADAS"
def:ArchiveLocationID="LF.ADQSADAS">
<Description>
<TranslatedText xml:lang="en">ADAS-Cog Analysis</TranslatedText>
</Description>
<ItemRef ItemOID="IT.ADQSADAS.STUDYID" OrderNumber="1" Mandatory="No"/>
…
<ItemRef ItemOID="IT.ADQSADAS.AVISIT" OrderNumber="18" Mandatory="No" KeySequence="3"
MethodOID="MT.ADQSADAS.AVISIT"/>
…
Note that the reference to the
</ItemGroupDef> derivation definition (method) is
from the ItemRef.
<ItemDef OID="IT.ADQSADAS.AVISIT" Name="AVISIT" SASFieldName="AVISIT" DataType="text" Length="16">
<Description>
<TranslatedText xml:lang="en">Analysis Visit</TranslatedText>
</Description>
<CodeListRef CodeListOID="CL.AVISIT"/>
<def:Origin Type="Derived"/>
</ItemDef>
© CDISC 2012
Exercise 2
1. Locate workbook saved from Exercise 1
2. Open (if not already opened)
3. Click on Exercise 2 tab (or use TOC)
4. Follow steps.
You are changing the Origin type for AGEGR1N to be
Derived (it starts out as Comment).
This involves changing the Origin in the Variables tab
and adding a row in the Methods tab.
© CDISC 2012
Controlled Terminology Types
• Enumerations
• Codelists
• Dictionary References
© CDISC 2012
Controlled Terminology Examples
Enumeration
Enumeration
CodeList
External Dictionaries
© CDISC 2012
Define-XML CodeList Components
attributes
CodeList
attributes @ OID
EnumeratedItem @ CodedValue
@ Name
@Rank
Alias @ DataType
@Order
attributes @ SASFormatName
CodeListItem @ CodedValue
@Rank
Decode
@Order
OR
TranslatedText
Alias
ExternalCodeList
Alias
81
© CDISC 2012
1 Enumeration Example
<EnumeratedItem CodedValue="RELATED"/>
</CodeList>
© CDISC 2012
1
Codelist Example
2
© CDISC 2012
Codelists vs. Enumerations
SDTM variables for coded values should use Define-
XML CodeList element:
© CDISC 2012
ADaM XXXN Variables should use Define-XML
Codelist element.
Enumeration
Codelist
XSL Stylesheet uses Decode value from Codelist.
CDISC Controlled Terminology
• CDISC Controlled Terminology (CT) is defined for
many SDTM variables, for a few ADaM variables
Maintained in the US National Cancer Institute (NCI)
Enterprise Vocabulary System
• For SDTM where there is a CDISC defined
Controlled Vocabulary, cross-references to the
relevant NCI codes should be provided in
codelists using the ODM Alias element.
Context attribute must be nci:ExtCodeID
• Some CDISC CTs are Extensible meaning
sponsor may add new terms
New terms should only be added if they are distinct in
meaning from the existing terms.
88
© CDISC 2012
CDISC Controlled Terminology
© CDISC 2012
CDISC Controlled Terminology Support
© CDISC 2012
ADaM Controlled Terminology
Examples
© CDISC 2012
Exercise 3
1. Click on Exercise 3 in Exercise workbook.
2. Follow the steps.
© CDISC 2012
Comments
© CDISC 2012
Comments
Comments may be added to dataset, variable and
value level metadata.
Comments should not be used to store information
about computational algorithms/derivations.
© CDISC 2012
def:CommentDef Components
MetaDataVersion
attributes
ItemGroupDef @def:CommentOID
attributes
ItemDef
@def:CommentOID
attributes
def:CommentDef @OID
Description
def:DocumentRef
99
© CDISC 2012
Where Can Comments Be Added?
