AGE, AGEU validation?

b Bedeoan

on October 26, 2012

At the validation of a dataset in which dm.xpt does not contain any AGE, AGEU, AGETXT columns, I get the following warning:

"AGEU SE0057 SEND Expected variable not found".

Note that AGE and AGETXT are Permisible in SEND3 specs, while AGEU is Expected. Why is the tool complaining about a missing "Expected" variable?

Also... if we want to make sure that we have AGE/AGETXT and AGEU always... shouldn't complain first about the AGE? Or, if the AGE is missing, AGEU should not be required to be populated...

Forums: Validation Rule Suggestions

s Sergiy

on October 26, 2012

Hi! 1. The Expected variables

Hi!

1. The Expected variables must be included into domains, even if information was not collected. This is one of basic requirements for SDTM Model compliance.

2. In SEND model either of two variables AGE or AGETXT can be used to store study subject age information. The usage of those two variables depends on your particular study design. You may collect age as one single numeric value (AGE), range value (AGETXT) or even both ways (AGE and AGETXT) for different subjects in your study. To achieve flexibility in usage of two similar variables the CDISC team made them Permissible.

OpenCDISC has special checks to handle right usage of AGE and AGETXT variables:

SE0016 "Both AGE and AGETXT are populated"

SE0017 "Missing values for both AGE and AGETXT, when AGEU is provided"

SD0018 "Missing value for AGEU, when AGE or AGETXT is populated"

I agree that we probably need to add a special check that at least one of two AGE or AGETXT Permissible variables (columns) should be included in DM dataset.

Regards,

Sergiy

b Bedeoan

on October 26, 2012

Thank you, Sergiy, for your

Thank you, Sergiy, for your fast reply (as always :) )...

By "The Expected variables must be included into domains, even if information was not collected." you mean that expected variables should be present in both define.xml and corresponding .xpt, is that correct?

As an additional observation: if both AGE and AGETXT are Permisible, why AGEU is Expected and not just Permissible in SEND specs? Wouldn't make more sense to have all 3 variables have the same "optionality" level?

s Sergiy

on October 26, 2012

1. Yes, you are correct.

1. Yes, you are correct. Expected variables must be presented in data and define.xml regardless your data collection process. It's strongly recommended to collect Expected variables. However the CDISC team realizes, that it's not always possible.

2. Consider a group of two (AGE and AGETXT) variables as one combined Expected variable. Not in opposite direction. AGE is an Expected variable. When SEND introduced AGETXT, the only technically available way to handle this case was to make a downshift of AGE/AGETXT Core value to Perm. The ADaM standard has a concept of "Conditionally Required" variables, which can be applicable for AGE/AGETXT variables.