t Troy
on

 

Hi,

SD0007 (Inconsistent value for standard unit) is currently flagged as an Error by the validator.  However, there are legitimate cases where there can be multiple units for a given test in a domain, such as the LB domain in SEND (one sample case is where differing values of LBCAT can have different standard units for the test within the LBCAT; but there are also other cases like LBMETHOD or a random SUPP variable are what keeps them unique).

Please change this rule to a warning.

Forums: SEND

s Sergiy
on August 21, 2014

Hi Troy, 

SD0007 check works to insure consistency in Standard Units within the same combination of Test, Category, Specimen and Method.

This is our understanding of Standard Units.

I've never heard about using SUPPQUAL data to define additional "test qualifiers".

Kind Regards,

Sergiy 

t Troy
on September 25, 2014

The rule is a good rule to have; we are only arguing that it should be a Warning, not an Error.

SD0007 check works to insure consistency in Standard Units within the same combination of Test, Category, Specimen and Method.

There are legitimate additional variables in addition to the 4 you've listed that can constitute the "key" of consistent standard units.  I may have complicated the point by bringing up SUPPQUAL - ignore that.  A much easier case to focus on is --SCAT (subcategory) - this is another categorization variable where different --SCAT cases should be allowed standard different units within them.

 

The point is that:

  1. There is at least one valid case where more variables are needed than the 4 currently listed in order to define consistency
  2. There isn't an explicit rule in the SDTM or SENDIG that units be consistent within the variables you've listed

 

Because of that, SEND members are asking for a loosening of this rule into a Warning, so that valid datasets do not fail validation.

 

a Ana
on February 12, 2015

Hi Sergiy 

AS I understand this rule to be very valid for datasets like LB, I am getting this ERROR for the DA dataset where the subjects are dispensed both tablets and Capsule at the same visit so their test cd is DISPAMT and since there is no DACAT specific since thi sis just study medication dispensing records, nwo I am getting errors since DASTRESU is TABLET and CAPSULE.  

If it is WARNING, we might jsut explain but since this is an ERROR, are we to manipulate the DACAT to be different?

Thanks

Ana

j Jozef
on February 13, 2015

I.m.o. this rule is just damned wrong.
I have explained this in a blog entry which can be found at:

http://cdiscguru.blogspot.com/2015/02/rule-fdac084-is-just-damned-wrong.html

For example, for the combination LBTESTCD=GLUC, LBCAT=URINANALYSIS, LBSPEC=URIN, LBMETHOD=TEST STRIP, you can get results either as "1+", "2+", "NEGATIVE" (so no units), as a concentraion in mmol/L or as a concentration in mg/dL, depending on which test (by LOINC code) exactly was done.

You might be able to standardize the latter 2 to the same unit (using the molecular weight of glucose) but you can't for the first. So if you have one lab having done the test qualitative/ordinal, and the other lab having done the test quantiatively, you will get a lot of false positives.

I haven't found this rule in the SDTM-IG either.

So I would suggest to downgrade the message from "ERROR" to "INFO", as it can be an inconsistency, but hasn't to be one.

More info can be found in the blog.

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.