1
Managing CDISC version changes: how & when to implement? Presented by Lauren Shinaberry, Project Manager Business & Decision Life Sciences 2
Content Standards Technical Standards SDTM v1.1 SDTM IG v3.1.1 Define.xml ODM ODM v1.1 v1.2.1 Updates over time SEP 2013 SDTM ODM BRIDG v1.3 SDTM IG v1.1 v3.1 ADaM ODM BRIDG / v2.0 v1.2 SDTM v1.2 SDTM IG v3.1.2 CDASH BRIDG v2.1 BRIDG v2.0 BRIDG v1.1.1 Semantics Therapeutic Areas ADaM v2.1 ADaM IG BRIDG v3.0 BRIDG v2.2 BRIDG v3.0.3 ADaM Val. Checks BRIDG v3.0.2 ODM v1.3.1 BRIDG v3.0.1 Protocol Model SDTM v3.1.2 Am.1 Alzheimer SDM.XML ADaM Val. Checks v1.1 CDASH v1.1 SEND v3.0 Virology Parkinson s Disease Devices BRIDG v3.2 ADaM Val. Checks v1.2 SDTM v1.3 SDTM IG v3.1.3 Tuberculosis Pain PRM Toolset CDASH UG BRIDG v3.1 SDTM QS Supplements CDASH E2B SAE IG ADaM MD Guide BRIDG v4.0 BRIDG UG v2 Define.xml v2.0 PKD CDASH v1.2 SDTM v1.4 SDTM IG v3.1.4 SDTM Associated Persons IG SEND v3.0.1 Asthma Alzheimer v1.1 ADaM IG v1.1 Protocol Concept Guide Cardiovascular Therapeutic Brain Injury Schizophrenia Oncology Virology- Hepatitis Diabetes C Multiple Sclerosis Define.xml IG Validation CDASH v2.0 SDTM v1.5 SDTM IG v3.1.5 ADaM General Occurrence Model ADaM Integration IG Extended ODM PRM XML Schema SDTM.xml SEND v3.1 SDTM Devices IG v1.1 SDTM Device Submission Pilot 2002 2005 20032006 2004 2007 2008 2009 2010 2011 2012 2013 2014
More complex and increasing challenges NEW CDISC RELEASE What is different compared to the legacy standards? How to update the legacy standards most efficiently? Etc. CLINICAL DATA STANDARDS CLINICAL TRIALS What standards version to use? What standards versions were used previously? What is the scope of the standardisation Etc. SUBMISSION DOSSIER How to evaluate overall CDISC compliance How can we compare easily between studies? Etc. 4
Up-versioning How do you weigh the potential costs? How much of an impact is there on existing study data? SDTM 1.3 SDTM 1.4 5
Overview Case study: effort comparison Converting data from one version of SDTM to another Converting data directly from source to new version of SDTM Case study: managing end-to-end versioning in a metadata repository Allows for tracking interdependent changes across CDISC standards Facilitates comparisons between different versions of libraries as well as at the study level 6
Case Study: Effort comparison Two Phase III studies Same compound and indication Comparable number of SDTM datasets created Converted to SDTM v1.2 One study was already in SDTM v1.1 One study was converted from source 7
Significant Reduction Grand Total Trail Design SDTM acrf QC Define.xml QC consistency of metadata Programming Mapping Familiarization with study materials 0 50 100 150 200 250 300 350 400 Source to SDTM v1.2 SDTM v1.1 to v1.2 Activity Percent Reduction Familiarization with study materials 50% Mapping 65% Programming 80% QC consistency of metadata 78% QC Define.xml 90% SDTM acrf 37% Trail Design 0% Total Reduction in Effort 71% 8
Summary Case Study 1 As expected, up-versioning was significantly less effort than converting directly to the new version in this case Sample size is very small and the exact reduction numbers may not be this low over many more studies The areas for reduction are expected to be consistent and an overall reduction of ~70% is possible How to get a more accurate assessment for your studies? 9
Case study: managing your metadata Previous case study only included up-versioning for SDTM End-to-end implementations require the ability to manage complex versioning and relationships between standards Using concept of Active From and Active To dates allows retrieval of CDASH and SDTM standards at any point in time Linking metadata in study metadata repository allows impact analysis 10
Case Study Overview The Data Standards Library enables definition of CDASH and SDTM metadata linked to your DCMs available for easy selection to create study metadata A flexible concept for linking metadata across the 3- level CDASH and 5-level SDTM libraries An innovative concept of metadata versioning based on active from and to dates providing full control and flexibility over library changes Reporting capability to create Comparison Reports across library versions as well as other reporting
Case Study - high level flow DATA COLLECTION AREA DATA MANAGEMENT AREA DATA STANDARDS LIBRARY DBA STUDY SPECIFICATION EDC STUDY METADATA SPECIFICATION STUDY METADATA REPOSITORY COMPARISON DM VALIDATION ISSUE REPORTS ISSUE REPORTS CRO STUDY SDTM DATA & METADATA DEFINE.XML
Data Standards Library The Data Standards Library contains : Data Collection Modules: using CDASH metadata with clustered SDTM metadata annotated for CDASH annotated for SDTM Metadata Definitions : SDTM Standards Therapeutic Area Standards DATA COLLECTION MODULES SDTM 1.3 SDTM IG 3.1.3 THERAPEUTIC AREA METADATA DATA STANDARDS LIBRARY
Data Collection Modules - CDASH DATA COLLECTION MODULES LIBRARY DM 1 DM 2 VS 1 VS 2 EG 1... CRF template using CDASH annotations
Data Collection Modules - SDTM DATA COLLECTION MODULES LIBRARY DM 1 DM 2 VS 1 VS 2 EG 1...
Metadata Definitions DATA COLLECTION MODULES DATA STANDARDS LIBRARY 5 levels of metadata SDTM IG 3.1.3 and TA metadata stored in the same physical tables SDTM 1.3 SDTM IG 3.1.3 THERAPEUTIC AREA METADATA METADATA DEFINITIONS DOMAIN METADATA COMPUTATIONAL ALGORITHMS VARIABLE CONTROLLED METADATA TERMINOLOGY VALUE LEVEL METADATA
Allows Time Travel for Librarians PAST PRESENT FUTURE DCM Domain CDASH metadata SDTM metadata Code Lists Code Value VLM Name VLM Value Computational Algorithm Data Validation Rule
Example of versioning Active from 2-OCT-2013 SIZE codelist SMALL MEDIUM LARGE EXTRA LARGE Active from 30-OCT-2013
Comparison of library metadata against study build 4 different statuses : Ok : no conflict in metadata Difference : difference in the metadata value Missing : metadata not in study data Addition : metadata not in selected library
Case Study Summary Can manage links between CDASH, SDTM, Controlled Terminology, etc. and accommodate differing version releases between them Can easily run reports to compare metadata from study specifications against new (or old) versions of standards Quickly identify if any studies are affected by the new standard Use this information along with the metrics on effort to up-version to make business decision for your submission 20
Questions? The first step towards change is awareness. - Nathaniel Branden 21