Customer Oriented CDISC Implementation · Cornerstones: CDISC recommendations FDA expectations...

10/20/2005 1

Customer Oriented CDISC Implementation

Edelbert Arnold and Ulrike Plank

Accovion, Germany

October 12, 2005

10/20/2005 2

Efficient Strategy in Clinical Development

Cornerstones:

CDISC recommendations

FDA expectations

Sponsor’s requirements

Accovion’s inventory

10/20/2005 3

Cornerstones (I)

CDISC recommendations:

FDA expectations:

• Study Data Tabulation Model (SDTM)

• Analysis Dataset Model (ADaM)

• SDTM and ADaM compliance recommended (less strict for ADaM)

• Documentation (metadata and links, preferably define.xml)

• Supplemental documents (annotated CRF, programs)

10/20/2005 4

Cornerstones (II)

Sponsor’s requirements

• In-house vs out-sourcing

• CDISC relevance

• Sponsor provided material:

CRFs only

CDMS based study data

Other source data (e.g. analysis data)

10/20/2005 5

Cornerstones (III)

Accovion’s inventory:

• Core functions

Database Programmers (DBP)

Statistical Programmers (SP)

• Software

CDMS: Clintrial (CT) and Oracle Clinical (OC)

Analysis software: SAS 8.2 and 9.1.3

10/20/2005 6

Accovion’s CDISC relevant tasks

Database set-up (DBP)

Generation of Full-SDTM and ADaM datasets (SP)

Integration of different data structures (SP)

Analysis output (SP)

10/20/2005 7

CT / OC“Pre-SDTM“

Annotated

CRF

Mappingprograms

SDTM

ADaM

Report, submissionAnalysis output

(tables, listings and graphs)

SAS macros, sample

programs

Submission to the FDA?

N

Standard tool

xpt

define. xml

Macro to create xpt files

YSponsor’s

dataData

Integration Tool

Our workflow

10/20/2005 8

Database set-up

Current situation:SDTM Implementation Guide (SDTM-IG) recommended by the FDA

Consequences:

Consider SDTM-IG

Fill gap, if CDMS is not 100% SDTM compliant

Implementation strategies differ between companies

10/20/2005 9

Strategy for Database Set-up

Use of metadata libraries based on SDTM

Efficient implementation of changes in data structure

Flexibility in case of complex derivations

Additional variables needed for data analysis

10/20/2005 10

Use global libraries including:

Clintrial /Oracle Clinical

data

Accovion’s strategy

(Pre-SDTM= SDTM pure+

SDTM comfort)Copy and adapt on study level

SDTM pure: SDTM variables and metadata

SDTM comfort: Additional variables foranalysis

CDMS generates a Pre-SDTM: Some variables will be derived in SAS

10/20/2005 11

SDTM comfort: Benefit in relation to SDTM pure

Numeric codes:

Date variables:

Used in analysis and listings

Used for convenient data selection and sorting

DSDECOD (SDTM pure) DSDECODN (SDTM comfort)

ADVERSE EVENT 1COMPLETED 2... ...NON-COMPLIANCE WITH STUDY DRUG 6... ...STUDY TERMINATED BY SPONSOR 13... ...OTHER 999

EGDTC (SDTM pure) EGDT+EGDC (SDTM comfort)

2005-10-12T09:15 12OCT2005 (16721)

12-OCT-2005

10/20/2005 12

SDTM and ADaM implementation

Topic SDTM ADaMTopic SDTM ADaM

Version 1.1 (SDS V3.1) General Considerations: 1.0, other models 0.x

Topic SDTM ADaM


Reviewer Medical Reviewer Statistical Reviewer

Topic SDTM ADaM



Requirements Standardization Standardization

Topic SDTM ADaM




Analysis friendly: “One PROC away”

Topic SDTM ADaM





Characteristics Domain concept Analysis oriented

Topic SDTM ADaM





Characteristics Domain concept

No redundancyAnalysis oriented

Common variables in each dataset

Topic SDTM ADaM






No redundancy

CRF data and trial design data

Analysis oriented


More derived data

Topic SDTM ADaM






No redundancy

CRF data and trial design data

Textual results

Analysis oriented


More derived data

Numeric codes

10/20/2005 13

Possible strategies for SDTM and ADaM implementation

• Independent creation: CDMS ⇒ SDTM, CDMS ⇒ADaM

• Consecutive creation: CDMS ⇒ SDTM ⇒ ADaM

• Other concepts

CDMS (=SDTM) ⇒ ADaM

If CDMS not 100% SDTM compliant:

10/20/2005 14

Selecting a strategy

What is available in the CDMS?

