Transcript 5-4Covidence

08/04/2016
Implementation of
CDISC Data Standards in a CRO
Kurt Loeffler, PhD, Managing Director
Herbert Noack, PhD, Head of CDM
Birgit Keller, Head Biostatistics Cardiology
Covidence GmbH, Eschborn (Frankfurt), Germany
Japan CDISC Group
January 28, 2004, Tokyo
1
08/04/2016
Objectives for CDISC implementation
at Covidence
Need for a data standard that has been widely accepted by suppliers,
sponsors and clients, and authorities in the ICH regions (EU, USA, J)
Reduction in complexity of clinical trial data models, and interfaces
between systems
Need for a standard data model as a basis for development of
re-usable solutions to data processing, data analysis and data reporting, e.g.
 Standard CRF modules and database modules
 Data validation and data cleaning
 Standard layout of tables, listings, figures
 Analysis programs, SAS macros
Better quality of data and results
Reduction in time and effort
Technological advantage
Reduction in cost
2
08/04/2016
Data transfer with CDISC standards
Standards to enable seamless data flow from patients to reviewers
ECG
Operational
Database
LAB
ODM
LAB
CRFs
SDS
ADaM
Metadata
Study data
Audit trail
EDC
CRO
eSubmission for
Regulatory Review
Metadata
CRT datasets
Analysis datasets
3
08/04/2016
Covidence Implementation of CDISC Standards
Metadata definitions based on SDS Version 2.0 and LAB 1.0.1
implemented, migration to SDS Version 3.1 planned
CRF modules and annotated CRFs based on SDS 2.0 implemented
Data Dictionary based on SDS 2.0 and LAB 1.0.1 implemented in CDMS,
codelists added according to CRO and sponsor definitions
Interface to SAS for SDS and LAB datasets implemented
Mapping of Clinical Trial Data to Analysis Datasets based on SDS 3.0
and ADaM requirements currently implemented
Tables Shells (CRO and sponsor defined) in preparation
Analysis Programs according to ADaM in preparation
Archiving of clinical trial data based on ODM in preparation
Solution for legacy standards, legacy projects ?
4
08/04/2016
Submissions Data Standards - Version 3
Version 3 of the CDISC Submission Data Domain
Models has been prepared by the CDISC Submission
Data Standards (SDS) Group to guide the organization,
structure, and format of the Case Report Tabulation
datasets for clinical trials submitted to the FDA. The
focus of the SDS Group has been on Case Report
Tabulation (CRT) datasets. The FDA is developing a
repository for all sponsor data and a suite of
standard review tools (e.g. PPD Patient Profiles, CSS
Informatics) to access, manipulate, and view, the CRT
data. V3 is intended to facilitate transfer of data from
sponsors to the FDA and subsequent loading into the
FDA repository (JANUS).
5
08/04/2016
Submissions Data Standards - Roadmap
Feb/Mar 2004: Finalization SDS Version 3.1
Apr 2004: HL7 Ballot => Acceptance as HL7 Standard (ANSI)
Thereafter: FDA will recommend the use of SDS v.3.1
for all CRT data submitted
Future: replace SAS-XPT files with XML files
6
08/04/2016
FDA Reviewer Tools (from Steve Wilson)
7
08/04/2016
FDA Reviewer Tools (from Steve Wilson)
8
08/04/2016
FDA Reviewer Tools (from Steve Wilson)
9
08/04/2016
It all starts with an annotated CRF: Demo
10
08/04/2016
It all starts with an annotated CRF: Vital signs
11
08/04/2016
Implementation of CDISC SDS + LAB
standards at Covidence
CLINTRIAL®
ORACLE-CLINICAL®
SDS data dictionary
SDS panel structure
SDS compliant context
panel
LAB model panels
SDS data dictionary
SDS data collection modules
SDS compliant key info
LAB data collection modules
SASenv
CONTEXT PANEL/Key information
Demo
Disposition
DM
DS
Expo- Phys.Ex MedHist Concom Adverse Vital
sure
Med
Event Signs
EX
PE
MH
CM
AE
VS
Efficacy
Lab
LB
12
08/04/2016
Electronic data in CDISC standards (XML)
and today’s DBMS
Problem:
• Transfer format is XML
• Clintrial and OC do not (yet) support data import and
export in XML-format
Solution:
• XML tools to convert files to other formats, e.g. XML spy
13
08/04/2016
CRO: Implementation of CDISC-SDS in DBMS
Codelist
CRO own
SDS core
Codelist
Sponsor A
Codelist
Sponsor B
Codelist
Sponsor C
SDS extension – Sponsor A
SDS extension – Sponsor B
SDS extension – Sponsor C
14
08/04/2016
CRO: Implementation of CDISC-SDS in SAS
CDISC
DBMS
Post-Processing
CDISC-SDS compliant
- including sponsor codelists
- including sponsor extensions
Demo Interventions
Demo
DM
EX
CM
Findings
Events
Expo- Concom Adverse Dispos
Med
sure
Event
AE
ADaM compliant
Analysis Datasets
e.g. CFB standard
DS
Non CDISC
DBMS
Mapping
Med.
