vigndata: Simulated data set used in the blupsurv package vignette

In blupsurv: Recurrent events with BLUPs

Description

A simulated data set for illustrative use in the package vignette. To make the discussion less abstract, covariates were given real-world names. The data represent a 10-year study during which patients in the 50 US states were monitored for events of severe pain or fever. The data were generated as follows:

• 50 clusters, labeled with US state names

• between 5 and 25 subjects per cluster

• 1 stratum

• frailties for the pain process are lognormal with cluster and subject- level variance 0.25

• frailties for the fever process are lognormal with cluster and subject- level variance 0.5

• subject-level frailty covariance is 0.15

• age is a rounded Normal variable with mean 50 and variance 10

• sex is a Bernoulli variable with mean 0.5

• true regression coefficients for the pain process are 0 for age and 1 for sex

• true regression coefficient for the fever process are 0.025 for age and 2 for sex

• both baseline hazards are Weibull with lambda=0., gamma=1.8

• censoring times are fixed at 10

used for package examples. The object vigndata.fit contains the fitted bivariate model for this data with full discretization.

Usage

vigndata

Format

vigndata is a data frame of 8 columns and 1982 rows, vigndata.fit and vigndata.quickfit are objects of class bivrec.

See Also

bivrec

blupsurv documentation built on May 2, 2019, 6:51 p.m.