Dual derivations

Inconsistencies

Reviewer’s needs

Potential of standardization

10/20/2005 15

CDMS DraftSDTM/ADaMData

Mapping

Data Mapping

Data Mapping

Note: SAS is used for Data Mapping

Accovion’s strategy

SDTM

ADaM

10/20/2005 16

SAS Post-Processing for Full-SDTM

Draft SDTM/ADaM ⇒ SDTM

• Delete comfort variables

• Add datasets, e.g. supplemental qualifier datasets SUPP~ (only if necessary)

CDMS ⇒ Draft SDTM/ADaM• Derivation of variables, e.g. treatment start

• Add records, where needed, e.g. baseline records

10/20/2005 17

ADaM Implementation

CDISC guidance for ADaM:

General Considerations V1.0

Subject-Level Analysis (ADSL dataset, draft)

Other guidance (draft)

ADaM guidelines are not as restrictive as those for SDTM!

10/20/2005 18

ADaM implementation at Accovion

• Add variables, e.g change from baseline

• Add ADSL information → supports “One PROC away”

• Derive variables only once

Use Draft SDTM/ADaM as input

Create ADSL dataset

Create other ADaM datasets

Define standards across projects

Few study/project specific changes

10/20/2005 19

Tools for SDTM and ADaM Generation

SAS program generator

Standardized data mapping using SAS

• Dataset mapping table (one per study)

• Variable mapping tables (one per domain)

• Macro creates programs and datasets

At Accovion:

10/20/2005 20

Dataset mapping table: Example study 1234

Studyid Domain Description Structure Sort order1234 ADSL Subject Level Key Information Special Purpose USUBJID1234 ADAE Adverse Events Events USUBJID, AESEQ1234 ADEF Efficacy Findings USUBJID, EFTESTCD, EFANLTMN

ADaM domain metadata information

Studyid Domain Variable Label Type Length Format SAS library

Dataset name

Variable name

Task Transform (SAS code)

Com-ments

Exec. order

1234 ADSL SEX Sex Char 1 SDTM DM SEX no change1234 ADSL TRTAN Actual Treatment

Group NumberNum SDTM EX EXTRT macro %trta(indata,

outdata)… 2

ADaM variable metadata information Source information Mapping information

Variable mapping tables: Example domain ADSL

Tables to feed program generator

10/20/2005 21

Automation of Documentation

For submission to the FDA:

Documentation at Accovion:

Datasets as xpt files

Documentation for SDTM and ADaM in define.xmlformat

Standard tool

Usage of mapping tables

Additional information, e.g. role

10/20/2005 22

Output 1: Domain level metadata, example study 1234

Dataset Description Structure Purpose Keys Location DocumentationADSL Subject Level

Key Information1 record per subject Analysis USUBJID ../1234/dds/

adsl.xptSAP and/or adsl.sas

ADAE Adverse Events 1 record per subject per event Analysis USUBJID, AESEQ ../1234/dds/adae.xpt

SAP and/or adae.sas

ADEF Efficacy 1 record per subject per parameter per analysis time window

Analysis USUBJID, EFTESTCD, EFANLTMN

../1234/dds/adef.xpt

SAP and/or adef.sas

ADaM Datasets for study 1234

Variable Label Type Controlled Terms or Format

Origin Role Comment

SEX Sex Char DM.SEX Result qualifier No changeTRTAN Actual Treatment

Group NumberNum 0 = ‘Placebo’

1 = 'Drug A'2 = 'Drug B'3 = 'Drugs A+B'

Derived from EX.EXTRT

Selection, result qualifier

For one subject either one or two treatments in EXTRT are possible

Variable Metadata for Dataset ADSL – 1 record per subject

Output 2: Variable level metadata, example dataset ADSL

Output of Accovion’s standard tool

10/20/2005 23

Integrating different data structures (I)

Deviating data structures:

• Legacy studies

• Sponsor’s structures

Accovion full service:

• Standardized “Pre-SDTM”

• SDTM

• ADaM

A migration to SDTM and ADaM may be necessary!

