1 PhUSE SDE, 28-May-2008 1 A SAS based Solution for define.xml Monika Kawohl Statistical Programming Accovion
Jan 31, 2016
1
PhUSE SDE, 28-May-2008 1
A SAS based Solutionfor
define.xml
Monika KawohlStatistical Programming
Accovion
2
PhUSE SDE, 28-May-2008 2
Presentation Topic – define.xml
Excerpt from the CDISC Draft Metadata Submission Guidelines define.xml Sample
3
PhUSE SDE, 28-May-2008 3
Presentation Outline
Electronic Submission Context & Purpose
XML Basics
define.xml Sections/Elements
define.xml Generation Process
Expected define.xml Enhancements
Summary & Conclusions
4
PhUSE SDE, 28-May-2008 4
Electronic Submission Context & Purpose
Define Document Mandatory when Submitting Data to FDA
Purpose
• Describe Structure and Contents of Data
• Facilitate Review via Standardized Metadata Format
Aim: More Efficient Overall Review Process
define.xml Preferred Data Definition Format for SDTM
• define.xml Human- and Machine-Readable
Benefit of define.xml not Restricted to Submissions
5
PhUSE SDE, 28-May-2008 5
define.xml Documentation/Samples
CDISC Case Report Tabulation Data Definition Specification (define.xml),Version 1.0, February 9, 2005
Sample define.xml Included
CDISC Metadata Submission Guidelines, Appendix to the SDTM IG V3.1.1, Draft Version 0.9, July 25, 2007
Sample define.xml Included as Part of Sample Submission
CDISC SDTM/ADaM Pilot (Pilot 1), January 31, 2008
Mock Submission Package Available for CDISC Members
6
PhUSE SDE, 28-May-2008 6
XML Basics
Schema (Extension: .XSD)
• Declaration of Elements and their Attributes
• Prerequisite for Machine-Readability
XML File (Extension: .XML)
• Data and Metadata in Machine-Readable Format
• Usage of Elements and Attributes as Defined in Schema
Style Sheet (Extension: .XSL)
• Definition of Layout in Browser Tool for Human-Readability
• Usage of Elements and Attributes as Defined in Schema
7
PhUSE SDE, 28-May-2008 7
<?xml version="1.0" encoding="ISO-8859-1"?><?xml-stylesheet type="text/xsl" href="define1-0-0.xsl"?><ODM ...<ItemGroupDef OID="DM" Name="DM" Repeating="No" IsReferenceData="No" Purpose="Tabulation" def:Label="Demographics" def:Structure="One record per subject" def:DomainKeys="STUDYID, USUBJID" def:Class="Special Purpose" def:ArchiveLocationID="Location.DM"> ...
Style Sheet
Reference
Interaction of XML, XSL, XSD
8
PhUSE SDE, 28-May-2008 8
define.xml Sections/Elements
Data Metadata (TOC)
Variable Metadata
Variable Value Level Metadata
Computational Algorithms
Controlled Terminology/Code Lists
Annotated CRF
Optional: Supplemental Data Definition Document
Navigation via Bookmarks and Hyperlinks
9
PhUSE SDE, 28-May-2008 9
Data Metadata
10
PhUSE SDE, 28-May-2008 10
Data Metadata – XML Code
<ItemGroupDef OID="LB" Name="LB" Repeating="Yes" IsReferenceData="No" Purpose="Tabulation" def:Label="Laboratory Tests" def:Structure="One record per lab test per time point per visit per subject" def:DomainKeys="STUDYID,USUBJID,LBTESTCD,VISITNUM,LBTPTNUM" def:Class="Findings" def:ArchiveLocationID="Location.LB"> ... <def:leaf ID="Location.LB" xlink:href="LB.xpt"> <def:title>lb.xpt</def:title> </def:leaf>
</ItemGroupDef>
11
PhUSE SDE, 28-May-2008 11
Variable Metadata
12
PhUSE SDE, 28-May-2008 12
Variable Metadata – XML Code<ItemGroupDef OID="LB" ... <ItemRef ItemOID="LB.LBTESTCD" OrderNumber="5" Mandatory="Yes" Role="Topic"/> ... <ItemRef ItemOID="LB.LBBLFL" OrderNumber="22" Mandatory="No" Role="Record Qualifier"/> </ItemGroupDef>... <ItemDef OID="LB.LBTESTCD" Name="LBTESTCD" DataType="text" Length="8" Origin="CRF" Comment="CRF Pages 5, 10, 15, 20" def:Label="LAB Test or Examination Short Name"> <def:ValueListRef ValueListOID="ValueList.LB.LBTESTCD"/> </ItemDef>...<ItemDef OID="LB.LBBLFL" Name="LBBLFL" DataType="text" Length="1" Origin="Derived" def:Label="Baseline Flag" def:ComputationMethodOID="COMPMETHOD.LBBLFL"> <CodeListRef CodeListOID="YF"/></ItemDef>
13
PhUSE SDE, 28-May-2008 13
Variable Value Level Metadata
14
PhUSE SDE, 28-May-2008 14
Variable Value Level Metadata – XML Code<def:ValueListDef OID="ValueList.LB.LBTESTCD"> <ItemRef ItemOID="LB.LBTESTCD.ALB" OrderNumber="1" Mandatory="No"/> ... </def:ValueListDef>... <ItemDef OID="LB.LBTESTCD.ALB" Name="ALB" DataType="float" Length="8" SignificantDigits="1" Origin="CRF" Comment="CRF Pages 5, 15" def:Label="Albumin" def:DisplayFormat="5.1"/>
15
PhUSE SDE, 28-May-2008 15
Computational Algorithms
Complex Derivations
Derivations Used More than Once
16
PhUSE SDE, 28-May-2008 16
Computational Algorithms – XML Code
