You are viewing an old version of this page. View the current version.

Compare with Current View Page History

« Previous Version 8 Current »

ADaM allows for user-specified variable names when creating additional variables not specified in this document or the ADaMIG. The ADaMIG has variable name fragments and other conventions, which are often needed with ADSL and BDS datasets. OCCDS datasets may have fewer derived variables, but there are some conventions and fragments described here to help build easily understood variable names.

Many variables from SDTM data are copied to ADaM with no change to the variable name, label, or content. ADaM follows a principle of harmonization (same name, same meaning, same values). This means that when an SDTM variable name is in an ADaM dataset, the values within that variable are exactly the same as the SDTM data.

Many times, a modified version of an SDTM variable is needed, such as when missing data is imputed. As shown in Section 2.9.7.8, Descriptive Variables, when creating an analysis version of an SDTM descriptive variable, the prefix “A” (for "analysis") replaces the 2-letter SDTM domain code. This same convention can be used to create analysis versions of variables not specified in Section 3.2.8. For example, if changes needed to be made to AEBODSYS for analysis, the variable name ABODSYS (with a prefix of “A” instead of “AE”) would convey that it is the analysis version of the AEBODSYS variable, and that the value of ABODSYS may not be the same as AEBODSYS on some records.

When variable content remains unmodified but the name of the SDTM variable cannot be used in the analysis dataset, the prefix “U” (for "unmodified") can be used instead of the 2-letter SDTM domain code. For example, if combining records from Adverse Events (AE) and Medical History (MH) into an analysis dataset, it would likely not be analysis-ready to have body-system information spread across 2 variables, AEBODSYS on the rows from the AE dataset and MHBODSYS on the rows from the MH dataset. Instead, a single variable that contains the value from AEBODSYS or MHBODSYS is needed. Using the “A” prefix and calling the variable ABODSYS implies that it is an analysis version of a variable, and that the value of ABODSYS may not be the same as AEBODSYS and MHBODSYS on some records. If no changes are made to AEBODSYS and MHBODSYS, using instead a “U” prefix allows for a single column called UBODSYS but makes it clear that no modifications to the content of the SDTM variables were made.

  • No labels