Metadata XML Element Usage
Component
Dataset definition ItemGroupDef Domain level
explanations
Variable definition ItemDef -Variable level
explanations
- Describes Origin
when OriginType is
Assigned
Value definition ItemDef Value level
explanations
VLM Where Clause def:WhereClauseDef Describe join between
two datasets
© CDISC 2012
Comments with Links to External Documents
1
2
<ItemDef
1 def:CommentOID="COM.ARMCD">
OID="IT.DM.ARMCD" Name="ARMCD" DataType="text" Length="8" SASFieldName="ARMCD"
<Description>
<TranslatedText xml:lang="en">Planned Arm Code</TranslatedText>
</Description>
<CodeListRef CodeListOID="CL.ARMCD"/>
<def:Origin Type="Assigned"/>
</ItemDef>
<def:CommentDef OID="COM.ARMCD">
<Description>
2 <TranslatedText xml:lang="en">Assigned based on Randomization Number. See Note 2.1</TranslatedText>
</Description>
3 <def:DocumentRef leafID="LF.ReviewersGuide"/>
</def:CommentDef>
<def:SupplementalDoc>
1 <def:DocumentRef leafID="LF.ReviewersGuide"/>
<def:DocumentRef leafID="LF.ComplexAlgorithms"/>
</def:SupplementalDoc>
© CDISC 2012
Exercise 4
1. Click on Exercise 4 tab in Exercise Workbook.
2. Follow the steps.
3. The first 9 steps will guide you as you browse the
ADaM Define-XML display.
4. The remaining steps will guide you through
setting up references to specific pages in the
SDTM Reviewers’ guide for the TI and TS
datasets.
© CDISC 2012
Value Level Metadata
The SDTM domain models are highly normalized
data structures. As a result there are some cases
where the contents of a column or variable cannot
be unambiguously defined through Variable
Metadata.
© CDISC 2012
Vital Signs Example
108
© CDISC 2012
Define-XML 2.0 Value Level Metadata
• Value Level Metadata is usually attached to
XXSTRESN or XXORRES but may be attached to
any SDTM variable whose values should be
interpreted differently depending on other
variables.
Variable Where
QSSEQ PARAM IN NOTIN
(“ACTOT”)
QSSEQ PARAM EQ ACTOT
110
© CDISC 2012
Representing Value Level Metadata
© CDISC 2012
Representing Value Level Metadata
<def:ValueListDef OID="VL.VS.VSORRESU">
<ItemRef ItemOID="IT.VS.VSORRESU.HEIGHT.DM.COUNTRY.CMETRIC"
OrderNumber="1" Mandatory="Yes">
<def:WhereClauseRef
WhereClauseOID="WC.VS.VSTESTCD.HEIGHT.[DM].COUNTRY.CMETRIC"/>
</ItemRef>
<ItemRef ItemOID="IT.VS.VSORRESU.HEIGHT.DM.COUNTRY.CNMETRIC“
OrderNumber="2" Mandatory="Yes">
<def:WhereClauseRef WhereClauseOID="WC.VS.VSTESTCD.HEIGHT.[DM].COUNTRY.CNMETRIC
</ItemRef>
<ItemRef ItemOID="IT.VS.VSORRESU.WEIGHT.DM.COUNTRY.CMETRIC"
OrderNumber="3" Mandatory="Yes">
<def:WhereClauseRef WhereClauseOID="WC.VS.VSTESTCD.WEIGHT.[DM].COUNTRY.CMETRIC"
</ItemRef>
<ItemRef ItemOID="IT.VS.VSORRESU.WEIGHT.DM.COUNTRY.CNMETRIC"
OrderNumber="4" Mandatory="Yes">
<def:WhereClauseRef WhereClauseOID="WC.VS.VSTESTCD.WEIGHT.[DM].COUNTRY.CNMETRIC
</ItemRef>
</def:ValueListDef>
© CDISC 2012
Each Value Definition Includes a
WhereClause
<def:ValueListDef OID="VL.VS.VSORRESU">
<ItemRef ItemOID="IT.VS.VSORRESU.HEIGHT.DM.COUNTRY.CMETRIC"
OrderNumber="1" Mandatory="Yes">
<def:WhereClauseRef
WhereClauseOID="WC.VS.VSTESTCD.HEIGHT.[DM].COUNTRY.CMETRIC"/>
</ItemRef>
<ItemRef ItemOID="IT.VS.VSORRESU.HEIGHT.DM.COUNTRY.CNMETRIC“
OrderNumber="2" Mandatory="Yes">
<def:WhereClauseRef WhereClauseOID="WC.VS.VSTESTCD.HEIGHT.[DM].COUNTRY.CN
