SD1247: ECDOSE is not greater than 0 when ECOCCUR does not equal 'N' and ECDOSTXT is null

ECDOSE is not greater than 0

My understanding of ECDOSE is that it is surely NOT the amount of active ingredient, but the amount given in units of ECDOSU.
Example: 1 tabled => ECDOSE=1, ECDOSU=TABLET. Other example: 5ml placebo => ECDOSE=5, ECDOSU=ml.
So, ECOCCUR=N really means that nothing at all was given, and is, as also indicated in the SDTM-IG, not compatible with ECDOSE/ECDOSTXT/ECDOSU.
When nothing was indeed given, set ECOCCUR to "N", and leave ECDOSE and ECDOSU empty.
And if your source database does have "0" for "not given", yes, then you should indeed set ECOCCUR to "N" and leave ECDOSE empty as well as ECDOSU.
However, I do still not understand what you mean by "manually update ECOCCUR". The above described logic should be part of your mapping process.
So, if you made an error there, I would suggest to correct it in the mapping process, and re-run the mappings.

Also note that there is another great forum "Validation of CDISC Data" for such questions on LinkedIn: https://www.linkedin.com/groups/12122426/
2 forums know more than one ... ;-)

With best regards,
Jozef Aerts, XML4Pharma

Hi Jozef,  Thanks for your…

Hi Jozef, 

Thanks for your kind reply. Actually, our study is a double-blinded study, we first set ECTRT = BLIND TREATMENT / PLACEBO and ECDOSE = 1200  collected from blinded raw data, but after Unblinding, here we have some records with ECTRT= PLACEBO and ECODSE = 0, since this is a PLACEBO treatment, so that the ECDOSE should be 0, but it may inconsistent with ECOCCUR = Y base on the CDISC validation rule. So, How do you think we should deal with these ECTRT = PLACEBO records?

ECDOSE is not greater than 0

Dear Phoebe,

What is given as placebo? If it is a tablet than ECTRT=PLACEBO, ECDOSE=1, ECDOSU=TABLET.
So, if anything was given, ECDOSE can never be "0".

With best regards,
Jozef

H Jozef,   Thanks for your…

H Jozef,

Thanks for your quick response.

We checked the SDTM IG3.2, it shows that the doses of placebo should be represented by EXTRT = ‘PLACEBO’ and EXDOSE = 0, so we also assign ECDOSE = 0 when ECTRT = PLACEBO.

By the way, we acctually derived EC into two parts, ECMOOD = SCHEDULED and ECMOOD =  PERFORMED, so after blinding, we manually update the ECTRT and ECDOSE when ECMOOD = PERFORMED per actual exposure datasets.

Per your reply, we are not sure if it is accptable for SDTM IG when assign ECDOSE >0  for ECTRT = PLACEBO and ECMOOD = PERFROMED. We did not find any examples in the SDTM IG 3.2 for now.

EC versus EX

--DOSE variables have a different meaning in EC and EX. In EX, it is related to the active ingredient, in EC, it isn't.
In EC, you don't know (yet), in the case of a blinded study, whether the subject received placebo or active ingredient, and ECDOSE is related to what (tablet, injection, inhalation, ...) the subject received. So, ECDOSE=0 may only be used when the suubject received nothing at all.

If you look at example 6 of the SDTMIG-3.2 on page 106-108, you see that in EX (p.108), EXDOSE=0 in case of placebo, which is indeed according to what you say and did.
For EC however (on page 107), you see that there is no mention of "placebo" at all, all you see for ECTRT is "BOTTLE1", and "BOTTLE2", and when comparing the data with those of EX, one can deduce that "BOTTLE1" corresponds "DRUGA", and "BOTTLE2" corresponds to "PLACEBO", but you don't see that in EC.

So, "PLACEBO" or the name of the active ingredient should essentially never occur for ECTRT in EC in the case of blinded studies, they can only occur for open label studies as in exampe 2 on page 107.

Yes, I know it is complicated and contra-intuitive. I have to look it up each time again too ...