New rule for --DTC before --ENDTC?

Hi Amelia,  Thanks for your

Hi Amelia, 

Thanks for your suggestion! It will be a useful check.

However we need to limit it to Findings domains only. In Events and Intervention domains there is no standard definition for --DY, DTC variables. Therefore their content may not be related to Event, Intervention duration period. E.g., Evaluation Date.

Kind Regards,

Sergiy 

Single Point vs Interval Collection

I scraped this info from the SDTM 3.1.2 IG.  Hopefully it is still valid for 3.1.3.   I would agree that the same can be applied to -DY variables. 

--DTC before --ENDTC

I think this should NOT be a rule.

There are enough examples were --DTC (datetime of collection) can be after --ENDTC. For example, in MH, --DTC will very often be after --ENDTC. Simple example, I have taken Ibuprofen against headache (caused by the unnecessary complexity of SDTM) from January 12th to February 28th, but the investigator collected this information during my visit at March 15th.
Also in the Findings domains (though --ENDTC is pretty seldom there I presume), such an occasion can occur.

About --DY. This is a derived variable that should not be present in the SDTM data model!!! This is a violation of the third normal form for data(bases) - i.e. it is a transitive dependency. It should be calculated "on the fly" by tools that are used e.g. to inspect the data.

Does OpenCDISC check whether a given --DY is correctly calculated from --DTC and --RFSTDTC ?

DTC before ENDTC only for FINDINGS LB and PC

The only time DTC should be before ENDTC is for FINDINGS domains which implement interval collection such as LB and PC speficied in the table above.

Currenltly the only times where you would have all 3 variables populated (DTC, STDTC, ENDTC) is for MH and CE. Serigy, myself and Joseph agree this rule does not apply for MH and CE since DTC can be after ENDTC.