</ItemRef>
<ItemRef ItemOID="IT.VS.VSORRESU.WEIGHT.DM.COUNTRY.CMETRIC"
OrderNumber="3" Mandatory="Yes">
<def:WhereClauseRef WhereClauseOID="WC.VS.VSTESTCD.WEIGHT.[DM].COUNTRY.CM
</ItemRef>
<ItemRef ItemOID="IT.VS.VSORRESU.WEIGHT.DM.COUNTRY.CNMETRIC"
OrderNumber="4" Mandatory="Yes">
<def:WhereClauseRef WhereClauseOID="WC.VS.VSTESTCD.WEIGHT.[DM].COUNTRY.CN
</ItemRef>
</def:ValueListDef>
© CDISC 2012
Value Level Details
<def:WhereClauseDef OID="WC.VS.VSTESTCD.HEIGHT.[DM].COUNTRY.CNMETRIC
def:CommentOID="COM.SUBJECTDATA-JOIN-DM">
<RangeCheck SoftHard="Soft" def:ItemOID="IT.VS.VSTESTCD" Comparator="EQ
<CheckValue>HEIGHT</CheckValue>
</RangeCheck>
<RangeCheck SoftHard="Soft" def:ItemOID="IT.DM.COUNTRY" Comparator="EQ
<CheckValue>USA</CheckValue>
</RangeCheck>
</def:WhereClauseDef>
© CDISC 2012
Where Clause Details
<def:WhereClauseDef OID="WC.VS.VSTESTCD.HEIGHT.[DM].COUNTRY.CNMETRIC
def:CommentOID="COM.SUBJECTDATA-JOIN-DM">
<def:CommentDef<RangeCheck SoftHard="Soft" def:ItemOID="IT.VS.VSTESTCD" Comparator="EQ
OID="COM.SUBJECTDATA-JOIN-DM">
<CheckValue>HEIGHT</CheckValue>
<Description>
</RangeCheck>
<TranslatedText xml:lang="en">Join any Subject Level dataset with the Demographics da
on <RangeCheck SoftHard="Soft" def:ItemOID="IT.DM.COUNTRY" Comparator="EQ
<CheckValue>USA</CheckValue>
[IG.datasetname]IT.USUBJID = [IG.DM]IT.USUBJID, assuming 'IG.datasetname' is the OID of t
ItemGroupDef </RangeCheck>
that defines the
subject-level</def:WhereClauseDef>
dataset to be joined with the Demographics dataset.</TranslatedText>
</Description>
</def:CommentDef>
© CDISC 2012
Adding Value Level MetaData to a
Variable
attributes
def:ValueListDef
OID
attributes ItemOID
ItemRef
attributes
def:WhereClauseRef
1.. WhereClauseOID
attributes OID
def:WhereClauseDef
attributes
RangeCheck
def: ItemOID
CheckValue Comparator
118
© CDISC 2012
Value Level Metadata - ADaM
© CDISC 2012
Value Level Metadata – ItemDef
<def:ValueListDef OID="VL.SC.SCORRES">
<ItemRef ItemOID="IT.SC.SCORRES.EDLEVEL" OrderNumber="1" Mandatory="No">
<def:WhereClauseRef WhereClauseOID="WC.SC.SCTESTCD.EDLEVEL"/>
</ItemRef>
<ItemRef
<ItemDef ItemOID="IT.SC.SCORRES.MARISTAT"
OID="IT.SC.SCORRES.EDLEVEL" OrderNumber="2"
Name="SC.EDLEVEL" Mandatory="No">
DataType="text" Length="24"
<def:WhereClauseRef
SASFieldName="EDLEVEL"> WhereClauseOID="WC.SC.SCTESTCD.MARISTAT"/>
</ItemRef>
<Description>
<ItemRef ItemOID="IT.SC.SCORRES.SUBJINIT"
<TranslatedText OrderNumber="3" Mandatory="No">
xml:lang="en">Education Level</TranslatedText>
<def:WhereClauseRef WhereClauseOID="WC.SC.SCTESTCD.SUBJINIT"/>
</Description>
</ItemRef> CodeListOID="CL.EDLEVEL"/>
<CodeListRef
</def:ValueListDef>
<def:Origin Type="CRF">
<def:DocumentRef leafID="LF.blankcrf">
<def:PDFPageRef PageRefs="6" Type="PhysicalRef"/>
</def:DocumentRef>
</def:Origin>
</ItemDef>
© CDISC 2012
Value Level Metadata def:WhereClauseRef
<def:ValueListDef OID="VL.SC.SCORRES">
<ItemRef ItemOID="IT.SC.SCORRES.EDLEVEL" OrderNumber="1" Mandatory="No">
<def:WhereClauseRef WhereClauseOID="WC.SC.SCTESTCD.EDLEVEL"/>
</ItemRef>
</def:ValueListDef>
<def:WhereClauseDef OID="WC.SC.SCTESTCD.EDLEVEL">
<RangeCheck SoftHard="Soft" def:ItemOID="IT.SC.SCTESTCD" Comparator="EQ">
<CheckValue>EDLEVEL</CheckValue>
</RangeCheck>
</def:WhereClauseDef>
