Validator 1.4.1 - some incorrect validations in PP, PC, LB, MI

b Bedeoan

on September 16, 2013

Hello all,

I just used the 1.4.1 version to validate on sone studies and validate them for SEND.

I notices some inconsistencies with the SEND 3.0 standard:

#1. SE0063 - SEND/dataset variable label mismatch :

- Validator expects "Planned Study Day of Collection" as label for PC VISITDY, but in SEND specs this is "Visit Day".

- Validator expects "Date/Time of Reference Time Point" as label for PC PCRFTDTC, but in SEND specs it is "Date/Time of Reference Point". Same for PP PPRFTDTC.

- Validator expects "Reference Range for Char Rslt-Std Units" for LB LBSTNRC, but in SEND specs it is "Reference Range for
Char Rslt-Std Unit" (no "s" at the end).

Or should we disregard the SEND standard and update the labels for these variables on our end?

#2. SD1079 Variable is in wrong order within domain

- Reported for PPSTINT, PP. But PPSTINT is wrongly placed as order in config-send-3.0.xml.

- Same obs for MIMETHOD in MI.

Forums: Validation Rule Suggestions

s Sergiy

on September 16, 2013

Hi Ameilia, Thank you for

Hi Ameilia,

Thank you for reporting your findings! Unfortunately there are still few incorrect metadata specs in validator.

Please confirm your validation results. I've double checked your examples in validator specifications. Some of them look confusing.

In OpenCDISC SEND specs the PC.VISITDY label is "Visit Day", the PCRFTDTC label is "Date/Time of Reference Point" as specified in SEND IG. I'm not sure why you got such reported messages.
Your finding about extra "s" character in the LBSRNRC label is correct.

We can update validation specifications on website. However I am not sure if it's the best option for quick fix of such minor findings. We need to be careful with version control.

Order of PPSTINT is incorrect. It should be in front of PPENINT and also PPDETECT. Unfortunately --STINT and --ENINT variables will be introduced only in SDTM v1.4 later this year, and their location inside Timing variables was not defined. SEND introduced --STINT, --ENINT variables several years before they are adopted by SDTM Model.
There is a conflict between SEND IG and SDTM Model about an order of MIMETHOD variable in a dataset. According to SEND IG #4.1.2 on page 21 "Variables for the three general observation classes should be ordered with Identifiers first, followed by the Topic Qualifier, and Timing variables. Within each role, variables are ordered as shown in Tables 2.2.1-2.2.5 of the SDTM." OpenCDISC Validator uses this rule. In SEND IG there are few cases (e.g., MIMETHOD) when an order of variables in standard domains' tables is not the same as specified by SDTM Model. It's confusing for users.

Kind Regards,

Sergiy

b Bedeoan

on September 24, 2013

Hi Sergyi, Please ignore my

Hi Sergyi,

Please ignore my comment related to PC.VISITDY and PCRFTDTC. It was an inconsistency on my end :) not in define.xml (where I was looking for the labels), but in headers of the xpts.

Thanks a lot for everything!