Extensible / not extensible CodeLists & non-standard values

b Bedeoan

on February 15, 2013

Hello all,

I'm a little puzzled what information should define.xml contain in the CodeList sections and non-standard values.

#1. The Codelist entry should contain all available values for that CodeList or jjust the ones used into the current dataset?

E.g. CL.C66731.SEX contains the F, M, U, UN values (see "SEND Terminology.xls"). If my dataset contains just M and F values for SEX, the Codelist in define.xml should contain:

a) entries just for M and F?

b) entries for all 4 values from SEND Terminology?

Going for option a) is easier and also has as result a smaller define.xml & define.pdf. However I've seen here: http://www.opencdisc.org/forum/codelist-extensible-vs-non-extensible recommendations for option b)...

#2. For an extensible CodeList the user can specify its own entries.

E.g. CL.C77808.SPECIES is extensible. So, if the user uses a species that is not in SEND terminology like "Primate" (instead of MONKEY which is in standard SEND CT), he can add an entry like below:

            <CodeList OID="CL.C77808.SPECIES" Name="Species" DataType="text">
                <CodeListItem CodedValue="Primate">
                    <Decode>
                        <TranslatedText>Primate</TranslatedText>
                    </Decode>
                </CodeListItem>
            </CodeList>

Is this correct?

#3. For a non-extensible CodeList, the user is not allowed to specify other entries than the standard ones. If he still wants to use both standard and non-standard entries, what can he add for that CodeList in define.xml?

a) add just standard SEND entries; ignore the non-standard ones?

b) add both standard and non-standard entries. But this would contradict the definition of "non-extensible"?

c) don't add an entry for that CodeList at all.

E.g. CL.C66781.AGEU is non-extensible. If the user uses as units both "DAYS" (in CT) and "SEMESTER" (not in CT), how the Codelist entry for AGE should look like?

Amelia

Forums: Define.xml

s Sergiy

on February 15, 2013

Hi Amelia, #1. The

Hi Amelia,

#1. The define.xml codelist should include only values used during data collection process.

E.g., only options listed on CRF. Your actual data can be a subset of those values. Actually in define.xml and annotated CRFs you need to fully describe your data collection process, but not only collected data. The later is a common data quality issue in our industry, because it's more simple to use collected data as input to define file (e.g., extract all collected terms by SAS proc freq) , rather than to extract metadata from EDC, data management system, etc.

The answer to your question is "all values presnted on your study CRF"

#3. You cannot use additional Terms in non-exta. It's severe violation. You need to map your original collected data into available standard terms