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

Gibbs sampling for long-term survival data with competing risks

E C Chao1

  • 1MathSoft, Inc., Seattle, Washington 98109, USA. echao@statsci.com

Biometrics
|April 17, 1998
PubMed
Summary

This study introduces a cure model extension for competing risks, offering more realistic cure probabilities than the 5-year survival rate. The model predicts a 27% cure probability for leukemia patients with acute graft-versus-host disease (GVHD) and 46% for the non-GVHD group.

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

Inhibition of discharge in inferior colliculus, AII cortex and Ep cortex after presentations of click stimuli.

Brain research·1994
Same author

Mineral-produced high-pressure striae and clay polish: key evidence for nonballistic transport of ejecta from ries crater.

Science (New York, N.Y.)·1976
Same author

Contacts with Earth Scientists in the People's Republic of China: Geological tasks are popular; collective participation and relevance to current problems are emphasized.

Science (New York, N.Y.)·1973
Same author

Petrology of unshocked crystalline rocks and shock effects in lunar rocks and minerals.

Science (New York, N.Y.)·1970
Same author

Shock effects in certain rock-forming minerals.

Science (New York, N.Y.)·1967
Same author

Nickel-iron spherules from aouelioul glass.

Science (New York, N.Y.)·1966

Area of Science:

  • Biostatistics
  • Survival Analysis
  • Cancer Research

Background:

  • Traditional 5-year survival probability is a limited measure of cure in long-term survival data.
  • Mixture models offer predictive probabilities of cure, providing a more realistic assessment of treatment effectiveness.
  • Existing cure models require extension to accommodate competing risks data.

Purpose of the Study:

  • To develop an extended cure model for competing risks data, enhancing the prediction of long-term patient survival.
  • To compare the predictive accuracy of the extended cure model against the traditional 5-year survival rate.
  • To apply the model to leukemia patient data following bone marrow transplant to estimate cure probabilities.

Main Methods:

  • Developed a finite mixture model extending cure models to competing risks, without assuming independence of cause-specific failure times.

Related Experiment Videos

  • Utilized Gibbs sampling to impute the status of cure via posterior predictive probability, defining cure as near-zero failure risk.
  • Applied the model to a dataset of leukemia patients undergoing bone marrow transplant, considering outcomes of cure, relapse, or non-relapse death.

    • Main Results:

    • The extended cure model estimated a 27% probability of cure for leukemia patients with acute graft-versus-host disease (GVHD) and 46% for the non-GVHD group.
    • For non-cured patients, the probability of relapse was 0.50 in the non-GVHD group and 0.34 in the GVHD group.
    • The non-GVHD group demonstrated a better survival chance, while the GVHD group had a lower relapse chance, illustrating the GVHD-versus-leukemia effect.

    Conclusions:

    • The extended cure model provides a more realistic measure of treatment effectiveness and long-term survival prediction compared to the 5-year survival rate.
    • The model successfully estimated cure probabilities in leukemia patients post-bone marrow transplant, highlighting differences based on GVHD status.
    • Findings underscore the importance of considering competing risks and the GVHD-versus-leukemia effect in survival analysis for cancer patients.