Statistical Graphical User Interface Plug-In for Survival Analysis in R Statistical and Graphics Language and Environment

Abstract

Introduction: R is a statistical and graphics language and environment. Although it is extensively used in command line, graphical user interfaces exist to ease the accommodation with it for new users. Rcmdr is an R package providing a basic-statistics graphical user interface to R. Survival analysis interface is not provided by Rcmdr. The AIM of this paper was to create a plug-in for Rcmdr to provide survival analysis user interface for some basic R survival analysis functions. Materials and Methods: The Rcmdr plug-in code was written in Tinn-R. The plug-in package was tested and built with Rtools. The plug-in was installed and tested in R with Rcmdr package on a Windows XP workstation with the "aml" and "kidney" data sets from survival R package. Results: The Rcmdr survival analysis plug-in was successfully built and it provides the functionality it was designed to offer: interface for Kaplan Meier and log log survival graph, interface for the log-rank test, interface to create a Cox proportional hazard regression model, interface commands to test and assess graphically the proportional hazard assumption, and influence observations. Conclusion: Rcmdr and R though their flexible and well planed structure, offer an easy way to expand their functionality that was used here to make the statistical environment more user friendly in respect with survival analysis.