Strategies for Implementing CDISC

3,989 views
3,794 views

Published on

Over the past decade, CDISC data standards have become the FDA preferred method for the data submission. In fact, the FDA is considering a proposed rule change that would mandate the submission of data in CDISC Study Data Tabulation Model (SDTM) and Analysis Data Model (ADaM) formats for all new marketing applications. However, the implementation of this standard has proved to be intimidating to many with only a very small percentage of drug companies actually developing and submitting data in this format.

During the webinar, Thomas Kalfas, an experienced data management professional and CDISC subject matter expert, shared his knowledge and strategies for implementing CDSIC. Topics included a brief review of CDISC, implementation challenges, and insight into the best timing for implementation.

Published in: Health & Medicine, Business

Strategies for Implementing CDISC

  1. 1. 2 0 1 1 FA L L B I O M E T R I C S W E B I N A R S E R I E S Strategies for Implementing CDISCDec. 13, 2011 Presented by Thomas Kalfas
  2. 2. Thomas Kalfas  Director, Global Biometrics Technical Operations  24+ years of technical data management and biostatistical programming experience in pharma/ biotech/CRO industries  Member: CDISC IAB/CAB, CDISC SDS and CDISC Validation teams since 2006  Focus on technical operations, standards development and implementation2011 FALL BIOMETRICS WEBINAR SERIES
  3. 3. CDISC Acronyms/Definitions3  ADaM – Analysis Data Model; statistical analysis data standards  CDASH – Clinical Data Acquisition Standards Harmonization; Case Report Form (CRF) standards  CDISC – Clinical Data Interchange Standardization Consortium; organization advocating global standards for clinical trial data  Define Doc aka Define.xml aka CRT-DD – dataset specifications; a dynamic table of contents for the submission datasets (SDTM and/or ADaM)  SDTM – Study Data Tabulation Model; clinical trial data standards 2011 FALL BIOMETRICS WEBINAR SERIES
  4. 4. Objectives4  Topics will include a brief review of CDISC, implementation challenges, and insight into the best timing for implementation.  We will not be going in-depth into actual conversions or creation of specific CDISC domains, but rather focusing on high-level requirements, issues and feedback from the FDA, common approaches for standards implementation, and our recommendations. 2011 FALL BIOMETRICS WEBINAR SERIES
  5. 5. Today’s Topics5 1) Business Justification for CDISC 2) Current CDISC Status 3) Methods for Producing CDISC Deliverables – Common Errors – Timing Considerations and Best Methods – Recommendations 4) Summary 2011 FALL BIOMETRICS WEBINAR SERIES
  6. 6. Business Justification for CDISC “ The Center for Drug Evaluation and Research (CDER) is strongly encouraging sponsors to submit data in standard form as a key part of “ its efforts to continue with advancement of review efficiency and quality. -CDER, May 20112011 FALL BIOMETRICS WEBINAR SERIES
  7. 7. Origins7 Critical Path Initiative (2004)  Streamline the submissions/review process, shorten the review cycles, decrease costs, and allow for easier data warehousing  FDA asked sponsors to voluntarily use SDTM and ADaM standards for the e-submissions  Powerful tools/software to be developed (based on these standards) to assist reviewers with their evaluations 2011 FALL BIOMETRICS WEBINAR SERIES
  8. 8. Business Justification for CDISC8 FDA encouraging sponsors to continue the learning curve on CDISC standards – No submission will be – Ultimately (~1-2 years), rejected for non- a minimum level of compliance compliance will be expected for all – If compliant, then review submissions, and if not (after training completed met, then these “CDISC- and reviewers have like” submissions would become familiar with be treated as non-CDISC format/tools) should be or “legacy” data (and quicker would take much longer to review) 2011 FALL BIOMETRICS WEBINAR SERIES
  9. 9. Business Justification for CDISC9 Process Improved Cost Value- data efficiencies quality savings added CDISC 2011 FALL BIOMETRICS WEBINAR SERIES
  10. 10. Cost Savings10  Estimated 30% clinical trial efficiencies gained (project startup, cleaning, programming and analysis) – Potentially reduce the study lifecycle by 8 months resulting in savings of approximately $9 billion annually Cycle Time in Months 8 Months Cycle Time Reduction/Trial  Estimated thata restricted implementation of these standards at tail- end submission stage would decrease the potential return by only 60% 2011 FALL BIOMETRICS WEBINAR SERIES Source: Gartner&CDISC (November 2006))
