Related Experiment Videos
A Weibull regression model with gamma frailties for multivariate survival data
S K Sahu1, D K Dey, H Aslanidou
1Statistical Laboratory, University of Cambridge, UK. s.sahu@statslab.cam.ac.uk
Lifetime Data Analysis
|January 1, 1997
Summary
This study introduces a new family of survival analysis models using frailty random effects to account for correlated survival times. A flexible baseline hazard model is proposed and compared to the standard Weibull model using Bayesian methods.
Area of Science:
- Biostatistics
- Survival Analysis
- Statistical Modeling
Background:
- Survival data often exhibits correlations due to unobserved covariates.
- Frailty models incorporate these unobserved factors as random effects.
- Dependency in survival times is crucial for accurate analysis.
Purpose of the Study:
- Introduce a new family of conditional proportional hazards models with frailty.
- Propose a flexible baseline hazard model using a correlated prior process.
- Compare the proposed model with the standard Weibull model for survival data.
Main Methods:
- Utilized conditional proportional hazards models with gamma frailty.
- Developed a flexible baseline hazard model based on a correlated prior process.
- Employed Markov Chain Monte Carlo (MCMC) methods for analysis and Bayesian model selection criteria for comparison.
Main Results:
- The proposed flexible baseline hazard model offers an alternative to the standard Weibull model.
- Model diagnostics and Bayesian model selection criteria were developed and applied.
- The methodologies were successfully applied to the kidney infection dataset.
Conclusions:
- The new frailty-based models provide a robust framework for analyzing correlated survival data.
- The proposed flexible baseline hazard model demonstrates utility and comparability with existing methods.
- The study highlights the importance of accounting for frailty in survival analysis.