Cdisc sdtm implementation_process _v1


Published on

Published in: Business, Technology
  • Be the first to comment

No Downloads
Total views
On SlideShare
From Embeds
Number of Embeds
Embeds 0
No embeds

No notes for slide

Cdisc sdtm implementation_process _v1

  1. 1. CDISC SDTMImplementationProcessMichael ToddPresidentNth Analytics
  2. 2. 2Overview• CDISC has several initiatives• Only SDTM will have the force of law– The FDA announced plans to requireimplementation of SDTM as a federalregulation• Therefore, near-term, only one will havesignificant impact• We will focus on SDTM in thispresentation
  3. 3. 3Pace of Change• The pharmaceutical industry changes veryslowly– Regulated industry– Two powerful groups: Government / Industrystasis• Slow implementation of SDTM is entirelyexpected• Probability of other CDISC initiativeshaving near-term impact is low
  4. 4. 4Convergence of Forces• Post Vioxx era is unique time when significantforces converge to produce change– FDA: Focus on safety– Industry:• Generic competition, cost pressures, resistance to priceincreases, resistance to me-too drugs, costly, safety-relatedfailures• Increased efficiency becomes important• Perceived benefits from investment in new methodologies• Cost-effective only with widespread implementation
  5. 5. 5FDA: Focus on Safety• FDA has been hammered by Congress andmedia (including blogosphere) for high-profilesafety failures• We see an increased emphasis on cross-compound class effects:– Epilepsy compounds and suicidality– Tumor necrosis factor (TNF) blockers and cancer• Analyses of the sponsors’ entire clinicaldatabase required to address these concerns
  6. 6. 6Regulatory Basis for use of SDTM
  7. 7. 7FDA to Mandate Use of SDTM• Federal Register– Vol. 71, No. 237 /Monday, December 11, 2006 / The RegulatoryPlan / Page 72784– “The Food and Drug Administration is proposing to amend theregulations governing the format in which clinical study data andbioequivalence data are required to be submitted”– “…The proposal would also require the use of standardized datastructure, terminology, and code sets contained in current FDAguidance (the Study Data Tabulation Model (SDTM) developedby the Clinical Data Interchange Standards Consortium) to allowfor more efficient and comprehensive data review.”– 2 year implementation period• A federal regulation has the force of law
  8. 8. 8eCTD Guidance• The eCTD guidance is the current regulatory basis forthe use of SDTM in FDA submissions• FDA Guidance: “Providing Regulatory Submissions inElectronic Format – Human and Pharmaceutical ProductApplications and Related Submissions Using the eCTDSpecifications–– Module 5 Datasets: ‘See associated document “Study DataSpecifications” for details on providing datasets and related files(e.g., data definition files, program files).’• Study Data Specifications–
  9. 9. 9eCTD: Study Data SpecificationsSDTM• Regulatory basis for use of SDTM:– “Specifications for the Data Tabulation datasets ofhuman drug product clinical studies, are located in theStudy Data Tabulation Model (SDTM) developed bythe Submission Data Standard working group of theClinical Data Interchange Standard Consortium(CDISC).”– “This folder is reserved for the datasets conforming tothe SDTM standard.”– “The latest release of the SDTM and implementationguides for using the model in clinical trials is availablefrom the CDISC web site.”
  10. 10. 10eCTD: Study Data SpecificationsDEFINE.XML• “The data definition file describes the format and contentof the submitted datasets.”• “The specification for the data definitions for datasetsprovided using the CDISC SDTM is included in the CaseReport Tabulation Data Definition (define.xml).”• “The latest release of the Case Report Tabulation DataDefinition Specification is available from the CDISC website (”• “Include a reference to the style sheet as defined in thespecification and place the corresponding style sheet inthe same folder as the define.xml file.”
  11. 11. 11eCTD: Study Data SpecificationsAnnotated CRF• “This is a blank case report form annotations that document thelocation of the data with the corresponding names of the datasetsand the names of those variables included in the submitteddatasets.”• “The annotated CRF is a blank CRF that includes treatmentassignment forms and maps each item on the CRF to thecorresponding variables in the database.”• “The annotated CRF should provide the variable names and codingfor each CRF item included in the data tabulation datasets.”• “All of the pages and each item in the CRF should be included.”• “The sponsor should write ‘not entered in database’ in all itemswhere this applies.”• “The annotated CRF should be provided as a PDF file. Name thefile blankcrf.pdf.”