© CDISC 2012
ADaM Parameter/Value Level Metadata
<def:ValueListDef OID="VL.ADQSADAS.DTYPE">
<ItemRef ItemOID="IT.ADQSADAS.DTYPE.ACITM01-ACITM14" Mandatory="No">
<def:WhereClauseRef WhereClauseOID="WC.ADQSADAS.DTYPE.ACITM01-ACITM14"/>
</ItemRef>
<ItemRef ItemOID="IT.ADQSADAS.DTYPE.ACTOT" Mandatory="No">
<def:WhereClauseRef WhereClauseOID="WC.ADQSADAS.DTYPE.ACTOT"/>
</ItemRef>
</def:ValueListDef>
<def:WhereClauseDef OID="WC.ADQSADAS.DTYPE.ACITM01-ACITM14">
<RangeCheck Comparator="NOTIN" SoftHard="Soft" def:ItemOID="IT.ADQSADAS.PARAMCD">
<CheckValue>ACTOT</CheckValue>
</RangeCheck>
</def:WhereClauseDef>
<def:WhereClauseDef OID="WC.ADQSADAS.DTYPE.ACTOT">
<RangeCheck Comparator="EQ" SoftHard="Soft" def:ItemOID="IT.ADQSADAS.PARAMCD">
<CheckValue>ACTOT</CheckValue>
</RangeCheck>
</def:WhereClauseDef>
© CDISC 2012
ADaM Parameter/Value Level Metadata
© CDISC 2012
Exercise 5
1. Select Exercise 5 tab in Exercise workbook.
2. In steps 1-5 you will view VLM for LBORRES.
3. In the remaining steps you will add VLM for
LBSTRESN.
© CDISC 2012
Topics
Introduction to Define-XML
Define-XML
XML essentials
Metadata representation
Dataset-XML
126
© CDISC 2012
Protocol CDASH SDTM ADaM
Define-XML
SDM-XML
SDM-XML ODM
ODM Define-XML
Dataset-XML
Glossary
Therapeutic Areas
BRIDG
SHARE Questionnaires
© CDISC 2012
SEND SDTM ADaM SEND SDTM ADaM
Metadata Metadata Metadata Data Data Data
Define- Dataset-
XML XML
ODM
128
© CDISC 2012
Define-XML
Dataset-XML
Define-XML provides Dataset-XML Metadata
© CDISC 2012
Datasets
© CDISC 2012
ODM Model
attributes
ODM
@ Description
@ FileType
Study MetaData @ Granularity
@ Archival
AdminData
@ FileOID
ReferenceData @ CreationDateTime
@ PriorFileOID
ClinicalData Data & Audit @ AsOfDateTime
@ ODMVersion
Association @ Originator
@ SourceSystem
ds:signature
@ SourceSystemVersion
@ ID
133
© CDISC 2012
ODM Model
attributes
ODM
@ Description
@ FileType
Study Define-XML @ Granularity
@ Archival
AdminData
@ FileOID
ReferenceData @ CreationDateTime
@ PriorFileOID
ClinicalData Dataset-XML @ AsOfDateTime
@ ODMVersion
Association @ Originator
@ SourceSystem
ds:signature
@ SourceSystemVersion
@ ID
134
© CDISC 2012
ODM MetaData
MetaDataVersion
attributes
Include
@ OID
Protocol
@ Name
StudyEventDef
@ Description
FormDef
ItemGroupDef
ItemDef
CodeList
Presentation
ConditionDef
MethodDef
135
© CDISC 2012
ODM Data
SubjectData
attributes
Audit
Signature @ SubjectKey
InvestigatorRef @ TransactionType
SiteRef Audit
Annotation Signature
Annotation
StudyEventData Audit
FormData Signature
ArchiveLayoutRef
Annotation Audit
Signature
ItemGroupData Annotation
ItemData
136
© CDISC 2012
CDISC “end-to-end” process
CRO Sponsor
SDTM+/-
Investigator
Subject LABs
137
CDISC “end-to-end” process
CRO Sponsor
SDTM+/- Archive
ODM
Investigator
Subject LABs
138
Dataset-XML
V1 Development Goals:
• Keep it simple
• Primary use case is transport
Dataset-XML is intended to be a transport format
1-1 replacement for XPT – 1 file per SEND or SDTM
domain or per ADaM dataset
Requires no change to content standards
• Facilitate “apples-to-apples” FDA Pilot
• Require minimal additional effort for sponsors
Sponsors develop SDTM or ADaM datasets as usual
then convert to XML (vs. convert to XPT)
© CDISC 2012
“Like all magnificent things, it’s very
simple.”
