automated sdtm creation and discrepancy detection jobs: the numbers tell the tale

20
AUTOMATED SDTM CREATION AND DISCREPANCY DETECTION AND DISCREPANCY DETECTION JOBS: THE NUMBERS TELL THE TALE Joris De Bondt PhUSE Conference 2014 12-15 Oct 2014

Upload: sgs

Post on 19-Jun-2015

412 views

Category:

Health & Medicine


1 download

DESCRIPTION

The FDA is advising use of data standards as early as possible in the study lifecycle. As a result, Data Management centers are using the Study Data Tabulation Model (SDTM) to drive operations from First Patient In till Database Lock. Many tools on the market allow for the creation of SDTM datasets via intuitive user interfaces. However, targeted tools are needed to manage nightly jobs taking care of data source downloads (eCRF, ePRO, Lab, etc), data uploads in a staging database, converting to SDTM and running edit checks before the Clinical Data Manager arrives in the morning.

TRANSCRIPT

Page 1: Automated SDTM Creation and Discrepancy Detection Jobs: The Numbers Tell The Tale

AUTOMATED SDTM CREATION AND DISCREPANCY DETECTIONAND DISCREPANCY DETECTION JOBS:THE NUMBERS TELL THE TALE

Joris De Bondt

PhUSE Conference 2014

12-15 Oct 2014

Page 2: Automated SDTM Creation and Discrepancy Detection Jobs: The Numbers Tell The Tale

OUTLINE

Setting the scene

OUTLINE

g• SGS’ Conversion Center

• SDTM data sources

• Timing for CDMS data refresh

SGS’ Conversion Center

• Glossary

• Technologies usedTechnologies used

• What is it?

• Code snippets• Code snippets

• Performance tuning

2PhUSE Conference, 12-15 Oct 2014

Page 3: Automated SDTM Creation and Discrepancy Detection Jobs: The Numbers Tell The Tale

SETTING THE SCENE: SGS’ CONVERSION CENTER

A central tool to manage the end-to-end flow of refreshing the

SGS CONVERSION CENTER

REQUEST g gData Management study databases

• REQUEST: sends an automatic data request via:

DOWN LOAD

INSTALL REQUEST: sends an automatic data request via:

Web browser, on a variety of vendor specific platforms

SAS Connect

INSTALL

CONVERT

SAS Connect

to obtain a new extract of the data

(unless scheduled transfer is available)

META DATA

REPORTS( )

• DOWNLOAD: downloads new data extract from:

PAGES 2 LOCK

A2O

Web browser, on a variety of vendor specific platforms

SAS ConnectSAS

CLIN

3PhUSE Conference, 12-15 Oct 2014

sFTP serverCLIN

TRIAL

Page 4: Automated SDTM Creation and Discrepancy Detection Jobs: The Numbers Tell The Tale

SETTING THE SCENE: SGS’ CONVERSION CENTER

• INSTALL: loads new data extract (Oracle dmp, CSV, SAS,

SGS CONVERSION CENTER

REQUEST ( pexcel, etc.) into the Data Management database

• CONVERT: converts all current study data into SDTM format

DOWN LOAD

INSTALL

• METADATA: generates new metadata files for the SDTM data

• REPORTS: generates reports for external distribution

INSTALL

CONVERT

REPORTS: generates reports for external distribution

• PAGES2LOCK: generates overview of eCRF pages that meet locking criteria

META DATA

REPORTS g

• A2O: refreshes output of checks and listings

• SAS: creates SAS datasets from the SDTM data

PAGES 2 LOCK

A2O • SAS: creates SAS datasets from the SDTM data

• CLINTRIAL: starts the synchronisation of the SDTM database to Clintrial

SAS

CLIN

4PhUSE Conference, 12-15 Oct 2014

ClintrialCLIN TRIAL

Page 5: Automated SDTM Creation and Discrepancy Detection Jobs: The Numbers Tell The Tale

SETTING THE SCENE: SDTM DATA SOURCESSDTM DATA SOURCES

Site specificsystems

eCRF Paper CRF eSource ePRO Lab, imaging, etc.

