frailty: Example of a simulation study on frailty survival models

In ellessenne/interest: Interactively Explore Results from Simulation Studies

Description

A dataset from a simulation study comparing frailty flexible parametric models fitted using penalised likelihood to semiparametric frailty models. Both models are fitted assuming a Gamma and a log-Normal frailty. One thousand datasets were simulated, each containing a binary treatment variable with a log-hazard ratio of -0.50. Clustered survival data was simulated assuming 50 clusters of 50 individuals each, with a mixture Weibull baseline hazard function and a frailty following either a Gamma or a Log-Normal distribution. The comparison involves estimates of the log-treatment effect, and estimates of heterogeneity (i.e. the estimated frailty variance).

Usage

frailty

Format

A data frame with 16,000 rows and 6 variables:

• i Simulated dataset number.

• b Point estimate.

• se Standard error of the point estimate.

• par The estimand. trt is the log-treatment effect, fv is the variance of the frailty.

• fv_dist The true frailty distribution.

• model Method used (Cox, Gamma, Cox, Log-Normal, RP(P), Gamma, or RP(P), Log-Normal).

Examples

data("frailty", package = "rsimsum")

ellessenne/interest documentation built on April 29, 2020, 1:07 p.m.