Analysis dataset structures in this guide provide a standard way to build analysis datasets. These structures are captured in the "CLASS OF DATASET" field in the ADaM metadata. Currently ADaM has 3 structures: the Subject-level Analysis Dataset (ADSL), the Basic Data Structure (BDS), and the Structure for Occurrence Data (OCCDS). Analysis datasets that follow the ADaM fundamental principles and other ADaM conventions but do not follow one of these 3 defined structures are considered to be ADaM datasets with a class of ADAM OTHER. Currently the ADaM standards do not cover non-subject data. A REFERENCE dataset structure has been developed for the TIG and will be considered for inclusion in future versions of ADaM.

For the TIG Product Description use case, ADaM fundamental principles will be followed with identifiers in addition to the currently allowed USUBJID and SPEVID. Current ADaM principles and standards are implemented for the Product Impact on Individual Health use case.  For the Product Impact on Population Health use case, the REFERENCE dataset structure addresses reference data which captures historic trends and results that are used as input for study analyses.

The following table describes ADaM dataset structures applicable to TIG use cases.

Use CaseDataset Structure(s)Implementation
Product description

ADAM OTHER

Follows ADaM fundamental principles but is not subject based

Product impact on individual health

ADSL, BDS, and OCCDS

Follows ADaM fundamental principles and current ADaM structures

Product impact on population healthREFERENCEFollows ADaM fundamental principles and the new dataset structure developed for the TIG
Analysis dataset specifications list standard ADaM variables and define the required characteristics of standard variables (columns) that are frequently needed in ADaM datasets. The ADaM standard requires that these variable names be used when a variable that contains the content defined in the pre-defined ADaM variables sections is included in any ADaM dataset, regardless of dataset class. It also requires these ADaM standard variables be used for the purposes indicated, even if the content of an ADaM variable is a copy of the content of an SDTM dataset variable.

The columns in an ADaM dataset specification table are:

RowDataset Specification ColumnPurpose of Column Content
1Variable NameSpecifies the name of the variable in the resulting dataset
2Variable LabelSpecifies a descriptive label for the variable
3Type

Specifies the data type of the variable. Values for in this column are:

  • Num for numeric data
  • Char for character or alphanumeric data
4Controlled Terms, Codelist, or Format

Specifies applicable controlled terminology or formats with which to populate the variable. Values in this column are:

  • Names of one or more CDISC Controlled Terminology codelists, with each codelist name in parentheses
  • Short references to external terminology (e.g., MedDRA)
  • The name of an external ISO 8601 format
5Core

Specifies expectations for inclusion of the variable in the resulting dataset. Values in this column are:

  • Req for variables which are Required and must be included in the resulting dataset. However, unlike SDTM a required variable may be null for any record. 
  • Cond for variables which must be included in the dataset in certain circumstances
  • Perm for variables for which it is Permissible to include or exclude the variable from the resulting dataset
6CDISC Notes

Provides additional context for the intended use of the variable and may include:

  • A description of the purpose of the variable and/or what the variable means
  • Guidelines for variable use including rules for when or how the variable should be populated, or how the contents should be formatted
  • Example values which could appear in the variable. Such values are intended to support understanding and are not intended to influence decisions regarding what values must be captured in the variable.

  • No labels