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

Compare with Current View Page History

« Previous Version 11 Next »

Model Documentation

Class: Operation

The Operation class is used in the operations attribute of the AnalysisMethod class to define the statistical operations to be performed for any analysis that references the analysis method.

Examples of operations include:

  • "Count of subjects by group" and "Percent of subjects by group" in a "Summary by group of a categorical variable" method
  • "Mean", "Standard Deviation", "Minimum Value", "Maximum Value", and confidence intervals in a "Summary by group of a continuous variable" method
  • "P-value" in a "Pearson's chi-square test group comparison for a categorical variable" method
Each operation generates a single result for each combination of groups in any "results-by-group" grouping factors included in the specification of an analysis that references the analysis method containing the operation (see the Analysis and OperationResult sections for more information about "results-by-group" grouping factors). Each result is generally expected to represent a single value, but sponsors may choose to define operations that generate composite results as a combination of more than 1 individual value. However, the disadvantage of composite results is that there will not be individual metadata elements to describe each component of a composite result. A better approach is to define operations that generate individual result values and then, if necessary, combine the result values from these operations to produce a composite value for display.

ER Diagram: Operation

Each operation is defined as an instance of the Operation class and, in each instance:

  • The id attribute contains the identifier value assigned for the operation.
  • The name attribute indicates the assigned name of the operation.
  • The label attribute contains a short label for the operation, such as might be shown in an output display (e.g., "n", "%", "Mean", "SD").
  • The description attribute may be used record a detailed description of the operation.
  • The referencedOperationRelationships attribute may be used to specify relationships to any other operations whose results are referenced in the calculation of the results for this operation, as described below.
  • The resultPattern attribute contains a textual representation for the default pattern or format used for the display of the results of this operation in output displays. The pattern should indicate the precision of the value to be displayed (e.g., "X.XXX") and may include parentheses or other additional characters (e.g., "(XXX.X)" or "±XX.XX"). In general, the pattern will usually be in the format that is displayed in output table shells, but sponsors may choose to use other pattern specifications.

Some "dependent" operations may be defined to reuse the results of other "referenced" operations in the calculation of their result values. For example, an operation that calculates a percentage may reuse the results of other "count of subjects" operations to act as the numerator and denominator in the calculation of the percentage. In such operations, the referencedOperationRelationships attribute of the dependent operation is used to specify 1 or more relationships to the other referenced operation(s) whose results are used in the dependent operation's calculation. Each relationship is defined as an instance of the ReferencedOperationRelationship class, in which:

  • The id attribute contains the identifier value assigned for the relationship.
  • The referencedOperationRole attribute contains an indication of the role that the results of the referenced operation play in the calculation of the dependent operation. The role is represented as an instance of 1 of the 2 specializations of the ExtensibleTerminologyTerm class for the OperationRoleEnum enumeration:
    • OperationRole, in which the controlledTerm attribute contains a role from the OperationRoleEnum enumeration
    • SponsorOperationRole, in which the sponsorTermId attribute contains the identifier of a sponsor term created in the terminology extension for the OperationRoleEnum enumeration. In this case, the submissionValue attribute of the referenced sponsor term contains the name of the role. 

See the TerminologyExtension and ExtensibleTerminologyTerm sections for more information about the definition and use of terminology extensions, including examples.

  • The operationId attribute contains the identifier value for the operation that is being referenced.
  • The analysisId attribute may contain the identifier of a "referenced analysis" (i.e., an analysis that uses the method that contains the referenced operation). The analysisId attribute is only used if the results of the referenced operation are always generated by the same analysis. If the referenced analysis may differ for each analysis that uses the method containing the dependent operation, then the referenced analysis may be specified in the referencedAnalysisOperations attribute of the analysis that uses the dependent operation (as described below).
  • The description attribute may be used to record details of the relationship, which may include:
    • In description of the referenced operation and how its results are used.
    • In indication of which analyses are expected to produce the required result values, which may be helpful in situations when the analysisId is not provided in the referenced operation relationship.

The identifier of the analysis that produces the results of the referenced operation (the "referenced analysis identifier") must be specified. If it is not included in the definition of the referenced operation relationship (i.e., the analysisId attribute of an instance of the ReferencedOperationRelationship class is not used), then the referenced analysis identifier must be specified as part of the definition of any analysis that uses the method containing the dependent operation. The referenced analysis identifier should not be specified in both the definition of the referenced operation relationship and the definition of an analysis that uses the dependent operation. When the referenced analysis identifier is specified as part of the definition of an analysis that uses the dependent operation, it is specified in the referencedAnalysisOperations attribute of the analysis in an instance of the ReferencedAnalysisOperation class, in which:

  • The referencedOperationRelationshipId attribute contains the identifier value of the referenced operation relationship.
  • The analysisId attribute contains the referenced analysis identifier.

Example

This example shows the specification of an analysis method that contains 2 operations, 1 of which is a dependent operation. The definitions of 2 analyses are also included, 1 of which is an analysis that uses the example method and the other is the analysis that produces some of the results used by the dependent operation.