  11. 11. Current CDISC Status  FDA Trends  CDISC Updates  Premier Research and CDISC2011 FALL BIOMETRICS WEBINAR SERIES
  12. 12. FDA CDISC Trends12  Committed to Standards Initiative (2009-2013) and actively working to “refine and maximize utility of CDISC standards”  Both CDER and CBER accepting/requesting SDTM *and* ADaM formatted datasets since December 2010  Number of eSubmissions has increased by ~2K per month since 2010  CDER now tracking number of submissions (SDTM and ADaM) on their website 2011 FALL BIOMETRICS WEBINAR SERIES
  13. 13. FDA CDISC Trends13 Source: http://www.accessdata.fda.gov/FDATrack/track?program=cder&id= CDER- OB-NDAs-BLAs-and-Efficacy-Supplements-with-Electronic-Datasets-Available 2011 FALL BIOMETRICS WEBINAR SERIES
  14. 14. CDISC Updates14  Working closely with the FDA to: – Determine “Supp-qual” data needed to be added to main domains – Have reviewers comfortable with the Implementation Guide (IG) – Build review automation for standard analysis  Providing further guidance to the industry to help navigate gray areas within the standard  Developing therapeutic standards, i.e., supplements to the IG to address specific implementations  Developing a set of Medical Device domain standards  Discussing/clarifying type/location of derived variables (SDTM vs. ADaM) 2011 FALL BIOMETRICS WEBINAR SERIES
  15. 15. Premier Research and CDISC15  Spending more time with our clients discussing the benefits of early-stage CDISC implementation  Premier has developed CRF and DB standards in line with CDASH and SDTM. Use of Premier Standards enables our operational staff (CRF- and DB-developers and programmers) to utilize these standards to help realize the efficiencies anticipated by Gartner in their projections of industry savings.  Number of early-stage CDISC (SDTM) projects for 2011 has tripled that of 2010 (~40% of our active projects) 2011 FALL BIOMETRICS WEBINAR SERIES
  16. 16. Premier Research and CDISC16  Spike in requests to add SDTM as a requirement for existing “legacy” projects  Also seeing an increase in requests for: – Early-stage feeds into a ISS/ISE-like datamart – CDISC training – Consulting services for implementation of CRF, DB and programming standards for existing and new clients 2011 FALL BIOMETRICS WEBINAR SERIES
  17. 17. Methods for Producing CDISC Deliverables  Common Errors  Timing Considerations and Best Methods  Recommendations2011 FALL BIOMETRICS WEBINAR SERIES
  18. 18. Common Errors (per FDA)18  “SDTM-Like” submissions  Non-compliant define.xml  Traceability issues  Define doesn’t validate  Invalid ISO 8601 date format  Required variable not found  Inconsistent value for standard units  Invalid value for MedDRA Term  All dates in the SDTM domains must conform to the ISO 8601 format  Begin Date must be ≤ End Date (e.g., CM or AE start dates that come after the end dates)  For a given test, all values of --STRESU should be the same. In some cases - -TESTCD may not be sufficient to uniquely indentify a test For a full list, go to:  http://www.fda.gov/downloads/Drugs/DevelopmentApprovalProcess/Form sSubmissionRequirements/ElectronicSubmissions/UCM254113.pdf  http://www.fda.gov/BiologicsBloodVaccines/DevelopmentApprovalProcess/ ucm209163.htm 2011 FALL BIOMETRICS WEBINAR SERIES
  19. 19. Common Methods19 Study Close DB Start-up Study Conduct Out Lock Late-Stage Conversion Early-Stage Mid-Stage Implementation Conversion 2011 FALL BIOMETRICS WEBINAR SERIES
  20. 20. Common Methods: Late-Stage Conversion (1)20 Usually as part of an NDA submission, but now being requested for individual or groups of studies earlier than the traditional NDA activities  Cost-effective? Might seem so, but…  Compliance issues common as this could take place months or even years after the study had completed (dependent on the quality/compliance issues introduced at the protocol and CRF layers)  Traceability issues (it is critical that the data is traceable from the CSR to the analysis datasets to the SDTM datasets to the raw datasets to the CRFs) 2011 FALL BIOMETRICS WEBINAR SERIES
  21. 21. Common Methods: Late-Stage Conversion (2)21  DB is locked!!!  Cleaning of DB may not have incorporated all SDTM compliance checks, e.g., start dates must be less than or equal to stop dates, etc.  Aside from the database, what else has been produced that will now need to be reconciled against the “new” SDTM datasets? Analysis datasets? TLG’s? ISS/ISE? CSR?  The further along you are with the study, the more work necessary to ensure traceability 2011 FALL BIOMETRICS WEBINAR SERIES