- Natalie Babbett, Tuck Everlasting
© CDISC 2012
“Nature is pleased with simplicity. And
nature is no dummy”
- Isaac Newton
© CDISC 2012
One xml file per dataset
© CDISC 2012
ClinicalData
SubjectData
StudyEvent
Form
Dataset-XML Data
ItemGroup
ClinicalData Item
ItemGroup
Item
143
© CDISC 2012
144
© CDISC 2012
SDS-XML Sample
145
© CDISC 2012
Transport Use Case
© CDISC 2012
Transport Format vs. Operational
Format
• Transport Format is optimized for transmission of
data between organizations
Dataset-XML further specialized for regulatory
submission use case
• Platform independent
• Vendor neutral
Archive Archive
© CDISC 2014
FDA Pilot Test Methodology
1. Sponsor converts the original Phase 3 SDTM and ADaM (if
available) XPT datasets into DS-XML format with Define.xml
and sends them to the FDA (Batch 1).
2. Sponsor modifies the original Phase 3 SDTM SAS dataset to
contain longer variable names, labels, and text fields. Sponsor
then converts the new extended (modified) SAS dataset into
DS-XML format and sends them with Define.xml to the FDA
(Batch 2).
3. FDA performs a Data Processing Test on Batch 1 and Batch
2 by converting the XML datasets into .sas7bdat.
4. FDA performs a Data Matching Test on Batch 1 by
comparing the converted SDTM and ADaM SAS datasets (from
XML) with the converted SDTM and ADaM SAS datasets (from
XPT) using PROC COMPARE. FDA also performs a Data
Matching Test on Batch 2 by comparing the
converted SDTM SAS dataset (from XML) with the original
SDTM© CDISC
SAS 2014 dataset using PROC
© CDISC 2014
© CDISC 2014
Benefits of a CDISC Dataset XML Standard
• Open, non-proprietary standard without field width or
variable naming restrictions of SAS V5 Transport files.
• Supports FDA goal of encouraging open source reviewer
tool development.
• Harmonized with BRIDG, CDISC Controlled Terminology.
• Supports representation of data relationships, metadata
versions and audit trails.
Note: not all of these will be available in the first release.
• Data elements include references to metadata in Define-
XML.
• Straightforward implementation starting from SDTM data
in SAS or Oracle.
155
© CDISC 2014
How Dataset-XML Improves Submissions
• Short term – quick wins
Provides XPT functionality without 1980’s limitations
• 8 character variable names
• 200 character text variables
Enables tighter link with Define-XML
• Improved validation capabilities
• Long term
Content standards can evolve to take advantage of more
modern technology
© CDISC 2014
Dataset-XML Use Case Example
• Case 2: Dataset-XML for operational uses
Sponsor Regulatory Agency
Dataset Dataset
-XML -XML
Define- Define-
XML XML
Archive Archive
attributes
Audit
Signature @ SubjectKey
InvestigatorRef @ TransactionType
SiteRef Audit
Annotation Signature
Annotation
StudyEventData Audit
FormData Signature
ArchiveLayoutRef
Annotation Audit
Signature
ItemGroupData Annotation
ItemData
158
© CDISC 2011
Dataset-XML Capabilitities
• Typed data
Automated content validation
© CDISC 2011
Supports only ‘untyped’ data (so far)
© CDISC 2012
Exercise 6
Explore Dataset-XML examples.