5PhUSE Conference, 12-15 Oct 2014

Page 6: Automated SDTM Creation and Discrepancy Detection Jobs: The Numbers Tell The Tale

SETTING THE SCENE: TIMING FOR CDMS DATA REFRESH

Refreshing

TIMING FOR CDMS DATA REFRESH

REQUEST g• the SDTM database

• the output of checks and listings

DOWN LOAD

INSTALL p g

impacts the work of Data Management (DM), so

INSTALL

CONVERT

impacts the work of Data Management (DM), so• the best time is overnight, when DM staff is not at work

• but also DM Programmers are not around

META DATA

REPORTSg

meaning that the entire end-to-end flow should run unattended

PAGES 2 LOCK

A2O g

SAS

CLIN

6PhUSE Conference, 12-15 Oct 2014

CLIN TRIAL

Page 7: Automated SDTM Creation and Discrepancy Detection Jobs: The Numbers Tell The Tale

SGS’ CONVERSION CENTERGLOSSARY

Windows Powershell

GLOSSARY

• Microsoft's task automation and configuration management framework

• Consists of a command-line shell and associated scripting language built on .NET FrameworkFramework

• Administrative tasks are generally performed by cmdlets (pronounced command-lets), which are specialized .NET classes implementing a particular operation

• Sets of cmdlets may be combined into scripts, executables (which are standaloneSets of cmdlets may be combined into scripts, executables (which are standalone applications).

jQuery• cross-platform JavaScript library designed to simplify the client-side scripting of

HTML

• free, open source software, licensed under the MIT License

• jQuery's syntax is designed to make it easier to navigate a document, select DOM elements, create animations, handle events, and develop Ajax applications.

VMWare

7PhUSE Conference, 12-15 Oct 2014

• provides cloud and virtualization software and services

Page 8: Automated SDTM Creation and Discrepancy Detection Jobs: The Numbers Tell The Tale

SGS’ CONVERSION CENTERTECHNOLOGIES USED

VMWare, Powershell

TECHNOLOGIES USED

REQUEST

VMWare, Powershell

VMWare Powershell Oracle PL/SQL

DOWN LOAD

INSTALL VMWare, Powershell, Oracle PL/SQL

VMWare, Powershell, Oracle PL/SQLjQuery

INSTALL

CONVERT

VMWare, Powershell, Oracle PL/SQL

VMWare, Powershell, Oracle PL/SQL, Crystal Reports

jQuery,PHP

META DATA

REPORTS

VMWare, Powershell, Oracle PL/SQL

VMWare, Powershell, Oracle PL/SQL

PAGES 2 LOCK

A2O , ,

VMWare, Powershell, Oracle PL/SQL, SAS

VMW P h ll O l PL/SQL

SAS

CLIN

8PhUSE Conference, 12-15 Oct 2014

VMWare, Powershell, Oracle PL/SQLCLIN TRIAL

Page 9: Automated SDTM Creation and Discrepancy Detection Jobs: The Numbers Tell The Tale

SGS’ CONVERSION CENTERWHAT IS IT?WHAT IS IT?

All end-to-end study process information one-click-away in a single overview

Powershell

9PhUSE Conference, 12-15 Oct 2014

Oracle

Page 10: Automated SDTM Creation and Discrepancy Detection Jobs: The Numbers Tell The Tale

SGS’ CONVERSION CENTERWHAT IS IT?WHAT IS IT?

History information

Error information Error information

Powershell

10PhUSE Conference, 12-15 Oct 2014

Oracle

Page 11: Automated SDTM Creation and Discrepancy Detection Jobs: The Numbers Tell The Tale

SGS’ CONVERSION CENTERWHAT IS IT?WHAT IS IT?

Performance statistics, for detecting problems and possible improvements

Performance tuning

11PhUSE Conference, 12-15 Oct 2014

Oracle

Page 12: Automated SDTM Creation and Discrepancy Detection Jobs: The Numbers Tell The Tale

SGS’ CONVERSION CENTERPOWERSHELL CODE SNIPPETPOWERSHELL CODE SNIPPET

Automatic ‘On demand’ request of Rave SAS datasets

Via “InternetE plorer Application”• Via “InternetExplorer.Application” object

• Setting the URL

• Adding username and password

12PhUSE Conference, 12-15 Oct 2014

Oracle

Page 13: Automated SDTM Creation and Discrepancy Detection Jobs: The Numbers Tell The Tale

SGS’ CONVERSION CENTERJQUERY CODE SNIPPETJQUERY CODE SNIPPET

Customized SGS tooltip enablement to allow for a web 2.0 experienceweb 2.0 experience

13PhUSE Conference, 12-15 Oct 2014

Oracle

Page 14: Automated SDTM Creation and Discrepancy Detection Jobs: The Numbers Tell The Tale

SGS’ CONVERSION CENTERPERFORMANCE TUNINGPERFORMANCE TUNING

Processing time for small phase I trials large phase III trialsg p g p

From minutes to hours

When to start tuning efforts?

