gms | German Medical Science

65th Annual Meeting of the German Association for Medical Informatics, Biometry and Epidemiology (GMDS), Meeting of the Central European Network (CEN: German Region, Austro-Swiss Region and Polish Region) of the International Biometric Society (IBS)

06.09. - 09.09.2020, Berlin (online conference)

Artikel

Artikel empfehlen

Testing contrasts of quantiles in general factorial designs

Meeting Abstract

Suche in Medline nach

  • Marc Ditzhaus - TU Dortmund University, Dortmund, Germany
  • Dennis Dobler - Vrije Universiteit Amsterdam, Amsterdam, Netherlands
  • Markus Pauly - TU Dortmund University, Dortmund, Germany

Deutsche Gesellschaft für Medizinische Informatik, Biometrie und Epidemiologie. 65th Annual Meeting of the German Association for Medical Informatics, Biometry and Epidemiology (GMDS), Meeting of the Central European Network (CEN: German Region, Austro-Swiss Region and Polish Region) of the International Biometric Society (IBS). Berlin, 06.-09.09.2020. Düsseldorf: German Medical Science GMS Publishing House; 2021. DocAbstr. 258

doi: 10.3205/20gmds020, urn:nbn:de:0183-20gmds0203

Veröffentlicht: 26. Februar 2021

© 2021 Ditzhaus et al.
Dieser Artikel ist ein Open-Access-Artikel und steht unter den Lizenzbedingungen der Creative Commons Attribution 4.0 License (Namensnennung). Lizenz-Angaben siehe http://creativecommons.org/licenses/by/4.0/.

Gliederung

Text

As quantities of interest, we consider the median survival time and more general quantiles in the survival set-up with possibly right-censored observations. Using corresponding estimands, we formulate null hypotheses and determine confidence regions for these survival endpoints. A Wald-type statistic is proposed in the context of general factorial designs, which allows studying main and interaction effects. The test's theoretical properties, such as being asymptotically exact under the null hypothesis and consistent under general alternatives, can be transferred to a permutation counterpart of the test. The latter performs significantly better than the asymptotic test in terms of type-1 error control in case of small sample sizes. Simulation results under various null hypotheses are complemented by power comparisons and an illustrative data analysis.

The authors declare that they have no competing interests.

The authors declare that an ethics committee vote is not required.