PhUSE Best Practices Working Group: define.xml versus Reviewer’s Guide

Meeting Date: 2015-07-07

Present:

  • Sandy (Astra Zeneca)
  • Kiran (Roche)
  • Steve (Cytel)
  • Wei (Gilead)
  • Mina (FDA)

Agenda:

  • Admin/Updates
  • Startdefine.xmlreview

Admin/Updates:

  • Pre-meeting Steve asked the team what their companies were naming the SDRG pdf file since the SDRG completion guidelines say to name it study-data-reviewers-guide.pdf.
  • Sandy replied: I have seen it named reviewersguide.pdf, dataguide.pdf, sdrg.pdf (or adrg.pdf)…
  • Wei replied: I have used dataguide for SDTM sndreviewersguide for Adam.
  • Linda Gau replied:Yes, we use study-data-reviewers-guide.pdf.
  • Mina mentioned during the meeting that it may eventually be named sdrg.pdf (and adrg.pdf for analysis data). She believes a Federal Register Notice (FRN) will be coming soon stating FDA’s intent to use the PhUSE templates, but is not sure if there will be a public comment period.

define.xml discussion:

  • Using the V2.0.0 example SDTM file: define2-0-0-example-sdtm.xml (or .html):
  • We started with the Table of Contents. It was straightforward and the discussion centered on the order of domains. Section 3.4.2 of the Define-XML-2-0-Specification.pdf specifies ordering by domain class then alphabetically. We had no further recommendations.
  • We next discussed the handling of long comments/derivations
  • The Derivation/Comment columns had a feature to be able to link to other documents, the examples showed links to complexalgorithms.pdf.
  • We decided that if the text in this column exceeded two lines then the first priority was tolink to the Computational Algorithms section.
  • However if the text in the Computational Algorithms exceeded five lines or needs a graphic then it made sense to directly link to the complexalgorithms.pdf file to the specific page and we recommend noting this in the reviewer’s guide (via a general statement, probably at the end of the overview in section 3.1).
  • The Comments section also provides a place to make non-derivation comments for variables. We also discussed the possibility if these exceeded five lines we could possibly link to the reviewers guide (including the text in an appendix or within the domain-specific section) or to a supplemental file (supplementaldatadefinitions.pdf).
  • Same rules should apply to long comments/derivations at the value level.
  • Origins, particularly Assigned versus Derived. Do any companies have rules for how to assign these?
  • Seems to be determined by the project team.
  • There are guidelines in the SDTMIG (see section 4.1.1.8 in SDTMIG 3.2).
  • Need to be careful in using these according to the guidelines and including comments for anything that was derived.

Next meeting: 2015-07-21