<def:ComputationMethod OID="COMPMETHOD.LBBLFL"> Derive mean of pre-treatment measurements. Create new record with result and flag LBBLFL='Y'</def:ComputationMethod> ... <ItemDef OID="LB.LBBLFL" Name="LBBLFL" DataType="text" Length="1" Origin="Derived" def:Label="Baseline Flag" def:ComputationMethodOID="COMPMETHOD.LBBLFL"> <CodeListRef CodeListOID="YF"/> </ItemDef>
Masking of Special Characters
• Ampersand, Apostrophe, Quote, Less Than, Greater Than
• ' ‘
17
PhUSE SDE, 28-May-2008 17
Controlled Terminology/Code Lists
External Dictionary References incl. Versions
• e.g., MEDDRA, WHODRUG
18
PhUSE SDE, 28-May-2008 18
Controlled Terminology/Code Lists– XML Code<ItemDef OID="LB.LBBLFL" Name="LBBLFL" DataType="text" Length="1“ Origin="Derived" def:Label="Baseline Flag" def:ComputationMethodOID="COMPMETHOD.LBBLFL"> <CodeListRef CodeListOID="YF"/> </ItemDef> ...<CodeList OID="YF" Name="YF" DataType="text"> <CodeListItem CodedValue="Y"> <Decode> <TranslatedText xml:lang="en">YES</TranslatedText> </Decode> </CodeListItem></CodeList>
19
PhUSE SDE, 28-May-2008 19
Annotated CRF (blankcrf.pdf)
20
PhUSE SDE, 28-May-2008 20
Supplemental Data Definitions
Optional
PDF Document
Additional Information Useful for Data Review
• General Assumptions
• Flowcharts
• Derivation Dependancies
• Reviewers' Guide
21
PhUSE SDE, 28-May-2008 21
define.xml - SAS Based Generation ProcessUse All Metadata Already Available in SAS
Provide Additional Information Required
• Set-up at Design and Specification Level
• Format: Excel Spreadsheets
• Contents: CDISC Terminology, Study Specific Metadata (CRF Pages)
Combine Metadata and Additional Information in SAS
Create XML File in SAS
Use Stylesheet Provided with CDISC Sample
22
PhUSE SDE, 28-May-2008 22
Available Metadata vs. Additional Information RequiredMetadata Section
Available Required
Dataset SAS Dictionary Tables:
Name, Label
CDISC Metadata Terminology
Class, Structure, Keys
Attributes for Machine-readability
Variable SAS Dictionary Tables:
Name, Label, Type, Format
Length, Significant Digits
CDISC Metadata Terminology
Role, (Origin)
Study Specific Metadata
(Origin), Comment
• CRF Page References
• Derivation Descriptions
23
PhUSE SDE, 28-May-2008 23
Available Metadata vs. Additional Information Req. - ContinuedMetadata Section
Available Required
Variable Value Level
Parent Dataset Contents:
Values of --TESTCD, --TEST
Values of QNAM, QLABEL
(Attributes of Parent Variables:)
e.g., Type, Length
e.g., CRF Page References
Adaptation on Value Level:
Type, Length, Format
Significant Digits, Display Format
Origin, Comment
Computational Algorithms - Reference Name
Derivation Description
Code Lists SAS Format Library:
Name, Code, Decode, Type-
24
PhUSE SDE, 28-May-2008 24
Excel File Format for Variable Metadata
25
PhUSE SDE, 28-May-2008 25
Process Flow
SASFormats
DatasetMetadata
VariableMetadata
Computa-tional
Algorithms
EXCEL
SASDatasets
EXCELDraft
Variable Value Level
Metadata
EXCELEdited
Variable Value Level
Metadata
DEFINE.XML
AnnotatedCRF
Supple-mentalData
Definitions
XPTFiles
26
PhUSE SDE, 28-May-2008 26
Automated Consistency Checks
SDTM Adherence Checks
• Availability of Datasets and Variables
• Order of Variables in Dataset
• Labels and Data Type
• Variables with Controlled Terminology(SAS Format Attached)
Consistency Checks for Well-formed XML Code
• Intra Document Links, e.g. Computational Methods
Additional Manual Checks Required
27
PhUSE SDE, 28-May-2008 27
Expected Enhancements – ADaM IntegrationAspects of CDISC Pilot 1 => CDISC define.xml Standard
Link to Analysis Metadata (s.b.)
Link to CSR Table
Link to Variable Metadata of ADSL
Link to SAP
28
PhUSE SDE, 28-May-2008 28
Expected Enhancements - Continued
Correction of Software Issues with 2007 define.xml Sample
Adaptation to Latest CDISC ODM Standard
• V2.0 => V3.0
Improved Printability
• Stylesheet Enhancements
• Alternative Options (define_xml_printable.pdf)
Extension for CDISC ADaM Specific Metadata
Executable Computational Algorithm?
29
PhUSE SDE, 28-May-2008 29
Summary & Conclusions
Presented Solution Just One of Many Options
• Driven by Available Skills
• XML Code Easy to Implement according to CDISC Standards
• Biggest Challenge: Process Set-up
Advantages
• Early Integration (Design and Specification Level)
Increased Consistency, Lower Risk of Redundancy
• SDTM Adherence Check
• Built-in Consistency with SAS Datasets
CDISC define.xml Standard is Work in Progress
PhUSE SDE, 28-May-2008 30
Thank you!
Questions?