STDM generation in Lundbeck - From Source to Target H. Lundbeck A/S 16 -Mar-18 1
Agenda • Background • Architecture & Tools • Design and Implementation of the SDTM conversion framework • Evaluation • Plan for further delvelopement • Questions/Discussion H. Lundbeck A/S 16 -Mar-18 2
Background • This is a presentation about: – The considerations and experiences Lundbeck has had when implementing their SDTM conversion process • This a not a presentation about – Hardcore SAS macros – A detailed explaination of the SDTM data model. H. Lundbeck A/S 16 -Mar-18 3
Background • The SDTM conversion project was initated and developed in 2007 -2008 as a part of the Lundbeck e. Clinical program. • Started with one programmer under the supervision of a project lead. • Current status: the SDTM conversion process is rolled out in Line of Business (the CDO Department) and involves five programmers and one project manager H. Lundbeck A/S 16 -Mar-18 4
Background • The goal was to implement a framework for converting any clinical study data (source) into an SDTM data format (target) • Benefits of framework: – Extend the use of industry standards (CDISC/SDTM) to the clinical data process flow – Eliminate data transformation / alignment both within and across studies and from external sources – Enable systematic and extensive use of current and historical data – Simplify data interchange with partners and regulatory authorities H. Lundbeck A/S 16 -Mar-18 5
Background • Source data from four Rave studies were selected as the basis for the implementation of the framework – Lundbeck project Lu. AA 21004, studies 11984 A, 11984 B and 11985 A and Lundbeck project Lu. AA 24530: study 11918 A H. Lundbeck A/S 16 -Mar-18 6
Background • The experiences from converting the first four studies will lay the basis for a Lundbeck SDTM data model – Domain definitions of all domains currently used in the SDTM conversion process – Convension rules /Algorithms – CDISC/Lundbeck Controlled Terminology H. Lundbeck A/S 16 -Mar-18 7
Architecture • Available tools for the SDTM conversion: – Clinical study data (a lot!) – An ETL-tool for data manipulation – The SDTM Implementation Guidelines version 3. 1. 1/3. 1. 2 – CDISC Controlled Terminology – In-house experience of a previously SDTM conversion project (Gaboxadol) H. Lundbeck A/S 16 -Mar-18 8
Architecture SOURCE TARGET Rave Report ing Tool Datafax SAS-Di Studio e. DM/e. DE ETL Tool SDTM Interface CROs SADS Partners H. Lundbeck A/S 16 -Mar-18 9
Design and Implementation • The shortest distance between two points is a stright line: A H. Lundbeck A/S B 16 -Mar-18 10
Design and Implementation • Divide the line into stages: SOUR CE H. Lundbeck A/S Stage 1 Stage 2 16 -Mar-18 Stage 3 Stage 4 11 TARG ET
Design and Implementation • Stage 1(Define): – Define the mapping from source to target • Annotate CRF according to the SDTM format – Implement the mapping • Build up data dictionaries (SAS format libraries) Example H. Lundbeck A/S 16 -Mar-18 12
Design and Implementation • Stage 2(Apply): – Apply the mappings to the source data • Apply the SAS user defined formats – Set the value of general variables • E. g. STUDYID, DOMAIN, VISITNUM, VISITDY – A SAS macro library H. Lundbeck A/S 16 -Mar-18 13
Design and Implementation • Stage 3(Derive): – Derive standard values based on the original collected result(Relevant for the Findings Observation Class). • E. g. the variables --STRESC and STRESN and standard units in –STRESU – Derive other values. • E. g. the variables --DY, VSTESTCD=BMI, --SEQ H. Lundbeck A/S 16 -Mar-18 14
Design and Implementation • Stage 4(Structure): – Create the final structure on datasets • E. g. create the supplemental qualifiers domains. – Create the CO (Comments Domain) – Create the Trial Design Domains (TI, TE, TV, TS, TA) H. Lundbeck A/S 16 -Mar-18 15
Design and Implementation • Depictured with work-demanded proportions: Derive Define Stage 3 SOUR CE Stage 1 TARG ET Stage 2 Stage 4 Apply H. Lundbeck A/S 16 -Mar-18 Structure 16
Evaluation • Evaluation of the benefits: – Extend the use of industry standards (CDISC/SDTM) to the clinical data process flow – Eliminate data transformation / alignment both within and across studies and from external sources – Enable systematic and extensive use of current and historical data – Simplify data interchange with partners and regulatory authorities H. Lundbeck A/S 16 -Mar-18 17
Further Development • Implement the Web. SDTM validation checks • Development and implementation of a CDASH library – Greater transparency in the data flow • SDTM knowledge/experience sharing between companies (Network Groups) H. Lundbeck A/S 16 -Mar-18 18
…and some memories ’Computer Charlie er en helt almindelig mand & han ta'r toget ind til City hver dag gemmer sit ansigt så godt han ka' bag et slør fra dagens første Havanna han ligner til forveksling alle andre i kupé'en der til gengæld ligner koksgrå lig & sidder pænt på række som på bunden af en grav mens station efter station bare suser forbi’ ’I en blågrå kupe’ af C. V. Jørgensen H. Lundbeck A/S 16 -Mar-18 19
Thanks to … • H. Lundbeck A/S • Nordic CIDSC User Group • Danish Association for Data Managers H. Lundbeck A/S 16 -Mar-18 20
Questions/Discussion ? H. Lundbeck A/S 16 -Mar-18 21
Скачать презентацию STDM generation in Lundbeck — From Source to
3c46236eeeaadc3262b4621827959c55.ppt