Engine FDA (2405.2) Potential Validation Bugs

Pinnacle 21 Product Suggestions Validation Rule Suggestions
1

Hi there,
I just wanted to bring to attention that recently performed some Validations on SDTM-IG 3.2 (FDA) as well as SDTM-IG 3.4 (FDA) where encountered 2 validation findings which I would consider bugs.

(1) CT2002: PRDECOD value not found in ‘Prodedure’ extensible codelist
→ IG 3.4 says “Standardized or dictionary-derived name of PRTRT. If the codelist “PROCEDUR” is not used, the sponsor is expected to provide the dictionary name and version used to map the terms in the external codelist element in the Define-XML document. If an intervention term does not have a decode value, then PRDECOD will be null.” Whereas IG 3.4 additionally declares “(PROCEDUR)” as expected codelist, IG 3.2 does not (only *). However, in both cases, the finding is displayed, although Define.XML unambiguously links this variable to MedDRA dictionary.

(2) DD0116: FATESTCD/FATEST mismatch in Codelist ‘Findings About Test Code’
→ The validator complains about FATESTCD = OCCUR and FATEST = Occurrence Indicator. I understand that that the basic --OCCUR variable has “Occurrence” as label, however, there is a registered SDTM CT entry for OCCUR/Occurrence Indicator where I would assume that the CT information prevails over the “default” approach based on the variable names/labels.

Kind regards,
Harald

2

Hi Harald,

I’ve responded to your inquiries below:

  1. The variable in question (PRDECOD) was assigned to the PROCEDUR Codelist in the 2405.2 Engine Suite. This change was made in SDTMIG v3.2, SDTMIG v3.3, and SDTMIG v3.4. The Codelist was added to the variable as part of a decision to add Codelists that were specified on certain variables in later IGs, to the same variables in older IGs, even though it has not been explicitly specified in those versions. This was done for internal variable or cross-standard consistency on using the same controlled terminology on the same variables. Figuring that clients are generally updating to newer CT as they go. The codelist references added for variables in v3.2 and v3.3 is also supported by CDISC’s CT Relationships document where codelists not listed in the SDTMIG are being backfilled for older versions of SDTM. CT2002 will check the PROCEDUR codelist within the configured CT version and if the value used does not exist as a term there, the rule will report an issue. If the value is an appropriate extension of the codelist, this can be explained in the Reviewers Guide.

  2. In the SDTM v1.4/SDTMIG v3.2 and earlier, the variable label for -OCCUR is simply ‘Occurrence’. The guidance to use the existing variable name and label in FATESTCD/FATEST is listed in FA section of the SDTMIG in Section 6.4.3 - Variables Unique to Findings About. So, if you are using an existing variable name/label in FATESTCD/FATEST, the corresponding variable name/label from the SDTM Model should be used, in this case from SDTM v1.4 and earlier. This is also covered by the CDISC rule, CG0173. The label for the -OCCUR variable did change to ‘Occurrence Indicator’ in SDTM v1.7/SDTMIG v3.3. If your Study is using SDTMIG v3.2 in this case, this is the reason that DD0116 fired and this is expected. The situation here is simply the ever-evolving nature of the standards from version to version. The change to the label of ‘Occurrence’ to ‘Occurrence Indicator’ is one of those cases where the label is being changed between versions but the rules for FA in Section 6.4.3 remain the same. If you chose to move to SDTM v3.3 and v3.4, this issue will no longer occur - in the meantime, you can explain that you’re using SDTM v3.2 which expects a different label in this case.

1 Like