September 12, 2024 Webinar
graphicalMCP: An R package for graphical multiple comparison procedures
Dong Xi, PhD
Abstract:
Multiple comparison procedures (MCPs) are widely used in confirmatory clinical trials to control the probability of making false positive claims. Graphical approaches provide a flexible framework to accommodate commonly used MCPs and to create more powerful procedures. To allow both flexibility and efficiency, we've created graphicalMCP, an R-native package that is lightweight, performant, and educational. Using an extensible approach to allow different test types in different parts of a graph, the package covers all existing R packages for graphical approaches. Using vectorization and shortcuts, it greatly reduces simulation times compared to other similar packages, without using more efficient languages such as C++, which allows better compatibility with different computing environments. It also has the capability to draw graphical MCPs using vertices and directed edges, giving users great visualization along with full quantitative details.
Short Bio:
Dr. Dong Xi is a Statistical Advisor in the Biostatistics Innovation Group at Gilead Sciences. He has been supporting development and implementation of innovative statistical methodologies in multiplicity and causal inference for drug development in various therapeutic areas.