Dear Adam,
probably a better forum for this questions is the LinkedIn "SDTM Experts" forum (https://www.linkedin.com/groups/2758697/) as all the SDTM gurus are on that one. So you might also want to post it there too.
Regarding the matter itsefl, I would also prefer option 2.
In principle, SDTM should not be about derived data, but unfortunately the model deteriorated over time "for ease of review". Baseline flags should in principle not be in SDTM, but reviewers wanted them, so they were added. As such, I see no harm in putting the average as a separate record, and flag it with "baseline". It is not because it is not on the CRF that it cannot appear in SDTM. There are many other examples (such as the --DY variables) that do not appear on the CRF but that are in SDTM anyway.
Best regards,
Jozef
My baseline value for analysis is an average. The average is not collected on the CRF. In my SDTM dataset, should I
1 - label all records taken to compute the average with --BLFL = Y.
OR
2 - derive the average as a separate record and label only this one with --BLFL = Y and --DRVFL = Y.
I have the feeling that option 2 is better. I also have to calculate averages at the different time points and having a record derived for each of these averages appears to make life easier for a reviewer.