CONVERSIONstart date: 2014-08-26 05:07

When to start tuning efforts?

CONVERSIONt t d t 2014 08 25 00 58

CONVERSIONstart_date: 2014-08-26 03:48end_date: 2014-08-26 05:25d ti 01 36 08

_end_date: 2014-08-26 05:12duration: 00:05:15status: FINISHEDerror: start_date: 2014-08-25 00:58

end_date: 2014-08-25 04:29duration: 03:31:13status: FINISHED

duration: 01:36:08status: FINISHEDerror:

error:Phase I

Phase IIB status: FINISHEDerror:

Phase IIB

Phase III

14PhUSE Conference, 12-15 Oct 2014

Oracle

Page 15: Automated SDTM Creation and Discrepancy Detection Jobs: The Numbers Tell The Tale

SGS’ CONVERSION CENTERPERFORMANCE TUNINGPERFORMANCE TUNING

Based on overall cross-study comparisony p

REQUESTDOWNLOAD

INSTALL CONVERTMETADATA

REPORTS PAGES 2 LOCK A2O SAS

CLINTRIAL

15PhUSE Conference, 12-15 Oct 2014

Oracle

Page 16: Automated SDTM Creation and Discrepancy Detection Jobs: The Numbers Tell The Tale

SGS’ CONVERSION CENTERPERFORMANCE TUNINGPERFORMANCE TUNING

Based on overall timing evolution on study levelg y

16PhUSE Conference, 12-15 Oct 2014

Oracle

Page 17: Automated SDTM Creation and Discrepancy Detection Jobs: The Numbers Tell The Tale

SGS’ CONVERSION CENTERPERFORMANCE TUNINGPERFORMANCE TUNING

Based on task-timing evolution on study levelg y

“Expected Time of Arrival”

17PhUSE Conference, 12-15 Oct 2014

Oracle

Page 18: Automated SDTM Creation and Discrepancy Detection Jobs: The Numbers Tell The Tale

SGS’ CONVERSION CENTERPERFORMANCE TUNINGPERFORMANCE TUNING

Allowing for targeted performance tuning

DOWNLOADDOWNLOAD

g g p g

Data download

DOWNLOADstart_date: 2014-04-15 23:30end_date: 2014-04-15 23:56duration: 00:26:46t t FINISHED

DOWNLOADstart_date: 2014-04-06 23:30end_date: 2014-04-07 00:27duration: 00:57:45t t FINISHED status: FINISHED

error:status: FINISHEDerror:

CONVERSIONstart date: 2014 04 25 12:25

CONVERSIONstart date: 2014 04 07 01:40

SDTM conversion

start_date: 2014-04-25 12:25end_date: 2014-04-25 16:03duration: 03:37:27status: FINISHED

start_date: 2014-04-07 01:40end_date: 2014-04-07 06:08duration: 04:28:08status: FINISHED

18PhUSE Conference, 12-15 Oct 2014

error:error: Oracle

Page 19: Automated SDTM Creation and Discrepancy Detection Jobs: The Numbers Tell The Tale

TAKE HOME MESSAGESTAKE HOME MESSAGES

Data Management centers• face wide variety of

• source datasource data• data provider platforms

process small phase I to large phase III trialsreq ire minimal impact of data refresh for DM staff require minimal impact of data refresh for DM staff

IT solutions can be put in place to run a sequence of DM IT solutions can be put in place to run a sequence of DM actions unattended overnight allowing for DM Programmers to

T k t d• Track unexpected errors• Tune the process performance• Manage all trials with minimal oversight effort

19PhUSE Conference, 12-15 Oct 2014

REQUESTDOWNLOAD

INSTALL CONVERTMETADATA

REPORTS PAGES 2 LOCK A2O SAS

CLINTRIAL

Page 20: Automated SDTM Creation and Discrepancy Detection Jobs: The Numbers Tell The Tale

THANK YOU FOR YOUR ATTENTIONTHANK YOU FOR YOUR ATTENTION

Joris De BondtLife Science ServicesHead EDC and Data StandardsWHERE EXPERIENCE SGS Belgium NVGeneraal De Wittelaan 19A b52800 Mechelen – BelgiumPhone: +32 (0)15 29 93 40Mobile: +32 (0)473 746477Fax: +32 (0)15 27 32 50E il j i d b dt@

WHERE EXPERIENCEMEETS SPEED

20PhUSE Conference, 12-15 Oct 2014

E-mail : [email protected]/CRO