Assessment of Optimal Selected Prognostic Factors

Science Direct Working Paper No S1574-0358(04)70335-4

23 Pages Posted: 20 Mar 2018

See all articles by Berthold Lausen

Berthold Lausen

affiliation not provided to SSRN

Torsten Hothorn

University of Erlangen-Nürnberg

Frank Bretz

affiliation not provided to SSRN

Martin Schumacher

University of Freiburg - Medical Center

Date Written: October 2002

Abstract

The identification and assessment of prognostic factors is one of the major tasks in clinical research. The assessment of one single prognostic factor can be done by recently established methods for using optimal cutpoints. Here, we suggest a method to consider an optimal selected prognostic factor from a set of prognostic factors of interest. This can be viewed as a variable selection method and is the underlying decision problem at each node of various tree building algorithms.We propose to use maximally selected statistics where the selection is defined over the set of prognostic factors and over all outpoints in each prognostic factor. The maximum of multivariate normally distributed random variables is the approximate null distribution for maximally selected rank statistics. We demonstrate that it is feasible to compute the approximate null distribution. Data of the German Breast Cancer Study Group and of a small study on patients with diffuse large B-cell lymphoma illustrate our method. Moreover, we discuss the wide range of possible applications of our suggestion.

Keywords: prognostic factor, clinical research, statistical computing, variable selection, maximally selected tests, -value adjusted regression trees

Suggested Citation

Lausen, Berthold and Hothorn, Torsten and Bretz, Frank and Schumacher, Martin, Assessment of Optimal Selected Prognostic Factors (October 2002). Science Direct Working Paper No S1574-0358(04)70335-4, Available at SSRN: https://ssrn.com/abstract=3142693

Berthold Lausen (Contact Author)

affiliation not provided to SSRN

Torsten Hothorn

University of Erlangen-Nürnberg ( email )

Schloßplatz 4
Erlangen, DE Bavaria 91054
Germany

Frank Bretz

affiliation not provided to SSRN

No Address Available

Martin Schumacher

University of Freiburg - Medical Center ( email )

Hugstetter Straße 49
Freiburg, 79106
Germany

Do you have negative results from your research you’d like to share?

Paper statistics

Downloads
60
Abstract Views
421
Rank
643,430
PlumX Metrics