© CDISC 2012
Topics
Introduction to Define-XML
Define-XML
XML essentials
Metadata representation
Dataset-XML
© CDISC 2012
Role of Analysis Results Metadata
• Facilitates review by documenting relationship
between analysis results and the analysis
datasets.
• Recommended in PMDA Technical Conformance
Guide
For key efficacy and safety results
Clinical study results that provide rationale for setting
dose and administration
© CDISC 2012
Components of Analysis Results
Metadata
• Analysis Display metadata definitions
Analysis Result metadata definitions
• Analysis Parameters
• Analysis dataset(s)
– Analysis variable(s)
– Selection criteria
• Documentation
• Programming Statements
© CDISC 2012
List of Displays
Display Details
© CDISC 2012
Structure of Define-XML Metadata
MetaDataVersion
def:SupplementalDoc
def:ValueListDef
def:WhereClauseDef
ItemGroupDef
ItemDef
CodeList
MethodDef
def:CommentDef
def:leaf
arm:AnalysisResultDisplays
© CDISC 2012
Structure of Define-XML Metadata
arm:AnalysisResultDisplays
MetaDataVersion
arm:ResultDisplay
def:SupplementalDoc
def:ValueListDef Description
def:WhereClauseDef
def:DocumentRef
ItemGroupDef arm:AnalysisResult
ItemDef
Description
CodeList
arm:AnalysisDatasets
MethodDef
def:CommentDef def:WhereClauseRef
def:leaf arm:AnalysisVariable
arm:AnalysisResultDisplays arm:Documentation
arm:ProgrammingCode
arm:Code
© CDISC 2012
2
1
<arm:AnalysisResultDisplays> 1
<arm:ResultDisplay OID="RD.Table_14-3.01" Name="Table 14-3.01">
<Description><TranslatedText xml:lang="en">
2
Primary Endpoint Analysis: ADAS-Cog - Summary at Week 24 - LOCF (Efficacy Population)
</TranslatedText> </Description>
<def:DocumentRef leafID="LF.Table-14-3.01">
<def:PDFPageRef PageRefs="2" Type="PhysicalRef"/>
</def:DocumentRef>
<arm:AnalysisResult OID="AR.Table_14-3.01.R.1" ParameterOID="IT.ADQSADAS.PARAMCD"
AnalysisReason="SPECIFIED IN SAP“AnalysisPurpose="PRIMARY OUTCOME MEASURE">
<Description><TranslatedText xml:lang="en">
3 Dose response analysis for ADAS-Cog changes from baseline
</TranslatedText></Description>
.
.
.
1
<arm:AnalysisResult
OID="AR.Table_14-3.01.R.1"
1 ParameterOID="IT.ADQSADAS.PARAMCD"
AnalysisReason="SPECIFIED IN SAP"
AnalysisPurpose="PRIMARY OUTCOME MEASURE">
<Description>
<TranslatedText xml:lang="en">Dose response analysis for ADAS-Cog changes from
baseline</TranslatedText>
</Description>
<arm:AnalysisDatasets>
<arm:AnalysisDataset ItemGroupOID="IG.ADQSADAS" >
1
<def:WhereClauseRef WhereClauseOID="WC.Table_14-3.01.R.1.ADQSADAS" />
<arm:AnalysisVariable ItemOID="IT.ADQSADAS.CHG"/>
</arm:AnalysisDataset>
<def:WhereClauseDef 1
OID="WC.Table_14-3.01.R.1.ADQSADAS">
</arm:AnalysisDatasets>
<RangeCheck Comparator="EQ" SoftHard="Soft" def:ItemOID="IT.ADQSADAS.PARAMCD">
1
<CheckValue>ACTOT</CheckValue>
</RangeCheck>
<RangeCheck Comparator="EQ" SoftHard="Soft" def:ItemOID="IT.ADQSADAS.AVISIT">
<CheckValue>Week 24</CheckValue>
</RangeCheck>
2
<arm:AnalysisResult
OID="AR.Table_14-3.01.R.1"
ParameterOID="IT.ADQSADAS.PARAMCD"
AnalysisReason="SPECIFIED IN SAP"
AnalysisPurpose="PRIMARY OUTCOME MEASURE">
<Description>
<TranslatedText xml:lang="en">Dose response analysis for ADAS-Cog changes from
baseline</TranslatedText>
</Description>
