Forums: Validation Rule Suggestions
Hi!
As of today the actual algorithm for SD0026 rule excludes any non-numeric --ORRES and '* Ratio' --TEST records from the validation of --ORRESU value. Therefore example from page 118 of SEND IG is OK. Do you have any particular issue with this check for PM domain?
SD0026 check is under FDA workgroup discussion. See http://www.phusewiki.org/wiki/index.php?title=Top_20_Validation_Rule_Failures_(CBER)
I think, that you brought another very good point about validation rule development, that a check assignment shoulld be the same level of importance as the check itself.
Best Regards,
Sergiy Sirichenko
This means that, with the latest config.xml file from SVN, we can have the SD0026 rule enabled for pm.xpt and not report warnings for situations like:
PMTESTCD = "DESCR"
PMORRES = "cyst"
PMORRESU = ""
Is that right?
Hi!
In your example --ORRES variable has a non-numeric value. The SD0026 check should skip such records and no warnings are expected. That approach has been implemented a while ago in previous versions of the validator.
Do you recieve Warning messages?
Thanks,
Sergiy
We just tested with the latest version of config.xml and I can confirm that SD0026 is no longer reported for PM with non-numeric --ORRES, which is great (definitively better than disabling the rule for this particular domain).
We'll check anyway SD0026 for other domains in the same situation (e.g. SC - see http://www.opencdisc.org/forum/se0015-missing-stresn-value-and-sd0026-missing-units-value).
Thanks!
I feel that the following rule should be disabled for PM (Palpable Masses):
SD0026: "Missing value for --ORRESU, when --ORRES is provided".
PMORRES can be populated while PMORRESU remains empty - see example from SENDIGv3.pdf - page 118.