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

Measures of explained variation for survival data.

E L Korn1, R Simon

  • 1Department of Biomathematics, UCLA School of Medicine 90024.

Statistics in Medicine
|May 1, 1990
PubMed
Summary
This summary is machine-generated.

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

Evaluation of the contribution of randomised cancer clinical trials evaluating agents without documented single-agent activity.

ESMO open·2020
Same author

Interim monitoring for non-inferiority trials: minimizing patient exposure to inferior therapies.

Annals of oncology : official journal of the European Society for Medical Oncology·2017
Same author

Assessing treatment efficacy in the subset of responders in a randomized clinical trial.

Annals of oncology : official journal of the European Society for Medical Oncology·2017
Same author

Statistical controversies in clinical research: assessing pathologic complete response as a trial-level surrogate end point for early-stage breast cancer.

Annals of oncology : official journal of the European Society for Medical Oncology·2015
Same author

Response to comment on: 'evaluation of chemoresponse assays as predictive biomarkers'.

British journal of cancer·2015
Same author

Evaluation of chemoresponse assays as predictive markers.

British journal of cancer·2015

Quantifying prognostic model predictive power is crucial. This study proposes explained variation measures, which are model-based and incorporate a time range, offering a better assessment than statistical significance alone.

Area of Science:

  • Biostatistics
  • Epidemiology
  • Clinical Research

Background:

  • Assessing the predictive power of prognostic variables in survival models is essential for clinical decision-making.
  • Existing methods often focus on statistical significance or model fit, which may not fully capture predictive performance.
  • There is a need for robust measures to quantify how well a prognostic model predicts outcomes over time.

Purpose of the Study:

  • To highlight the importance of quantifying the predictive power of prognostic models.
  • To introduce and discuss measures of explained variation as a method for quantifying predictive power.
  • To present a novel approach for measuring predictive power that is model-based and time-dependent.

Main Methods:

  • Developed measures of explained variation specifically for survival time models.

Related Experiment Videos

  • Ensured the proposed measures are entirely model-based, relying on the model's predictions.
  • Incorporated the ability to specify a relevant time range for evaluating predictive power.
  • Utilized null models derived from mixtures of predicted distributions for comparison.

    • Main Results:

    • Proposed measures of explained variation provide a quantitative assessment of a prognostic model's predictive power.
    • The measures are flexible, allowing for the incorporation of a specific time horizon of interest.
    • The model-based nature of the measures ensures they are directly related to the model's predictions.

    Conclusions:

    • Quantifying predictive power is distinct from and essential alongside statistical significance and model fit.
    • Measures of explained variation offer a valuable tool for assessing the predictive performance of survival models.
    • The proposed approach enhances the interpretability and utility of prognostic models in research and practice.