  12. 12. 12SDTM Implementation
  13. 13. 13SDTM Implementation:Costs and Benefits• Staggering cost of compliance– Analogous to Y2K– Will require a federal mandate to achieve critical mass• However, with widespread implementation, industryshould realize substantial cost savings:– ETL tools will largely automate creation of reporting and analysisdatasets– Industry-wide standard databases will enable development ofcommercial reporting and analysis tools• Will shave weeks off timelines– SDTM will greatly simplify data transfer between clients andCROs
  14. 14. 14SDTM Implementation: Phases• Phase I: understand the rules of STDM– that has been done.• Phase II: tool implementation– ongoing• Tools cannot do it all• Implementation requires individuals withdata management and SDTM experience tomap with authority
  15. 15. 15The Role of the Data Manager
  16. 16. 16Data Manager:Key to Mapping Success• Reviews the SAP, Protocol and CRFs• Applies his/her experience with study and CRF design• Uses CDISC SDTM Guidance rules to annotate CRFswith SDTM domains and variables• Decides which SDTM domains are required based onCRF data• Assigns CRF data to SDTM variables• Creates new protocol specific domains when needed• Creates supplemental domains when needed• Participates in validation of SDTM domains
  17. 17. 17CRF Annotation
  18. 18. 18Sample Mapping Spreadsheet
  19. 19. 19Validation: SDTM Structure• SDTM compliance checks– Conformance with Implementation Guiderules*• Metadata structure and domain models• Keys, topic variables• Variable names, labels, type etc. are correct• All required variables have values, etc.• All expected variables are present• All collected relevant and derived data* Study Data Tabulation Model Implementation Guide v3.1.1 – Section 3.2.3 Conformance
  20. 20. 20Validation: Source Data• Verify metadata– Data manager is the expert for this task– Includes a manual review process• Techniques– Independent programming– For each source dataset, verify dataconverted correctly to SDTM for a randomsample of subjects
  21. 21. 21Validation: All Source DataMapped?• Leverage metadata• Use metadata to query source datastructure and determine differences– List of source data variables not mapped– List of source data values not mapped– Source data variables differing from SDTMvariables
  22. 22. 22SDTM: New Job Roles
  23. 23. 23Mapping Specialist• New role, but really an old role in a newtime• The mapping specialist decides how toconvert the raw data into SDTM domainsand variables• Senior position• Historically done by SAS programmers orstatisticians
  24. 24. 24Data Manager/Mapping Specialist:Job DescriptionsCreation of data definition files andCase Report Forms annotated toCDISC SDTMProtocol review regarding what, when,and how data are collectedQuality Control of CDISC SDTMdomainsDevelopment of data specificationsand CRF designDevelopment of CDISC SDTMdomains utilizing a data conversiontoolBuilding, testing and validation ofclinical database using standardizedsoftware packagesAuthoring of mapping specificationsfrom source (raw) data to target(CDISC SDTM)Ensure the completeness, accuracyand consistency of the dataMapping SpecialistClinical Data Manager
  25. 25. 25Mapping Specialist:Qualifications• In-depth knowledge of CDISC SDTMV3.1.1 and V3.1.2 Implementation Guides• Clinical data management experience• Knowledge of ETL tool• Basic SAS knowledge
  26. 26. 26Data Integration Specialist:Job Function• Develop conversion jobs in ETL tool• Write any SAS macros required forcomplex derivations that can’t be handleddirectly in the tool• Execute the job and review log foranomalies
  27. 27. 27Data Integration Specialist:Qualifications• Knowledge of clinical data– Not as much as mapping specialist• ETL experience• Intermediate SAS skills, especially macrolanguage and PROC SQL– Not as much as traditional SAS programmer– ETL tool is doing most of the work– Mostly writing code fragments• SDTM knowledge– Not as much as mapping specialist
  28. 28. 28Summary• Convergence of forces will drive SDTMimplementation• As implementation becomes widespread,organizations using ETL tools will have adecided edge• Data management expertise is critical tosuccessful SDTM automation