Hi Jozef, Thanks for reporting this issue! CDISC CT 'ECG Result' (C71150) is probably the most notorious one for its non-practical implementation. Having experience with several hundreds of different studies I’ve never seen any examples of usage this CDISC CT for EGSTRESC. Today OpenCDISC validator applies (EGSTRESC) CT on Variable level, rather than on Value Level. Such approach produces almost 100% of false-positive messages. The old CT0012 check ignored all numeric values to reduce amount of false-positive messages. However as I understand the referenced 'ECG Result' (C71150) CT is relevant only for particular records. CDISC definition as "Terminology codelist used with ECG Findings and Abnormalities within CDISC" is not helpful to identify applicable ECG Tests, Category, etc. My preference would be to remove 'ECG Result' (C71150) CT from validation process until CDISC team clarifies its intended usage. Kind Regards, Sergiy
Thanks Sergiy,
I think indeed the old implementation CT0012 was more appropriate.
This is a rule that (with some limitations) can easily be implemented in XQuery, or using the following algorithm:
- read the value
- check whether it is a number
- do a lookup in the define.xml whether a codelist is associated, and whether the value is in the codelist (the value can also be an integer)
- step 3 can be skipped when the datapoint is a non-integer, floating point number
- when not in the codelist, and not a floating point number, give out an error message
We are currently working on RESTful web services for such rules that are on the value level. We did already do that for VSPOS and parially also for EGSTRESC. You can find the information at:
http://www.xml4pharmaserver.com/WebServices/index.html
With best regards,
Jozef
Using OpenCDISC 2.0 Community Edition, configuration SDTM-IG 3.1.2, using Dataset-XML files and a define.xml 2.0 I get an Warning for almost every record "EGSTRESC value not found in 'ECG Result' extensible codelist" - CT2002 / FDAC341 "EGSTRESC value not found in 'ECG Result' extensible codelist".
The records that produce the error are all EGSTRESC values that represent a numeric value.
I looked into the SDTM-IG 3.1.2 but and indeed states that there is codelist (!!!) but also states "Contains the result value for all findings, copied or derived from EGORRES in a standard format or standard units. EGSTRESC should store all results or findings in character format; if results are numeric, they should also be stored in numeric format in EGSTRESN."
Looking at the examples in the SDTM-IG, there are many EGSTRESC values that represent a number.
With best regards,
Jozef