• Decrease font size
  • Return font size to normal
  • Increase font size
U.S. Department of Health and Human Services

Vaccines, Blood & Biologics

  • Print
  • Share
  • E-mail

Supplemental Information for Planning A CDISC Formatted Submission

Note:  This page is under revision (Feb. 2014).

  1. SDTM
    1. All data that is collected is expected to be submitted.  The definition of permissible variable is one that should be used when collected or derived.  "The Sponsor does not have the discretion to not submit permissible variables when they contain data (SDTM IG v3.1.2, sec. 4.1.1.5)".  (Update 12/22/10)
    2. Trial Design datasets provide a standard way to describe the planned conduct of a clinical trial.  At a minimum, the Trial Summary (TS) domain should be submitted whenever possible.  Additional variables not listed in Appendix C3 of the SDTM IG v3.1.2 may be added as long as they are explained in the define.xml (Update 12/22/10)
    3. Subject Elements(SE) and Subject Visits (SV) desribe actual elements and visits for each subject.  These datasets should be submitted whenever possible. This is especially important when the trial includes multiple treatment periods. (Update 12/22/10)
    4. Baseline flags are defined (Update 12/22/10):
      1. As described in the analysis plan.  If there is no definition, then
      2. As defined in the protocol.  If there is no definition, then
      3. The last observation prior to experimental intervention 
    5. Baseline flags should be documented as such in SDTM and if the flag is derived it should be identified as such and explained in the define.xml (Update 12/22/10)
    6. SUPPQUAL is a dataset domain in SDTM.  It is intended to include data variables that are not specified in SDTM.  Discussion with the review office is required to ensure that variables are placed in the appropriate domain rather than using SUPPQUAL as a catch-all for those variables the sponsor/applicant is unsure how to fit into an existing data domain. (Update 8/12/10)
      • This means that although variables may not be included in the published domain they may be added to the domain if they are from the same observation class (interventions, events or findings)
      • For example, the location of a rash has been collected but the --LOC variable is not part of the published CE domain.  The variable --LOC from the Events Observation Class table 2.2.2 in the SDTM standard allows for this variable to be added to the CE domain rather than create a CESUPPQUAL dataset
      • The same rules apply for timing or identifier variables published in the standard; they can be added to the published domain rather than creating SUPPQUAL datasets.  This is the preferred method of reporting
    7. If no data is collected for a specific domain, annotate on the CRF and the define.xml but do NOT submit an empty dataset (Update 8/12/10)
    8. Identify efficacy data location on the SDTM Planning Checklist (Update 8/12/10)
    9. Required variables must be present and populated in SDTM datasets.  These variables cannot be null for any record.
    10. Expected variables must be present whether they are populated with data or not.  This means if data was collected it should be reported otherwise the record will be null
    11. Dates must conform to ISO 8601 formats.
    12. Begin date MUST be less than or equal to end date.
    13. Sponsors should follow section 3.1.2 of the SDTM Implementation Guide for general conformance requirements.
    14. CDISC SDTM is an evolving standard and there may not be clear direction in the current implementation guide on how certain clinical trial data should be represented.  In circumstances such as this, sponsors/applicants should discuss their proposed solution with the review office.
    15. The SDTM Implementation Guide allows for the the creation of custom domains if the data do not fit into an existing domain. Creation of custom domains should be discussed and planned for with the review office.  Follow the recommendations in the SDTM Implementation Guide section 2.6 for creation of custom domains.
    16. Do not include imputed data in SDTM datasets. Imputed data should be placed in the analysis dataset.
    17. Ensure the values for standard units are consistent
  1. define.xml
    1. The style sheet for the define.xml MUST be located in the same folder as the define.xml in the eCTD (Module 5) to enable the view and navigation of the define file.
  2. Controlled Terminology
    1. CDISC controlled terminology can be located via the link for the NCI Enterprise Vocabulary Services (EVS)
  3. Analysis Data Sets
    1. Analysis datasets should be derived from the clinical tabulation datasets that are submitted as part of the regulatory submission (Update 12/22/10)
      1. Sponsors may submit the program files used to create ADSL from the source datasets to provide evidence of traceability. (Update 12/22/10)
    2. Analysis datasets are always required unless approved by the Review Office (Update 8/12/10)