Working with ADaM Datasets

analysis datasets are designed to meet the goals and fundamental principles of the cdisc analysis data model (adam) and provide analysis ready datasets you can configure and format assets (for example, derived variable assets) in line with the cdisc adam principles ryze also provides published standards that can be used to create adam datasets these published standards have pre configured traceability to sdtm and provide a good starting point for building analysis datasets in practice, adam datasets can be much larger and include more variables than sdtm in this guide we’ll look at how ryze helps manage adam datasets in line with cdisc guidelines traceability ryze provides published adam datasets that can be imported into your standard or study these published datasets have pre configured links to sdtm for example, an adam variable might have an origin\ predecessor property that points to an sdtm variable identifier such as dm age ryze helps provide clear traceability from crf to sdtm and from sdtm to adam datasets analysis dataset creation when creating an analysis dataset, you’ll want to include analysis variables and classes such as adsl, bds, ods and so on published standards include these classes already and provide things like a prefixed identifier (e g bds studyid) and origin properties but you can manually build your own variables the following options are available when working with adam datasets use the published adam standard published adam datasets have pre configured links to sdtm as well as provide imputed and derived variables, so they provide a good starting point for building adam datasets you can also extend adam datasets by using cloning functionality to clone individual variables; these cloned variables retain links to sdtm variables for example, if you want to extend a grouped or pooled variable such as, pooled age group y (n), cloning functionality can be used creating other classes ryze provides the following adam dataset classes by default adsl bds occds adtte these datasets include the applicable identifier prefix, for example, occds studyid you can also create other classes and apply identifier prefixes by cloning the published adam content you must then change the identifier at the variable level, for example, bds studyid to advs studyid clone variables from sdtm and modify if you want to create a specific adam variable that already exists in your repository as an sdtm variable, you can import the sdtm variable as a clone and change the origin property this allows you to keep the link to the sdtm variable as well as quickly create an adam variable when cloning variables from sdtm, you must set the origin property to predecessor and add the object identifier (oid) of the source sdtm variable to the description field create variables manually alternatively, you can create adam variables and manually link them to sdtm variables in another standard this is not recommended if you are creating copies of sdtm variables and the cloning method described previously should be used analysis dataset generation analysis result metadata (arm) is used to tie together adam datasets, outcome tables, and documentation that describes the analysis the purpose of the define xml analysis results metadata extension is to support the interchange of cdisc adam key analysis results metadata for clinical research applications in a machine readable format ryze allows you to generate the define xml with arm and there are two assets that are used to manage arm results display and analysis result results display this asset describes which result is displayed, for example which outcome table is in view it can include a description of the outcome as well as a reference to the analysis result analysis result this asset describes the key result or exploratory outcome measure (analysis purpose) analysis result describes which variables are used for the analysis (analysis datasets) and includes a description of what is displayed (description) these assets can also include external documents such as the statistical analysis plan (document) and links to programming code used to derive the primary outcome measure (code) working with arm analysis results metadata (arm) provides traceability for an analysis result to the specific adam data used as input, it also provides information about the analysis method used and the reason the analysis was performed since arm describes a result that relates to adam datasets, before specifying arm, the adam dataset metadata must be added in the https //www cdisc org/standards/foundational/define xml/analysis results metadata arm v1 0 define xml v2 0 , cdisc outlines the key components as follows analysis display metadata definitions analysis result metadata definitions analysis parameter(s) analysis dataset(s) analysis variable(s) selection criteria documentation programming statements if you have an asset group with adam datasets, you can create a result display using the following workflow create the results display enter details for label, name, identifier and description click validate create document the result display has documentation provided to further describe the analysis description provides a short local text description, and a document reference is used to link to an external document (e g pdf) with page references click create document enter details for label, name and identifier to attach an external document, click on the upload icon under file name add page reference when a document is attached, you can create one or more page references the references can cite page numbers, a page range or named destinations within the document \[optional] enter some descriptive text click validate create analysis result click create analysis result enter details for label, name, identifier, analysis reason, analysis purpose add code the analysis result captures the programming code used to derive the result this can be expressed in two ways direct code entry into the “code” fields code>context and code>expression an external document reference, linking to code for example in a pdf file click validate associate datasets, in this step you reference the adam dataset providing the source data click associate datasets and select the source dataset under selection criteria , click add criteria add a label, name and identifier add a where clause, select a variable within the associated dataset, then specify a selection criteria using the comparator and value you can open the source dataset in another tab for reference you can add more than one selection criteria add a parameter when the dataset has been associated, it is possible to link to a valuelist item using the " parameter " property add analysis documentation for each result display, a separate documentation link can be created, giving a local description and external document reference export to define xml the docid\ kj0myd7uwa6 sfyeieluh tab is used to export to define xml exports must be generated at asset group level