pharmaverseadam

Test data (ADaM) for the pharmaverse family of packages

Purpose

To provide a one-stop-shop for ADaM test data in the pharmaverse family of packages.

Package Contents

The ADaM contents of this package is populated by an script that executes the {admiral}, {admiralonco}, {admiralophtha}, {admiralvaccine}, {admiralpeds}, {admiralmetabolic} and {admiralneuro} templates and saves the resulting datasets here. This script can be executed by the package maintainers in two scenarios:

• Regularly, upon new releases of these packages;
• Ad-hoc, whenever templates in the above packages have been updated but releases are far away in the calendar. In this case, the ADaM datasets are created using a development version of the templates.

Please check the Changelog to see the versions of the packages used to generate the ADaM datasets in current or past version of {pharmaverseadam}. Please see the Dataset Creation and Documentation Process for details on the script.

Installation

The package is available from CRAN and can be installed with:

install.packages("pharmaverseadam")

To install the development version of the package from GitHub run:

pak::pkg_install("pharmaverse/pharmaverseadam", dependencies = TRUE)

Dataset Creation and Documentation Process {#doc}

The execution of the ADaM templates and creation of the package documentation in {pharmaverseadam} is semi-automated for consistency and ease of maintenance. Metadata for each dataset such as names, labels, descriptions, authors, and sources, is managed in a centralized XLSX file (inst/extdata/adams-specs.xlsx) and used to generate .R documentation files. These are then used to populate the reference pages in the package documentation.

The workflow consists of two main steps:

1. Metadata Preparation

Firstly, the user reviews inst/extdata/adams-specs.xlsx and ensures its contents is up-to-date. If a new template has been added to an existing package, or if a new extension package has been created whose ADaM templates they wish to execute in {pharmaverseadam}, then the user should update adams-specs.xlsx with specifications for the new dataset(s). Note that any missing metadata fields will be set by default to "No label/description/source available."

2. New Therapeutic Area Addition

The addition of a new Therapeutic Area to the package requires its inclusion in:

• _pkgdown.yml and data-raw/create_adams_data.R, where it is used for the titles and grouping on the Reference page.
• vignettes/articles/preview-adam.Rmd, which is used to generate Data Exploration previews. The grouping of datasets in this vignette is based on the Therapeutic Area.

3. Execute create_adams_data.R

Secondly, the user runs data-raw/create_adams_data.R. This script handles the installation of each package, the execution of the templates, the saving of the ADaM datasets and the creation of the documentation pages.

Script steps

1. Save specs as JSON
   The script saves the specs stored in inst/extdata/adams-specs.xlsx as a JSON file located here: inst/extdata/adams-specs.json. This is so that the specs themselves, as well as any diffs across commits, are easily viewable on Github and R Studio.

2. Installs Required Packages
   The script installs the following packages:

   • {admiral}
   • {admiralonco}
   • {admiralophtha}
   • {admiralvaccine}
   • {admiralpeds}
   • {admiralmetabolic}
   • {admiralneuro}
   • {pharmaversesdtm}

   By default, the latest development versions of each package will be used, but the user can also select a different version instead (e.g. a released version) with which to refresh {pharmaverseadam}. The user may also deselect some packages whose templates they do not wish to run. If a new extension package has been created, this should be added to the script in each relevant section (see the script itself for more details).

3. Executes Templates from Each Package

4. Processes and Saves Datasets into {pharmaverseadam}

5. Generates Dataset Documentation
   For every dataset, a matching .R file is created in the R/ folder containing:

   • Dataset name and title,
   • Variable-level documentation (names and labels),
   • Source information indicating which template and package generated it,
   • Example usage with data("<dataset>").

Acknowledgments

Along with the authors and contributors, thanks to the following people for their work on the package:

G Gayatri, Daphne Grassely, Sadchla Mascary, Kangjie Zhang and Zelos Zhu.