Model Documentation
Class: Analysis
The Analysis class is used in the reporting event's analyses attribute to define the analyses for the reporting event. Each analysis is defined as a set of operations (defined in a referenced analysis method) performed on a specified analysis variable for any specified subset of data records (defined in a referenced data subset) for a given subject population (defined in a referenced analysis set) which may be subdivided by one or more factors (defined in any referenced grouping factor). The results of each analysis are represented as a property of the analysis and each result is associated with references to all analysis components needed to give the result context (i.e., the operation used to calculate the result and a specific value for each of the grouping factors used for the analysis, such as treatment, sex, age group, etc.).
An instance of the Analysis class is created for each analysis and, in each instance:
- The id attribute contains the identifier value assigned for the analysis.
- The version attribute contains the version number of the analysis.
- The name attribute indicates the assigned name of the analysis.
- The description attribute may be used to record a detailed description of the analysis.
- The reason attribute is used to record the rationale for the analysis, and is equivalent to the arm:AnalysisReason attribute described in the ARM for Define-XML specification. The reason attribute contains an instance of 1 of the two specializations of the ExtensibleTerminologyTerm class for the AnalysisReasonEnum enumeration:
- AnalysisReason, in which the controlledTerm attribute contains a value from the AnalysisReasonEnum enumeration.
- SponsorAnalysisReason, in which the sponsorTermId attribute contains the identifier of a sponsor term created in the terminology extension for the AnalysisReasonEnum enumeration. In this case, the submissionValue attribute of the referenced sponsor term contains the value that would be shown as the rationale for the analysis.
- The purpose attribute is used to record the purpose of the analysis, and is equivalent to the arm:AnalysisPurpose attribute described in the ARM for Define-XML specification. The purpose attribute contains an instance of 1 of the 2 specializations of the ExtensibleTerminologyTerm class for the AnalysisPurposeEnum enumeration:
- AnalysisPurpose, in which the controlledTerm attribute contains a value from the AnalysisPurposeEnum enumeration.
- SponsorAnalysisPurpose, in which the sponsorTermId attribute contains the identifier of a sponsor term created in the terminology extension for the AnalysisPurposeEnum enumeration. In this case, the submissionValue attribute of the referenced sponsor term contains the value that would be shown as the purpose of the analysis.
For both reason and purpose, refer to the TerminologyExtension and ExtensibleTerminologyTerm sections for more information about the definition and use of terminology extensions, including examples.
- One or more references to associated documentation for the analysis may be included in the documentRefs attribute, as described in the DocumentReference section.
- One or more category identifier values may be included in the categoryIds attribute to categorize the analysis, as described in the AnalysisOutputCategorization section.
- The dataset and variable attributes are used to indicate the analysis variable that is the focus of the analysis. In general:
- The USUBJID variable in the ADSL dataset should be used as the analysis variable for analyses that represent summaries of unique subjects numbers.
- The same dataset variable should not be used as both the analysis variable and a grouping variable.
- The analysisSetId attribute contains the identifier value of an instance of the AnalysisSet class to indicate the subject population for the analysis. For analyses performed for an unrestricted subject population, it is recommended that an "All Subjects" population is defined and referenced. Refer to the AnalysisSet section for more information about defining analysis sets.
- The dataSubsetId attribute may contain the identifier value of an instance of the DataSubset class to indicate any restriction of the data records included in the analysis. Refer to the DataSubset section for more information about defining data subsets.
- The orderedGroupings attribute may be used to reference any grouping factors used for the analysis, as described below.
- The methodId attribute contains the identifier value of an instance of the AnalysisMethod class to indicate the analysis method used to perform the analysis. Each method may contain 1 or more operations, each of which produces a separate result value. Refer to the AnalysisMethod section for more information about analysis methods.
- The referencedAnalysisOperations attribute may be used to identify analyses that produce result values that are used in the calculation results for this analysis. For more information about referenced operation results, including examples, refer to the Operation section.
