QSEVALID value not found in 'Medical Evaluator' extensible codelist

See SDTM IG 3.3.

Hi Deepika, 

See SDTM IG 3.3.

--EVAL variable is assigned to (EVAL) CT

--EVALID variable is assigned to (MEDEVAL) CT

--EVALID is used to distinguish multiple evaluators with the same role recorded in --EVAL. Examples: RADIOLOGIST1 or RADIOLOGIST2

Evaluator initials and other supportive information should be stored in SUPPQUAL datasets.

Kind Regards, 

Sergiy

I checked the SDTM IG 3.3…

Thanks for pointing me to IG3.3. I checked the SDTM IG 3.3 and below is what it says which means EVALID can also be mapped to Rater's Initials when used with clinical classification data. Also QRS - the Hamilton Anxiety Rating Scale also says you can map Initials to QSEVALID. If we use both EVAL and MEDEVAL Controlled Terminology to map QSEVAL - can we represent this in Define.xml?

The Evaluator Identifier variable (RSEVALID) can be used in conjunction with RSEVAL to provide additional detail of who is providing the assessment. For example RSEVAL = "INDEPENDENT ASSESSOR" and RSEVALID = "RADIOLOGIST 1" may further qualify the RSEVALID variable may be subject to controlled terminology but may also represent free text values depending on the use case. When used with disease response data, RSEVALID is subject to MEDEVAL controlled terminology. When used with clinical classification data, RSEVALID = "INDEPENDENT ASSESSOR". RSEVAL must be populated when RSEVALID is populated. The RSEVALID variable may be subject to controlled terminology but may also represent free text values depending on the use case.

1. When used with disease response data, RSEVALID is subject to MEDEVAL controlled terminology.
2. When used with clinical classification data, RSEVALID may represent free text values such as rater's initials.

QSEVALID value not found in 'Medical Evaluator' extensible codel

This is a severe misinterpretation of the standards.
Neither SDTM-IG 3.2 nor 3.3 mention QSEVALID. Some other domains implement --EVALID (like TU,TR) and mention the codelist MEDEVAL.
In case a variable like QSEVALID is not mentioned in the SDTMIG, and one wants to use it, one needs to look into the SDTM Model (v.1.4 or 1.7). The SDTM model does not make any statement about a specific codelist to be used with --EVALID variables.
It is an incorrect interpretation of the standard to assume that when a codelist needs to be assigned to a specific variable like TUEVALID, that then the same codelist needs to be assigned to another, similar variable of which the SDTMIG does not say anything (like QSVALID). Such an interpretation is inherently wrong.
Furthermore, the define.xml is of utmost importance here. If in your case, if you do not assign a codelist to QSEVALID in the define.xml(which is your good right), the validator should see this and accept your decision. If you do assign a codelist to QSEVALID in the define.xml (your own or MEDEVAL, which you then probably extend), you should of course take care that it is properly populated. Once again: "the define.xml is the sponsor's truth about the metadata".

P.S. Also the combination --EVALID - MEDEVAL has some problems. "RADIOLOGIST" is surely not an ID, it is a role. When the radiologist "Raymond" goes home and cooks for his family, then he is not a radiologist anymore, but plays the role of "COOK", He then still has the ID of "Raymond". This difference between ID and role is something the CT team does not seem to have understood.

Jozef Aerts
XML4Pharma and CDISC volunteer

I checked the SDTM model 1.4…

I checked the SDTM model 1.4 and --EVALID is part of findings observation class. I agree Define should represent what is done and not necessarily what is expected to be done.

Thank you for your response. I appreciate it.