.png){kind=icon}
.png){kind=icon}
Return
Efficient estimation of random effects Cox models with application to lung cancer data
.png){kind=icon}
PRE
AI
G
S
M
DOI：10.1080/00949655.2026.2642825.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Analysing right-censored clustered survival data is critical when studying treatment differences across correlated observations in biomedical research, such as multi-centre clinical trials, recurrent events and environmental studies. The Cox proportional hazards (PH) model is widely used for such data; however, some of the covariates in the model may not be relevant for accurately predicting survival times. To address this, we explore pretest and shrinkage estimation methods for the Cox PH model. We employ two models: an unrestricted model encompassing all covariates and a restricted model containing a smaller subset. By optimally combining estimators from both, we establish pretest and shrinkage estimators. Their performance is assessed using mean squared error (MSE) and relative MSE. Through extensive simulation studies and an application to lung cancer data, we demonstrate that the proposed shrinkage estimators exhibit lower risk compared to estimators derived from the full model when the shrinkage dimension exceeds two.
Keywords:
Proportional hazard
partial likelihood
random effects
shrinkage and pretest
time-to-event data
Journal
J
IF:
1.2
Papers:
114
Citations:
4.1K
Organization
Cited Papers