- The programming code used to generate the results of the analysis may be recorded or referenced in the programmingCode attribute, as described in the AnalysisOutputProgrammingCode section.
- The results attribute contains the results of the analysis, as described in the OperationResult section.
The subject or data groupings used for the analysis are each referenced in the orderedGroupings attribute. Although grouping factors are generally defined with reference to a single grouping variable (as described in the GroupingFactor section), multiple separate grouping factors may be referenced in the specification of an analysis, so it is possible to define analyses of subjects and data records that are grouped by multiple factors. In the orderedGroupings attribute each grouping factor is referenced in an instance of the OrderedGroupingFactor class, in which:
- The order attribute contains an integer value that is used to order the referenced grouping factor with respect to other referenced grouping factors (i.e., "1" for the first grouping factor, "2" for the second, and so on).
- The groupingId attribute contains the identifier value of the referenced grouping factor.
- The resultsByGroup attribute contains either "true" or "false" to indicate whether each operation in the analysis method referenced in the methodId attribute of the analysis is (resultsByGroup = "true") or is not (resultsByGroup = "false") expected to produce a separate result value for each of the groups in the referenced grouping factor. In this guide, referenced grouping factors are referred to as "results-by-group" grouping factors when resultsByGroup is "true"; i.e., an operation result value is expected for each group in the grouping factor. In analyses that generate summary statistics, all referenced grouping factors are usually results-by-group grouping factors. However, in analyses that generate comparative statistics across groups, some or all of the referenced grouping factors will not be results-by-group grouping factors. Refer to the OperationResult section for more information about the generation of operation result values for groups in grouping factors.
Example
- documentRefs: refer to the DocumentReference section examples of informational document references specified for an output.
- categoryIds: refer to the AnalysisOutputCategorization section for examples of the categorization of outputs.
- referencedAnalysisOperations: refer to the Operation section for an example of the definition and use of referenced operation relationships.
- programmingCode: refer to the AnalysisOutputProgrammingCode section for examples of the representation of programming code for analyses and outputs.
- results: refer to the OperationResult section for examples of the representation of analysis results.
In the following line captions, a label is shown in parentheses next to each referenced identifier value. These labels were obtained from the definition of each of the referenced components even though these definitions are not included in the example.
Lines 2-30: | Show the definition of the "Comparison of Height by Treatment" analysis, in which:
|
Lines 31-62: | Show the definition of the "Comparison of Height by Treatment" analysis, in which:
|
analyses: - id: An03_06_Height_Comp_ByTrt version: 1 name: Comparison of Height by Treatment description: ... reason: controlledTerm: SPECIFIED IN SAP purpose: controlledTerm: PRIMARY OUTCOME MEASURE documentRefs: ... categoryIds: ... dataset: ADSL variable: HEIGHTBL analysisSetId: AnalysisSet_02_SAF dataSubsetId: ... orderedGroupings: - order: 1 resultsByGroup: false groupingId: AnlsGrouping_01_Trt methodId: Mth04_ContVar_Comp_Anova referencedAnalysisOperations: ... programmingCode: ... results: ... - id: An08_01_Obs_Summ_ByTrt version: 1 name: Summary of Observed Value by Treatment, Parameter and Visit description: ... reason: controlledTerm: SPECIFIED IN SAP purpose: controlledTerm: PRIMARY OUTCOME MEASURE documentRefs: ... categoryIds: ... dataset: ADVS variable: AVAL analysisSetId: AnalysisSet_02_SAF dataSubsetId: Dss09_VS_AnRec orderedGroupings: - order: 1 resultsByGroup: true groupingId: AnlsGrouping_01_Trt - order: 2 resultsByGroup: true groupingId: AnlsGrouping_08_Param - order: 3 resultsByGroup: true groupingId: AnlsGrouping_09_Visit methodId: Mth02_ContVar_Summ_ByGrp programmingCode: ... results: ... ...