s Shiming


I have an external file of calculated SF-12 score. It cannot be added to QS because the input for that file is from SDTM QS, so adding the calculated score to QS will cause recursive programming. Is it ok if I generate a custom domain for the calculated score? I am not sure whether I can do that because this external file has the same structure as QS.  and the SDTMIG 3.2 said we can only generate a custom domain if none of the existing domain satisfy the need. 

Image removed.

Forums: SDTM

j Jozef
on October 29, 2022

There is a huge amount of information about how to treat total scores on the CDISC website, especially in the SDTMIGs, and the QRS pages. I propose that you read that first.
You statement "it cannot be added to QS because the input for that file is from SDTM QS ... and will cause recursive programming" is not valid: there are so many variables in SDTM that use information from SDTM itself, for example, all the --DY variables that use information from DM (RFSTDTC). Also, for example the --LOBXFL flag variables (last observation before first exposure) (unfortunately) requires recursive programming. And in QS itself, the FDA requirement for "logically skipped questions" may also require recursive programming or a post-programming step.

I don't know how you generate your SDTM datasets, but there are some excellent software packages on the market that take care of all this, including calculating and/or adding total scores. Some of these software programs are extremely user-friendly, mostly using drag-and-drop and wizards, with a little amount of programming and no knowledge of SAS required.

If your "custom" dataset has the same structure as the QS dataset, and you cannot do any "inserts" into QS, an alternative is to generate a second instance of QS that has the total scores, and then merge it back into the original QS domain.

Important: Failure to be able to program something can never be an excuse for a custom domain!

on November 8, 2022

Hi Shiming, 

In general, all derived information should go to ADaM data rather than SDTM.

Kind Regards,

j Jozef
on November 9, 2022

With all respect, but Sergiy's statement is not entirely correct.
Indeed, derived data should essentially go into ADaM, but there are so  many exceptions in SDTM where the data is derived. Reason for this is that FDA reviewers wanted to have these in SDTM (although they are derived and essentially should go into ADaM) for "ease of review".
For example, all "DY" variables (--DY, --STDY, --ENDY) in SDTM are clearly derived, not collected. The record sequence numbers (--SEQ) are derived or assigned, not collected.
Another example is the --LOBXFL (Last Observation Before Exposure Flag) which is clearly derived, even requiring a post-processing step.
For questionnaires, the SDTMIG is clear on that some records may be derived, such as "total scores".
From the SDTMIG v.3.3 page 27:
"For example, some records in a Findings dataset such as QS contain values collected from the CRF; other records may contain derived values, such as a total score."

With best regards,

SDTMIG v.3.3 page 27

Want a demo?

Let’s Talk.

We're eager to share and ready to listen.

Cookie Policy

Pinnacle 21 uses cookies to make our site easier for you to use. By continuing to use this website, you agree to our use of cookies. For more info visit our Privacy Policy.