SD1021 for continuation variables COVAL1 COVAL2 ..., and TSVAL1 ... checks

  Hi Joan,  1. We follow a

Hi Joan, 

1. We follow a general recommendation from SDTM/SEND IG documents:

“When splitting a text string into several records, the text should be split between words to improve readability.”

(SDTM IG #4.1.5.3.2  page 51 or SEND IG #4.5.3.2 page 37)

2. Yes, there is no info about labels for COVAL1, COVAL2, … variables.  As a temporary solution we use the same label as specified for COVAL variables: “Comment”. I would prefer “Comment 1” or “Comment2” instead. But we decided to not add anything to SDTM standard.

Best Regards,

Sergiy Sirichenko

splits >200 char COVAL1, ...

Hi Sergiy

The page 51 split instruction is about text being split to different records, when it's split between parent domain record and SUPP-- record.  Rather than adjoining columns.  Should the initial leading blank condition apply for this situation?

For the labels, thanks for the information on that.

Best,

Joan

Different handling of this with and without the define.xml

Using the configuration file for *Amendment 1* (latest update posted to svn area), I'm seeing now that when I've selected no define to input to the validation of xpt files, then the COVAL1 - COVAL4 continuation leading space character does not create an issue with validation; the same is true for continuation TSVAL1.

But only when a define (attributes of datatype=text; length=200 for these fields) is submitted along with the xpt files being validated, then the validator complains with warning.  What do you think?

Thanks very much,

Joan

splitting between words still leaves space between words problem

Hi, 

I think splitting between words still do not solve the issue of the blank space, let's say for example "...patient said it forgot to take the pill..." and the first letter of 'said' is already at 198 characters, but for readability we can:

- Split after "it " and preserving the trailing space, then COVAL1 starts with "said" with no leading space

- Split after "it" without preserving the trailing space but still splitting between words, then COVAL1 starts with " said", i.e, with a leading space.

Removing the space in between the words shouldn't be an option as that would mean the comment is 'patientsaid' which is not correct either. So what's the approach taken in this case? Usually I tell the programmers to follow the first option even if it leaves a trailing blank space in COVAL as trade off of having COVAL1 without a leading blank space. Not to mention this could be more complicated if next to the blank space we add the possibility of punctuation marks

splitting between words

And then some people still swear that we should keep using SAS-XPT and not move to XML ... (where there is no 200 character limitation).

Leading blanks

I think we do not need leading or trailing blanks. The application that combines the COVAL, COVAL1, COVAL2, ... variables should just insert a blank between 2 parts.

But most of all we need a modern transport format.