Validation Rule Suggestions

Hello all,

SEND specs allows you to specify some numeric results (e.g. for weighings) in both "collected" and "standardized" form, where:

a) the "collected" form is "Result of the measurement or finding as originally received or collected." (--ORRES and --ORRESU variables);

b) the "standardized" form is the "the result value for all findings or derived from ORRES in a standard format or standard units" (--STRESC, --STRESN, --STRESU variables).

Hello,

I am using Validator 1.3 and I ran it on a study that did not have IDVAR and IDVARVAL in SUPPDM. No error was produced in the report. I think this should be a check in OpenCDISC validator.

Regards,

David.

Hi all,

Do we need a rule for limitating the number of decimals used to represent the --STRESN variable for domains that contain measurements - e.g. OM or BW?

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.

I haven't found any validation rule for the relation between SE and EX records.

I feel that there is a one-to-many relation between SE and EX (1 SE to many EX). Shouldn't this be illustrated by a consistency rule (like SD0070 is)?

Some rule suggestions in relation to:

- "SD0023" - "Missing value for --STAT, when --REASND is provided";

- SD0036 - "Missing value for --STRESC, when --ORRES is provided";

- SD0047 - "Missing value for --ORRES, when --STAT or --DRVFL is not populated";

- SD0048 - "Value for --ORRES is populated, when --STAT is 'NOT DONE'";

Please comment if the suggestions are not valid for some reason:

1. When --STAT="NOT DONE", --STRESC should also be empty (like --ORRES).

I don't think that checks related to lengths of variables should be included.  FDA is pushing for reducing the length of each variable in a submission to the minumum length needed to accomodate the longest value.  Therefore, for example, if the longest value of ARMCD in a submission is 4, the length should not be forced to 20.

The rule CT1023 - "Value for --SEV not found in (SEV) CT codelist" is enabled into config-send-3.0.xml  for most of the domains, including CL (for CLSEV).

However, into the SENDIGv3.pdf standard the CLSEV variable is not controlled by any Terminology (SENDIGv3.pdf - page 70). Other --SEV variables (MASEV, MISEV) variables are controlled by the CL.C90000.SEV Terminology list.

Should rule CT1023 be disabled for CL, to reflect the SENDIGv3.pdf standard?

I notice that SD1029 is a new rule in the recent version of OpenCDISC and is marked as an error. Can you provide where this requirement came from. I understand that it is good practice not to use extended ascii characters but in order to convince our CROs of this it would help to be able to refer to a regulatory requirement.

.... SD1029 Non-ASCII or non-printable characters Character variable value must not include non-ASCII or non-printable characters (outside of 32-127 ASCII code range)