Graphical Multiple Comparison Procedures

Multiple comparison procedures (MCPs) control the familywise error rate in clinical trials. Graphical MCPs include many commonly used procedures as special cases; see Bretz et al. (2011) , Lu (2016) , and Xi et al. (2017) . This package is a low-dependency implementation of graphical MCPs which allow mixed types of tests. It also includes power simulations and visualization of graphical MCPs.

graphicalMCP

Introduction

Graphical approaches for multiple comparison procedures (MCPs) are a general framework to control the family-wise error rate strongly at a pre-specified significance level $0<\alpha<1$. This approach includes many commonly used MCPs as special cases and is transparent in visualizing MCPs for better communications. graphicalMCP is designed to design and analyze graphical MCPs in a flexible, informative and efficient way.

Installation

Release

You can install the current release version from CRAN with:

install.packages("graphicalMCP")

Development

You can install the current development version from GitHub with:

# install.packages("pak")
pak::pak("openpharma/graphicalMCP")

Documentation

• For basic usage instructions, see vignette("graphicalMCP")
• To become familiar with graphical MCP terminologies, see vignette("glossary")
• To learn examples of how to use graphicalMCP,
  • see vignette("shortcut-testing") for sequentially rejective graphical multiple comparison procedures based on Bonferroni tests
  • see vignette("closed-testing") for graphical multiple comparison procedures based on the closure principle using Bonferroni, Hochberg, parametric and Simes tests
  • see vignette("graph-examples") for common multiple comparison procedures illustrated using graphicalMCP
  • see vignette("internal-validation") for internal validation via power simulations for methods used in graphicalMCP
  • see vignette("generate-closure") for rationales to generate the closure and the weighting strategy of a graph
  • see vignette("comparisons") for comparisons to other R packages
• To view vignettes in R after properly installing graphicalMCP, we can build vignettes by devtools::install(build_vignettes = TRUE), and then use browseVignettes("graphicalMCP") to view the full list of vignettes

Related work

• Graphical MCPs - gMCP
• Lighter version of gMCP which removes the rJava dependency - gMCPLite
• Graphical MCPs with Simes tests - lrstat

Built upon these packages, we hope to implement graphical MCPs in a more general framework, with fewer dependencies and simpler S3 classes, and without losing computational efficiency.

Acknowledgments

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

Keaven Anderson, Frank Bretz, Nan Chen, Yao Chen, Spencer Childress, Chelsea Dickens, Ekkehard Glimm, Michael Grayling, Willi Maurer, Colleen McLaughlin, Friedrich Pahlke, Matt Roumaya, Daniel Sabanés Bové, Alex Spiers, Gernot Wassmer, Jeremy Wildfire, Nan Xiao, Yanyao Yi, Ron Yu, and Ying Zhang

We owe a debt of gratitude to the authors of gMCP for their pioneering work, without which this package would not be nearly as extensive as it is.