DM Domain - Error 'RACE value not found in 'Race' non-extensible codelist' - Warning versus Error

s Sam

on September 4, 2019

Hi All,

Sorry to ask this question - I did check the existing content but could not find a similar question.

Anyway, in August, we ran community version 2.2 (FDA validation rules) on our data and received the following error:

'RACE value not found in 'Race' non-extensible codelist' (Pinnacle 21 ID = CT2001, Publisher ID = FDAC340)

This is fine and I've seen before. However, in September, we ran the same data with the same metadata (Community version 2.2, same dictionary versions, same CT version etc) and this time we received the same message but as a WARNING rather than an ERROR:

'RACE value not found in 'Race' non-extensible codelist' (Pinnacle 21 ID = CT2002, Publisher ID = FDAC341).

Does any know why this may occur/any previous experience?

Thanks!

Forums: SDTM

Trevor

on September 5, 2019

P21

Hi, This is actually one of…

Hi,

This is actually one of the most commonly occurring instances of CT2002 based on our metrics. I'll do my best to explain. The SDTM Implementation Guide states: RACE is subject to controlled terminology found in non-extensible codelist named RACE (C74457). The problem with this is that there are discrepancies between the values within the non-extensible codelist and the SDTM IG. For example:

The IG states that if multiple races are collected, RACE should be MULTIPLE and the additional information will be included in SUPPDM.
The IG states that if race was collected via an "Other, specify" field, RACE should be OTHER and the details will be included in SUPPDM.
The IG states that if a subject refuses to produce race, RACE should be UNKNOWN.

Of these 3 special cases described in the IG, only UNKNOWN is covered as a valid term based on the RACE codelist. This leaves MULTIPLE and OTHER as problems...

Pinnacle 21 addresses this by treating RACE as extensible in our own terminology file. In addition, we grade CT2002 as a Warning as opposed to Error.

I would suggest you explain this conformance rule in your Reviewer's Guide and explain what value or values of RACE are triggering the rule. It may just be as simple as explaining that the SDTM IG and codelist disagree with each other.

I hope this helps,

Trevor
Product Manager

j Jozef

on September 6, 2019

'RACE value not found in 'Race' non-extensible codelist'

As Trevor already explained, this is due to a discrepancy between the SDTM-IG and the CDISC Controlled Terminology (CDISC-CT).
The background is a lack of agreement between the two development teams within CDISC. The CDISC-CT team states that "multiple" and "other" semantically are not races, and so does not want to add them to the list (well "unknown" isn't either).

Personally, for codelists, I am in favor of a pragmatic approach, as also HL7 is doing, i.e. to include all possible and allowed values in codelists.

And anyway, "Race" is a very questionnable concept: we are all citizens of the world ...

Jozef