8    Representing Relationships and Data

8.4    Relating Non-Standard Variables Values to a Parent Domain

8.4.4    Alternative Representation of Non-Standard Variables

8.4.4.1   Background

As an alternative to submitting non-standard variables (NSVs) in separate SUPP-- datasets, they may be represented in the parent datasets. This method provides some additional benefits. These include the following:

• Permitting direct viewing of standard variables and NSVs within the same structure, eliminating the need for tools or the writing of programs to display the data together.
• Eliminating some current SUPP-- structural limitations by allowing:
  • Numeric NSVs to be represented in a numeric data type
  • Character NSVs to be defined with an appropriate length for each variable, rather than the typical default of $200 for QVAL
  • Allowing metadata for NSVs (including Controlled Terminology) to be applied at the variable level instead of the value-level.

8.4.4.2   Implementation Rules

In order to use the alternative described above, the following implementation rules must be followed:

• The list domains to which NSVs may be added does not change. NSVs may be added only to general-observation-class domains and Demographics.
• The same considerations must be used in assessing the number of NSVs represented, regardless of whether they are represented in SUPP-- datasets or in the parent domains. In other words, sponsors must not seek this implementation as a method for representing NSVs that would not have previously been represented in SUPP-- datasets.
• Complete metadata must be provided for each NSV in the same way it’s required for standard variables.
• Variable lengths for character NSVs should be set to the appropriate length for that variable, as with all standard character variables.
• Metadata roles for NSVs are defined appropriately, based upon the SDTM variables roles, and must be one of the following Roles:

1. 1. Non-Standard Identifier
   2. Non-Standard Qualifier
   3. Non-Standard Timing

• NSVs would be ordered after the standard variables, and ordered as shown in the bullet above. The order within each of the non-standard Roles should be consistent within the dataset and the define.xml.

Example

This is an example of hospitalization data with non-standard variables.

This example shows three hospitalization records, two for Subject 0001, and one for Subject 0002. In the collected data, there are standard variables from the SDTM: Identifiers from Table 2.2.4, Event Qualifiers from Table 2.2.2 of the SDTM, Timing variables from Table 2.2.5. In addition, there are the following non-standard variables:

Two methods are shown for the representation and submission of these data: the method of submitting NSVs in SUPPHO and the new method of submitting NSVs in the parent domain.

1. Existing Method Using SUPP--

ho.xpt: Healthcare Encounters

suppho.xpt

2. New Method with Non-Standard Variables Included in Parent Domain

ho.xpt: Healthcare Encounters

HO NSV Metadata

8.4.5    When Not to Use Non-Standard Variables

Examples of data that should not be submitted as Supplemental Qualifiers are the following:

• Subject-level objective data that fit in Subject Characteristics (SC). Examples include Marital Status, Skin Classification, Education Level.
• Findings interpretations that should be added as an additional test code and result. An example of this would be a record for ECG interpretation where EGTESTCD = “INTP”, and the same EGGRPID or EGREFID value would be assigned for all records associated with that ECG (see Section 4: 4.1.5.5, Clinical Significance For Findings Observation Class Data).
• Comments related to a record or records contained within a parent dataset. Although they may have been collected in the same record by the sponsor, comments should instead be captured in the CO special-purpose domain.
• Data not directly related to records in a parent domain. Such records should instead be captured in either a separate general observation class or special purpose domain.