<arm:AnalysisDatasets>
<arm:AnalysisDataset ItemGroupOID="IG.ADQSADAS" >
<def:WhereClauseRef WhereClauseOID="WC.Table_14-3.01.R.1.ADQSADAS" />
2
<arm:AnalysisVariable ItemOID="IT.ADQSADAS.CHG"/>
2
</arm:AnalysisDataset>
<ItemDef OID="IT.ADQSADAS.CHG" Name="CHG" SASFieldName="CHG" DataType="integer" Length=
</arm:AnalysisDatasets>
<Description> 2
<TranslatedText xml:lang="en">Change from Baseline</TranslatedText>
</Description>
<def:Origin Type="Derived"/>
</ItemDef>
3
<arm:AnalysisResult OID="AR.Table_14-3.01.R.1"
ParameterOID="IT.ADQSADAS.PARAMCD"
AnalysisReason="SPECIFIED IN SAP" 3
4 AnalysisPurpose="PRIMARY OUTCOME MEASURE">
<Description>
<TranslatedText xml:lang="en">Dose response analysis for ADAS-Cog changes from
baseline</TranslatedText>
</Description>
<arm:AnalysisDatasets>
<arm:AnalysisDataset ItemGroupOID="IG.ADQSADAS" >
<def:WhereClauseRef WhereClauseOID="WC.Table_14-3.01.R.1.ADQSADAS" />
<arm:AnalysisVariable ItemOID="IT.ADQSADAS.CHG"/>
</arm:AnalysisDataset>
</arm:AnalysisDatasets>
5
6
<arm:AnalysisResult
OID="AR.Table_14-3.01.R.1"
ParameterOID="IT.ADQSADAS.PARAMCD"
<def:WhereClauseDef OID="WC.Table_14-3.01.R.1.ADQSADAS"> 6
AnalysisReason="SPECIFIED IN SAP"
<RangeCheck Comparator="EQ" SoftHard="Soft" def:ItemOID="IT.ADQSADAS.PARAMCD">
AnalysisPurpose="PRIMARY OUTCOME MEASURE"> <ItemGroupDef OID="IG.ADQSADAS"
<CheckValue>ACTOT</CheckValue>
<Description> Name="ADQSADAS"
</RangeCheck>
<TranslatedText xml:lang="en">Dose response analysis for ADAS-Cog changes from
SASDatasetName="ADQSADAS"
<RangeCheck Comparator="EQ" SoftHard="Soft" 5 def:ItemOID="IT.ADQSADAS.AVISIT">
baseline</TranslatedText> Repeating="Yes"
<CheckValue>Week 24</CheckValue>
</Description> IsReferenceData="No"
</RangeCheck>
<arm:AnalysisDatasets> 5
Purpose="Analysis"
<RangeCheck Comparator="EQ" SoftHard="Soft" def:ItemOID="IT.ADQSADAS.EFFFL">
<arm:AnalysisDataset ItemGroupOID="IG.ADQSADAS" >
def:Structure="One record per subject per param
<CheckValue>Y</CheckValue>
6 <def:WhereClauseRef WhereClauseOID="WC.Table_14-3.01.R.1.ADQSADAS" />
</RangeCheck> def:Class="BASIC DATA STRUCTURE"
<arm:AnalysisVariable ItemOID="IT.ADQSADAS.CHG"/>
def:CommentOID="COM.ADQSADAS"
<RangeCheck Comparator="EQ" SoftHard="Soft" def:ItemOID="IT.ADQSADAS.ANL01FL">
</arm:AnalysisDataset> def:ArchiveLocationID="LF.ADQSADAS">
<CheckValue>Y</CheckValue>
</arm:AnalysisDatasets> …
</RangeCheck>
</def:WhereClauseDef>
7
8
Topics
Introduction to Define-XML
Define-XML
XML essentials
Metadata representation
Dataset-XML
© CDISC 2012
Tooling Options
• Define-XML Creation
Commercial applications
• SAS Clinical Standards Toolkit
• Pinnacle21 Community Version
• Formedix
• Xclinical Tabulator
• XML4Pharma SDTM-ETL
Toolkits provided by consultants
CROs
• Validation tools
Pinnacle21
Define.xml Checker (XML4Pharma)
© CDISC 2012
Stylesheet Challenges
• Defining display requirements
Goal is to help reviewers understand data contents
• Platform dependencies
Browsers
• Link navigation
• Printing
© CDISC 2012
© CDISC 2012
A note about XSL
• XSL is a programming language that is extremely
useful for transforming the information in an XML
file into a readable display
• XSL generated displays are key for reviewers.
