Simulates interval censored data from regression model with a Weibull baseline

Simulates interval censored data from a regression model with a Weibull baseline distribution. Used for demonstration

Usage

simIC_weib(
  n = 100,
  b1 = 0.5,
  b2 = -0.5,
  model = "ph",
  shape = 2,
  scale = 2,
  inspections = 2,
  inspectLength = 2.5,
  rndDigits = NULL,
  prob_cen = 1
)

Arguments

n

Number of samples simulated

b1

Value of first regression coefficient

b2

Value of second regression coefficient

model

Type of regression model. Options are 'po' (prop. odds) and 'ph' (Cox PH)

shape

shape parameter of baseline distribution

scale

scale parameter of baseline distribution

inspections

number of inspections times of censoring process

inspectLength

max length of inspection interval

rndDigits

number of digits to which the inspection time is rounded to, creating a discrete inspection time. If rndDigits = NULL, the inspection time is not rounded, resulting in a continuous inspection time

prob_cen

probability event being censored. If event is uncensored, l == u

Details

Exact event times are simulated according to regression model: covariate x1 is distributed rnorm(n) and covariate x2 is distributed 1 - 2 * rbinom(n, 1, 0.5). Event times are then censored with a case II interval censoring mechanism with inspections different inspection times. Time between inspections is distributed as runif(min = 0, max = inspectLength). Note that the user should be careful in simulation studies not to simulate data where nearly all the data is right censored (or more over, all the data with x2 = 1 or -1) or this can result in degenerate solutions!

Author

Clifford Anderson-Bergman

Examples

set.seed(1)
sim_data <- simIC_weib(n = 500, b1 = .3, b2 = -.3, model = 'ph',
                      shape = 2, scale = 2, inspections = 6,
                      inspectLength = 1)
#simulates data from a cox-ph with beta Weibull distribution.

ic_sp_ph(Surv(l, u, type = 'interval2') ~ x1 + x2, data = sim_data)
#> Call:
#> ic_sp_ph(formula = Surv(l, u, type = "interval2") ~ x1 + x2, 
#>     data = sim_data)
#> 
#> Coefficients:
#>          x1          x2 
#>  0.04103429 -0.35065942 
#> 
#> Log-likelihood: -753.3359
#> Number of iterations: 9
ic_sp_po(Surv(l, u, type = 'interval2') ~ x1 + x2, data = sim_data)
#> Call:
#> ic_sp_po(formula = Surv(l, u, type = "interval2") ~ x1 + x2, 
#>     data = sim_data)
#> 
#> Coefficients:
#>          x1          x2 
#> -0.08955382  0.47309560 
#> 
#> Log-likelihood: -759.4884
#> Number of iterations: 8

#'ph' fit looks better than 'po'; the difference between the transformed survival
#function looks more constant