Lines 2-32:

Show the "Summary by group of a categorical variable" analysis method, which contains 2 operations:

  • Lines 8-12: Show the definition of the "Count of subjects" operation.
  • Lines 13-32: Show the definition of the "Percent of subjects" operation, which is a dependent operation; it references the results of 2 other operations to act as the numerator and denominator in the calculation of the percentage:
    • Lines 19-25: Show the definition of the referenced operation relationship that specifies that the results of the operation identified as "Mth01_CatVar_Summ_ByGrp_1_n" (line 22) act as the numerator (lines 20-21) in the calculation of the percentage. The analysisId attribute is not included, which indicates that the referenced analysis identifier differs depending on the analysis that uses this operation (so this referenced analysis identifier must be specified in the definition of any analysis that uses this operation). The description (lines 23-25) indicates the type of analysis that can provide the numerator for the calculation.
    • Lines 26-32: Show the definition of the referenced operation relationship that specifies that the results of the operation identified as "Mth01_CatVar_Count_ByGrp_1_n" (line 29) act as the denominator (lines 27-28) in the calculation of the percentage. Again, the analysisId attribute is not included, which indicates that the referenced analysis identifier differs depending on the analysis that performs this operation (so this referenced analysis identifier must also be specified in the definition of any analysis that uses this operation). The description (lines 30-32) indicates the type of analysis that can provide the denominator for the calculation.
Lines 35-45:

Show the definition of the "Summary of Subjects by Treatment" analysis, which is identified as "An01_05_SAF_Summ_ByTrt". This analysis definition is included in the example because it uses the method identified as "Mth01_CatVar_Count_ByGrp", which contains the referenced operation that provides the denominator for the calculation of the percentage, as described above and referenced as "Mth01_CatVar_Count_ByGrp_1_n" on line 27. The definition of the method identified as "Mth01_CatVar_Count_ByGrp" is shown in Example 1 in the AnalysisMethod section.

Lines 46-65:

Show the definition of "Summary of Subjects by Treatment and Age Group" analysis, which is identified as "An03_02_AgeGrp_Summ_ByTrt" (line 46) and includes 2 referenced analysis operations:

  • Lines 61-62: Show that the analysis identified as "An03_02_AgeGrp_Summ_ByTrt" (line 62)—i.e., this same analysis—provides the results used as the numerator in the "Percent of subjects" operation (as specified in the referenced operation relationship identified as "Mth01_CatVar_Summ_ByGrp_2_pct_NUM" (line 61) whose definition is shown on lines 19-25). 
  • Lines 63-64: Show that the analysis identified as "An01_05_SAF_Summ_ByTrt" (line 64)—i.e., the analysis defined on lines 35-45—provides the results used as the denominator in the "Percent of subjects" operation (as specified in the referenced operation relationship identified as "Mth01_CatVar_Summ_ByGrp_2_pct_DEN" (line 63) whose definition is shown on lines 26-32). 

YAML Example
methods:
- id: Mth01_CatVar_Summ_ByGrp
  name: Summary by group of a categorical variable  
  description: Descriptive summary statistics across groups for a categorical variable,
    based on subject occurrence
  label: Grouped summary of categorical variable
  operations:
  - id: Mth01_CatVar_Summ_ByGrp_1_n
    name: Count of subjects
    label: n
    order: 1
    resultPattern: XXX
  - id: Mth01_CatVar_Summ_ByGrp_2_pct
    name: Percent of subjects
    label: '%'
    order: 2
    resultPattern: ( XX.X)
    referencedOperationRelationships:
    - id: Mth01_CatVar_Summ_ByGrp_2_pct_NUM
      referencedOperationRole:
        controlledTerm: NUMERATOR
      operationId: Mth01_CatVar_Summ_ByGrp_1_n
      description: The count operation whose result provides the numerator for calculation
        of the percentage. The referenced analysis should be the analysis that contains
        this percent operation.
    - id: Mth01_CatVar_Summ_ByGrp_2_pct_DEN
      referencedOperationRole:
        controlledTerm: DENOMINATOR
      operationId: Mth01_CatVar_Count_ByGrp_1_n
      description: The count operation whose result provides the denominator for calculation
        of the percentage. The referenced analysis should have the same analysis set
        and grouping(s) as the analysis containing this percent operation.
  ...
analyses:
- id: An01_05_SAF_Summ_ByTrt
  version: 1
  name: Summary of Subjects by Treatment
  methodId: Mth01_CatVar_Count_ByGrp
  dataset: ADSL
  variable: USUBJID
  analysisSetId: AnalysisSet_02_SAF
  orderedGroupings:
  - order: 1
    resultsByGroup: true
    groupingId: AnlsGrouping_01_Trt
- id: An03_02_AgeGrp_Summ_ByTrt
  version: 1
  name: Summary of Subjects by Treatment and Age Group
  methodId: Mth01_CatVar_Summ_ByGrp
  dataset: ADSL
  variable: USUBJID
  analysisSetId: AnalysisSet_02_SAF
  orderedGroupings:
  - order: 1
    resultsByGroup: true
    groupingId: AnlsGrouping_01_Trt
  - order: 2
    resultsByGroup: true
    groupingId: AnlsGrouping_03_AgeGp
  referencedAnalysisOperations:
  - referencedOperationRelationshipId: Mth01_CatVar_Summ_ByGrp_2_pct_NUM
    analysisId: An03_02_AgeGrp_Summ_ByTrt
  - referencedOperationRelationshipId: Mth01_CatVar_Summ_ByGrp_2_pct_DEN
    analysisId: An01_05_SAF_Summ_ByTrt
  ...
...

  • No labels