Hist
Incl.
Excl.
Vital
Signs
Lab
ECG
Test
Phys.
Exam
MH
IE
VS
LB
EG
PE
Exten.
Efficacy
CRT, Tables, Listings, Graphs
15
08/04/2016
Analysis and reporting
SAS macros based on SDS and ADaM standards
Mapping from CDMS standards to SDS ,
(e.g., from CDISC DD based on SDS 2.0 to SDS 3.0)
Case Report Tabulations based on SDS data structure
Standard Table shells based on SDS structures for SDS
Domains
Analysis Programs based on currently available three
ADaM data models:
• Time-to-event data
• Change from Baseline
• Categorical data
16
08/04/2016
Mock Output:
Time to Disease Progression
Table: Time to Disease Progression
Patient Group: All Randomized Subjects
Drug A Alone
(N=xxx)
Drug A+B
(N=xxx)
xxx
xxx
No. Of Subjects with Disease Progression
xxx
xxx
No. Of Subjects without Disease Progression (censored)
xxx
xxx
Percent Censored
xx.x
xx.x
xx.xx
xx.xx
(xx.xx, xx.xx)
(xx.xx, xx.xx)
25 – 75 percentile
xx.xx – xx.xx
xx.xx – xx.xx
Minimum - Maximum
xx.xx – xx.xx+
xx.x – xx.xx+
No. of Randomized Subjects
Time to Disease Progression (days)
Median
(95% c.i.)
Hazard Ratio (relative to Drug A Alone)
(95% c.i.)
xx.xx
(xx.xx, xx.xx)
P-value
Logrank
0.xxxx
Wilcoxon
0.xxxx
17
08/04/2016
18
Recommendations for legacy projects
Project
status
CDMS
SAS
Environment
SAS
analysis
progs
Recommendation
Before
submission
NonCDISC
NonCDISC
NonCDISC
Map integrated SAS-DB to SDS and
ADaM; rerun each study report using
ADaM datasets for validation
Last Phase
III study,
reporting
started
NonCDISC
NonCDISC
Phase III
running
NonCDISC
Map into SDS, develop analysis datasets
from SDS according to ADaM;
develop integrated SAS-DB according to
SDS and ADaM and rerun each study
report using ADaM datasets for validation
Map into SDS, develop analysis datasets
from SDS according to ADaM;
prepare integrated SAS-DB of phase II
studies according to SDS and ADaM and
rerun each study report using ADaM
datasets for validation
08/04/2016
CDISC implementation in summary
Annotated CRF modules (SDS)
Data dictionary (SDS)
Clintrial® panel structure (SDS+LAB)
OC®-Data Collection Modules (SDS+LAB)
standardized SAS datasets (compliant with SDS and ADaM)
CRT programs compliant with SDS
Table shells and analysis programs
compliant with SDS and ADaM standards
Future:
- a uniform data model from data acquisition to data submission:
- availability of ODM / XML for data import and export and
interchange between systems ?
19
08/04/2016
Observations and preliminary experience
Wide interest in the biopharmaceutical industry in EU
Most companies still rely on proprietary standards
CDISC standards increasingly used for new projects,
to map data to CRT and to produce analysis datasets in submissions
Some sponsor start to ask for compliance to CDISC
standards in rfp‘s
Further implementation in the industry expected after release of SDS 3.1
Availability and re-use of CDISC compatible standards and modules
allows for significant cost savings in CRF design, database set-up, data
validation programming, and statistical programming
Thank you !
[email protected]
20