© CDISC 2012
Define.XML Validation
Define-XML conformity is define by
• XML Syntactic & Semantic requirements defined by the
Define-XML schema
• Define-XML Semantic requirements & Business rules
defined by the Define-XML 2.0 Specification
ValidationTools
• Validating parsers
• Pinnacle 21
Resources
• Whitepaper www.cdisc.org/define-xml
184
© CDISC 2012
Define-XML Conformity
Compliant with normative content in Define-XML
specification
&
Validates to Define-XML Schema
185
© CDISC 2012
External File References
© CDISC 2012
Annotated CRF
4
<def:leaf ID="LF.acrf" xlink:href=“aCRF.pdf">
<def:title>Annotated Case Report Form</def:title>
</def:leaf>
© CDISC 2012
Reviewers’ Guide (SDTM)
<def:SupplementalDoc>
<def:DocumentRef leafID="LF.ReviewersGuide"/>
<def:DocumentRef leafID="LF.ComplexAlgorithms"/>
</def:SupplementalDoc>
</def:leaf>
© CDISC 2012
Dataset References
© CDISC 2012
OIDs
• OIDs are used throughout the Define-XML to
provide unique identifiers for elements.
• OID references are used to reference specific
element definitions.
ItemOID
CodelistOID
MethodOID
ValueListOID
…
• The Define-XML sample files use a naming
convention designed to make examples easier to
read. This naming convention is not part of the
standard.
• OIDs must be machine generated.
© CDISC 2012
Building Define-XML Q&A
© CDISC 2012
Thank You!
Merci!
Danke!
Dankie
Grazie!
Gracias!
Sally.Cassells@gmail.com
195
© CDISC 2012
Thank you for attending.
Please complete the feedback survey that will be
sent to your inbox within the next couple of days.
PDF certificates of attendance will shortly follow.
© CDISC 2012
CDISC Membership Benefits
• Gold Membership
Access to Member’s Only area of website
Access to Members Only Webinars
20% off all CDISC-hosted conferences and education
events
Opportunity to obtain CDISC ODM Certification
Opportunity to obtain CDISC Registered Solution
Provider
Opportunity to participate in CDISC licensed course
program
Personalized company plaque
© CDISC 2012
Membership Benefits Cont’d
• Platinum Membership
All Gold Membership benefits
40% (already includes 20% discount for Gold
Membership) discount on CDISC-hosted events and
Education courses
Representation on CDISC Advisory Council
• Networking opportunities with clients, visionaries, and other peers
• Opportunity for Board Committee Representation (Financial,
Strategy, Technical)
• Voting privileges for Board of Directors
• Participation in town hall meetings with regulators and CDISC
Board Members
• Free one-time access to F2F training of Global Approach to
Accelerating Medical Research course
© CDISC 2012
Contact Information and Helpful Links
• For membership inquiries, please contact
membership@cdisc.org
• For education inquiries, please contact
training@cdisc.org
• For general CDISC inquiries, please contact
info@cdisc.org
© CDISC 2012
Exercise Materials
• Define-XML Examples folder
• Dataset-XML Examples folder
• ExerciseWorkbook.xslx – Step by step instructions
• SDTM-Metadata-Worksheet.xlsx – Excel
workbook generated using Open CDISC (V2.0.2)
• ADaM-Worksheet-Worksheet.xslx – Excel
workbook generated using Open CDISC (V2.0.2)
• ExerciseOutput folder
Define-XML displays with ‘answers’ to exercises 3 & 5
Excel files for Exercises 3,4,5
© CDISC 2014
Exercises
Exercise 1 - Locate sample files then open and browse
Exercise 3 – CodelistMetadata
Exercise 6 -- Dataset-XML
202
Timing
8:30-10:45 - Introduction, XML essentials, Datasets and
variables, Exercise 1
203
Timing (Osaka)
9:00-10:45 - Introduction, XML essentials, Datasets and variables, Exercise
1 (including short break)
Lunch 11:30-13:00
204