Issues with labels in Pinnacle 21 Community ver. 2.1

Labels = headaches

Inclusion/Exclusion Criterion Short Name is exactly 40 characters

With EXTRT, in 3.1.3 it was called Name of Actual Treatment

In 3.2, with the advent of EC (Exposure as Collected), it really should have been kept that name.  Instead we see the following.   

I agree that we should stay true to the guide, but guide in this sense appears wrong and makes no reference to revision history.  

Even In some cases when labels do change AND they include it in revision history it doesn't make sense : 

Updated label for REASND to read "Reason Medical History Not Done or Not Occurred"

But that is 47 characters long and we couldn't possibly mandate that since every validation would fail using SAS xpt.

Bottom line, validation of labels wastes alot of time and resources and is compounded by the fact that they live in both the define.xml and sas transport files with an imposed limit of 40 characters. 

I would bring this up to CDISC and have them ensure labels are consistent in each IG.   For now if you get the error with EXTRT, you can say it is a false positive, but in reality "Name of Actual Treatment" is a better representation now that EC exists.   ECTRT <> EXTRT in all cases.

Labels

I looked it up in the SDTM-IG 3.2 and the label for EXTRT should be "Name of Treatment".

Actually, the purpose of the variable label is to provide the reviewer information about what the variable means and is about. If it MUST be a (case sensitive!) 1:1 match with the variable name, I see no purpose of providing it as the FDA tools can look it up anyway (e.g. using a web service).
So, i.m.o. (and that of many users), labels should get more "freedom".

Here is a blog I wrote about this:

"SDTM Labels: freedom or slavery"
http://cdiscguru.blogspot.com/2015/12/sdtm-labels-freedom-or-slavery.html

SDTM label mismatch for AESEV

and also wrong for

LBSTNRC, Reference Range for Char Rslt-Std Units.

and there are some different cases:

1, when a variable was split into multiple variables, e.g. COVAL1-COVALn, TSVAL1-TSVALn, should the label for all split variables be the same or not?

2. when the split variables are stored in SUPP, should the QLABEL be the same as the main domain?

STUDYID  RDOMAIN  USUBJID  IDVAR  IDVARVAL  QNAM  QLABEL  QVAL  QORIG  QEVAL 
12345  MH  99-123  MHSEQ  6  MHTERM1  Reported Term for the Medical History  2nd 200 chars of text  CRF    
12345  MH  99-123  MHSEQ  6  MHTERM2  Reported Term for the Medical History  last 100 chars of text  CRF    

Thanks!

Labels again

AESEV is being fixed as well.  

This was a mixed up with the EVENTS --sev vs FINDINGS --sev model variables vs the IG implemenations.

Technically there are 15 explicit concepts of "severity".   These 4 x the number of model and IG versions.   Why can't it be defined once where appropriate.  Why is FINDINGS different ?   You will see this for many other variable labels