  22. 22. Common Methods: Late-Stage Conversion (3)22 Need to: Determine need for standardization of values/units, e.g., labs, and coding dictionaries (if multiple studies, should have same version of dictionary) Completely document steps needed to produce SDTM and ADaM…you’ll need not just for the conversion, but also for the Define Docs Fully assess risk/impact of any review/compliance findings on downstream deliverables Documentation is critical (Annotated CRFs, Dataset Specifications, Code Lists, Change Logs, Define Docs) Reproduce numbers from your CSR! Need to show that the analysis/results can be reproduced from SDTM 2011 FALL BIOMETRICS WEBINAR SERIES
  23. 23. Common Methods: Late-Stage Conversion (4)23 Need to (continued): Maintain Change Logs to track programming changes (in addition to maintenance of your “living” documents, i.e., SDTM Dataset Specifications -> Define Doc Reproduce numbers from your CSR! Need to show that the analysis/results can be reproduced from SDTM – If significant issues, then need resolution – If differences can be explained, then consider adding text explanation to the “Reviewer Notes” portion of the Define Doc Same as above for your compliance checks, i.e., any oddities need to be explained in the reviewer notes 2011 FALL BIOMETRICS WEBINAR SERIES
  24. 24. Common Methods: Mid-Stage Conversion24 Usually as a late CDISC consideration in preparation for NDA activities  Same issues/challenges as for the Late-Stage Conversion; however…  DB is active, not locked!  So, while there are most likely still conversion challenges with non-standard CRF/DB setup, it is now easier to address compliance/review findings as part of you normal DM/cleaning processes! 2011 FALL BIOMETRICS WEBINAR SERIES
  25. 25. Common Methods: Early-Stage Implementation (1)25 Standards compliance from the very start  Protocol: – Controlled Terminology, e.g., AE Severity  CRFs: – CDASH: for items not covered by SDTM directly, e.g., date component fields – SDTM: want to get as close to 100% SDTM compliance as possible from the CRFs and DB – Controlled Terminology → reduces the need for additional programming/mapping/conversions (and QC) 2011 FALL BIOMETRICS WEBINAR SERIES
  26. 26. Common Methods: Early-Stage Implementation (2)26  Database: – Standard modules – Standard checks (cleaning and compliance) – Standardization of values/units, medical coding, etc.  Programming: – Still requires thorough documentation (Annotates, Specs, Define Docs), but development/maintenance is much easier as standard templates can be used – Standard programs refine the CDISC-like (both SDTM and ADaM) datasets into 100% compliant datasets – Standard programs used to load into ISS/ISE, produce standard TLGs, etc. 2011 FALL BIOMETRICS WEBINAR SERIES
  27. 27. Common Methods: Early-Stage Implementation (3)27  Efficiencies are realized due to standards adherence  Traceability is inherent within this process due to a more traditional SDLC methodology (i.e., specs first, then development, then validation, then these feed into the next deliverable, e.g., ADaM, ISS/ISE, and TLGs)  Value-added: reliance on standards, allows for CDISC outputs to be produced earlier in the study “life-cycle” and made available for data warehousing, data mining (as ISS/ISE) and DSMB/DMC requirements 2011 FALL BIOMETRICS WEBINAR SERIES
  28. 28. Summary  CDISC SDTM & ADaM standards are gaining traction  Requirements for CDISC datasets getting stronger  While late- and mid-stage conversions can and will continue to be done, the FDA cautions against this in favor of early-stage CDISC standards implementation  Early-stage implementations of CDISC not only allow for efficiencies to be realized, but also make value- added scenarios possible2011 FALL BIOMETRICS WEBINAR SERIES
  29. 29. Webinars Series29 Listen to past webinars: ▪ The Role of Data Monitoring Committees Speaker: Ron Kershner, Ph.D. ▪ IVR/IWR…More than just Randomization Speaker: Ryan Michaud ▪ Streamlining Data Management Start-up Speaker: Cheryl Silva 2011 FALL BIOMETRICS WEBINAR SERIES
  30. 30. Questions? Thomas Kalfas Director, Global Biometrics Technical Operations Telephone: 847.420.2622 thomas.kalfas@premier-research.com2011 FALL BIOMETRICS WEBINAR SERIES

×