Each year, a competition is run by the SCTO’s Statistics & Methodology Platform for statisticians in the CTU Network environment to develop code or programmes that will ultimately help improve clinical research and bring fresh solutions to persisting difficulties. The winners of the 2020 competition are Dr Marco Cattaneo and Arnaud Künzi for the following statistical programming packages:
selcorr by Dr Marco Cattaneo
The package’s purpose is to correct the p value upwards after stepwise selection of variables for a multivariable regression analysis. The uncorrected p value is too
optimistic because it does not account for the fact that the variables
have been selected for goodness-of-fit out of a larger set of variables. The package should therefore help yielding p values and confidence
intervals that are more adequate. The programme is written in R.
sts_graph2 by Arnaud Künzi The packages serves to conduct and visualise a landmark analysis in Stata. Landmark analyses are a popular observation method for comparing failure time that depend on group membership at the time of analysis. It is usually presented graphically using a Kaplan-Meier graph, supplemented by a table showing the risk population. Until now, landmark analysis in Stata has been very time-consuming, so a special, ready-to-use tool package is highly appreciated by the community.
0 Comments
Add a new comment