Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Videos

Rank tests for clustered survival data when dependent subunits are randomized.

Jong-Hyeon Jeong1, Sin-Ho Jung

  • 1Department of Biostatistics, University of Pittsburgh, PA 15261, USA. jeong@nsabp.pitt.edu

Statistics in Medicine
|September 15, 2005
PubMed
Summary

This study adjusts standard rank tests for clustered survival data, accounting for intracluster correlations in clinical trials. The new variance formulas improve testing for differences in survival distributions between treatment groups.

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Alirocumab Plus Cemiplimab in Immunorefractory NSCLC: A Single Arm Phase 2 Study.

Research square·2026
Same author

Alirocumab plus cemiplimab in advanced immuno-refractory metastatic non-small cell lung cancer: an ongoing multi-center phase II study.

Future oncology (London, England)·2026
Same author

A note on rank-preserving structural failure time models to account for crossover.

Clinical trials (London, England)·2026
Same author

Novel niclosamide stearate prodrug therapeutic shows potential efficacy against naturally occurring canine osteosarcoma in a clinical feasibility study.

American journal of veterinary research·2025
Same author

Real-world prescribing in accordance to ACC/AHA guidelines for lipid-lowering therapy in high-risk primary and secondary prevention of ASCVD: Real-World Prescribing for Lipid-Lowering Therapy.

American journal of preventive cardiology·2025
Same author

Design and analysis of individually randomized group-treatment trials with time to event outcomes.

Lifetime data analysis·2025

Area of Science:

  • Biostatistics
  • Clinical Trials Methodology
  • Survival Analysis

Background:

  • Clustered survival data exhibit positive intracluster correlations.
  • Standard rank tests assume independent samples, requiring adjustments for clustered data.
  • Clinical trials often randomize subunits from the same cluster to different treatments.

Purpose of the Study:

  • To derive general variance formulas for rank tests in clustered survival data with inter-group randomization.
  • To evaluate the performance of these adjusted variance formulas via simulation.
  • To compare the proposed non-parametric tests with shared frailty models.

Main Methods:

  • Derivation of general variance formulas for logrank, Gehan-Wilcoxon, and Prentice-Wilcoxon tests.
  • Adjustment for intracluster correlations within and between treatment groups.

Related Experiment Videos

  • Extensive simulation studies to assess small sample performance.
  • Comparison with optimal semi-parametric testing procedures (shared frailty model).

    • Main Results:

    • A general form of simple variance formulas for rank tests was derived.
    • Simulation studies demonstrated the performance of the adjusted variance formulas.
    • The non-parametric rank tests with adjusted variances showed comparable results to shared frailty models under specific conditions.

    Conclusions:

    • The derived variance formulas provide a practical adjustment for rank tests in clustered survival data.
    • These adjusted tests are valuable for comparing marginal survival distributions in clinical trials with clustered randomization.
    • The non-parametric approach offers a robust alternative, especially when assumptions of shared frailty models are not met.