discfrail-package: discfrail: Cox Models for Time-to-Event Data with...

Description Details Author(s) References See Also

Description

discfrail: Functions for fitting Cox proportional hazards models for grouped time-to-event data, where the shared group-specific frailties have a discrete nonparametric distribution.

Details

The methods are fully described and illustrated in Gasperoni et al. (2018).

Groups of individuals are clustered into a number of latent populations of groups, each with a common frailty that is assumed to act multiplicatively on their hazard. Covariates can be included through proportional hazards, in the manner of the standard Cox model. The baseline hazard is left unspecified as in the Cox model, and estimation is performed by maximum partial likelihood through an EM algorithm.

There are also functions for simulating from these models, and from similar models with a parametric baseline survival function.

npdf_cox fits nonparametric discrete frailty models. The number of latent populations of groups can either be fixed, or estimated through an automated model comparison and selection procedure.

plot.npdf plots fitted survival or cumulative hazard curves from the fitted models.

sim_npdf simulates from a proportional hazards model with an arbitrary baseline survival distribution and discrete shared frailty terms.

sim_weibdf simulates from a Weibull survival model with discrete shared frailty terms.

Author(s)

Maintainer: Francesca Gasperoni francesca.gasperoni@polimi.it

Authors:

References

F. Gasperoni, F. Ieva, A.M. Paganoni, C. Jackson, L. Sharples. (2018) Nonparametric frailty Cox models for hierarchical time-to-event data. Biostatistics.

See Also

Useful links:


discfrail documentation built on May 2, 2019, 9:58 a.m.