10/20/2005 24

Integrating different data structures (II)

Study specific source data:

• CDMS structure

• SDTM

• Analysis datasets

Migration to fixed target data:

• Meta-analysis: ADaM only

• Data for submissions: SDTM + ADaM

10/20/2005 25

Issues for consideration

Programming effort

Validation effort

Reusability of programs/potential of standardization

Risk of errors

Risk of inconsistencies

10/20/2005 26

Usage of standard mapping tool for data integration!

CDMS

SDTM

ADS

SDTM

ADaM

1. Direct Migration

CDMS

2. Migration to “Pre-SDTM”

Pre-SDTM(= SDTM pure+SDTM comfort)

Data Integration Tool

Accovion standard mapping

Accovion’s proposals

Data Integration Tool

10/20/2005 27

Analysis output

• Review of SAP and TLG shells

• “Analysis-ready” datasets

Programming based on analysis datasets (ADaM)

ADaM standards support usage of standard analysis tools

At Accovion a set of 20 standard tables was developed

Consider analysis output for ADaM set-up:

10/20/2005 28

Example “Change from Baseline in Primary Efficacy Variable – ITT population”:

Use efficacy dataset ADEF with selection criteria “WHERE (EFTESTCD=‘PEF’AND ITT=‘Y’);” T a b l e 1 . 1 C h a n g e f r o m B a s e l i n e i n P r i m a r y E f f i c a c y V a r i a b l e – I T T p o p u l a t i o n

N o t e : 1 p - v a l u e s f r o m p a i r w i s e c o m p a r i s o n A B v e r s u s A r e s p . B f r o m t - t e s t ; I T T = i n t e n t i o n - t o - t r e a t ; t r e a t m e n t g r o u p s a s t r e a t e d .

A n a l y s i s S t a t i s t i c R e s u l t C h a n g e f r o m B a s e l i n eW e e k _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ A B A B A B A B ´ ( N = x x ) ( N = x x ) ( N = x x ) ( N = x x ) ( N = x x ) ( N = x x ) _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ B a s e l i n e N w i t h d a t a x x x x x x x x x N m i s s i n g x x x x x x x x x M e a n x x x . x x x x . x x x x . x S D x x x . x x x x x . x x x x x . x x M i n x x x x x x x x x M e d i a n x x x . x x x x . x x x x . x M a x x x x x x x x x x 2 N w i t h d a t a x x x x x x x x x x x x x x x x x x N m i s s i n g x x x x x x x x x x x x x x x x x x M e a n x x x . x x x x . x x x x . x x x x . x x x x . x x x x . x S D x x x . x x x x x . x x x x x . x x x x x . x x x x x . x x x x x . x x M i n x x x x x x x x x x x x x x x x x x M e d i a n x x x . x x x x . x x x x . x x x x . x x x x . x x x x . x M a x x x x x x x x x x x x x x x x x x x p - v a l u e 1 x . x x x x x . x x x x 9 5 % C I [ x x . x ; x x . x ] [ x x . x ; x x . x ] e t c .

EFANLTMN

TRTAN

EFSTRESN

EFSTRESN=EFBLRESN

EFCHGBL

Relationship ADaM ⇔ analysis output

10/20/2005 29

FDA: Recommendation for CDISC analysis metadata structure not yet available

Analysis Name Description Reason Dataset DocumentationTable 1.1: ef00001t.lst Change from Baseline in Primary

Efficacy Variable – ITT populationPrespecified in Protocol

../1234/dds/adef.xpt SAP, Section X.Y and/or ef00001t.sas

Table 1.2: ef00002t.lst Proportion of Responders at Endpoint – Per-Protocol population

Prespecified in Protocol

../1234/dds/adef.xpt SAP, Section X.Y and/or ef00002t.sas

Analysis-Level Metadata for study 1234

Documentation of analysis

Follow ADaM General Considerations:

10/20/2005 30

Summary

CDISC impacts the clinical development process

CDISC standardization:• Efficiency ↑

• Risk of errors ↓

Accovion’s modular approach:• Benefit from standardization

Study specific needs

Changes in SDTM and ADaM models

Customer specific needs

• Flexibility:

10/20/2005 31

Thank you!

Thank you for your attention!

Thanks to our colleagues for all the efforts todevelop this strategy!

Edelbert Arnold and Ulrike PlankAccovion GmbH, Germany

[email protected] and [email protected]

Date post:	27-May-2020
Category:	Documents
Upload:	others
View:	6 times
Download:	0 times

Customer Oriented CDISC Implementation · Cornerstones: CDISC recommendations FDA expectations...

Documents