Simulate a Node Using Parametric Survival Models

Data from the parents is used to generate the node using either an accelerated failure time model, an accelerated hazard model, an extended hazard model, a proportional odds model or a Young and Prentice model, as implemented in the rsurv package (Demarqui 2024).

Usage

node_aftreg(data, parents, betas, baseline, dist=NULL,
            package=NULL, u=stats::runif(nrow(data)),
            cens_dist=NULL, cens_args, name, as_two_cols=TRUE,
            ...)

node_ahreg(data, parents, betas, baseline, dist=NULL,
           package=NULL, u=stats::runif(nrow(data)),
           cens_dist=NULL, cens_args, name, as_two_cols=TRUE,
           ...)

node_ehreg(data, parents, betas, phi, baseline, dist=NULL,
           package=NULL, u=stats::runif(nrow(data)),
           cens_dist=NULL, cens_args, name, as_two_cols=TRUE,
           ...)

node_poreg(data, parents, betas, baseline, dist=NULL,
           package=NULL, u=stats::runif(nrow(data)),
           cens_dist=NULL, cens_args, name, as_two_cols=TRUE,
           ...)

node_ypreg(data, parents, betas, phi, baseline, dist=NULL,
           package=NULL, u=stats::runif(nrow(data)),
           cens_dist=NULL, cens_args, name, as_two_cols=TRUE,
           ...)

Arguments

data: A data.table (or something that can be coerced to a data.table) containing all columns specified by parents. Passed to the argument of the same name in raftreg, rahreg, rehreg, rporeg or rypreg.
parents: A character vector specifying the names of the parents that this particular child node has. Converted into a formula and passed to the argument of the same name in raftreg, rahreg, rehreg, rporeg or rypreg.
betas: A numeric vector with length equal to parents, specifying the causal beta coefficients used to generate the node. Passed to the beta argument in raftreg, rahreg, rehreg, rporeg or rypreg.
phi: A numeric vector with length equal to parents, specifying the phi beta coefficients used to generate the node. Only required for extended hazard and Yang and Prentice models. Passed to the phi argument in rehreg or rypreg.
baseline: A single character string, specifying the name of the baseline survival distribution. Passed to the argument of the same name in raftreg, rahreg, rehreg, rporeg or rypreg.
dist: An alternative way to specify the baseline survival distribution. Passed to the argument of the same name in raftreg, rahreg, rehreg, rporeg or rypreg.
package: A single character string, specifying the name of the package where the assumed quantile function is implemented. Passed to the argument of the same name in raftreg, rahreg, rehreg, rporeg or rypreg.
u: A numeric vector of quantiles of length nrow(data). Usually this should simply be passed a vector of randomly generated uniformly distributed values between 0 and 1, as defined by the default. Passed to the argument of the same name in raftreg, rahreg, rehreg, rporeg or rypreg.
cens_dist: A single character naming the distribution function that should be used to generate the censoring times or a suitable function. For example, "runif" could be used to generate uniformly distributed censoring times. Set to NULL to get no censoring.
cens_args: A list of named arguments which will be passed to the function specified by the cens_dist argument.
name: A single character string specifying the name of the node.
as_two_cols: Either TRUE or FALSE, specifying whether the output should be divided into two columns. When cens_dist is specified, this argument will always be treated as TRUE because two columns are needed to encode both the time to the event and the status indicator. When no censoring is applied, however, users may set this argument to FALSE to simply return the numbers as they are.
...: Further arguments passed to raftreg, rahreg, rehreg, rporeg or rypreg.

Details

Survival times are generated according to the specified parametric survival model. The actual generation of the values is done entirely by calls to the rsurv package. All arguments are directly passed to the corresponding function in rsurv. Only the censoring is added on top of it. These convenience wrappers only exist to allow direct integration of these data generation functions with the interface provided by simDAG. Please consult the documentation and excellent paper by Demarqui (2024) for more information about the models and how to specify the arguments.

References

Demarqui Fabio N. Simulation of Survival Data with the Package rsurv. (2024) arXiv:2406.01750v1.

Author

Robin Denz

Value

Returns a data.table of length nrow(data) containing two columns if as_two_cols=TRUE and always when cens_dist is specified. In this case, both columns start with the nodes name and end with _event and _time. The first is a logical vector, the second a numeric one. If as_two_cols=FALSE and cens_dist is NULL, a numeric vector is returned instead.

Examples

library(simDAG)

set.seed(3454)

if (requireNamespace("rsurv")) {

library(rsurv)

# accelerated failure time model
dag <- empty_dag() +
  node(c("A", "B", "C"), type="rnorm", mean=0, sd=1) +
  node("Y", type="aftreg", formula= ~ -2 + A*0.2 + B*0.1 + A:B*1,
       baseline="weibull", shape=1, scale=2)
data <- sim_from_dag(dag, n_sim=100)

# accelerated hazard model
dag <- empty_dag() +
  node(c("A", "B", "C"), type="rnorm", mean=0, sd=1) +
  node("Y", type="ahreg", formula= ~ -2 + A*0.2 + B*0.1,
       baseline="weibull", shape=1, scale=2)
data <- sim_from_dag(dag, n_sim=100)

# extended hazard model
dag <- empty_dag() +
  node(c("A", "B", "C"), type="rnorm", mean=0, sd=1) +
  node("Y", type="ehreg", formula= ~ -2 + A*0.2 + B*0.1,
       baseline="weibull", shape=1, scale=2,
       phi=c(-1, 1))
data <- sim_from_dag(dag, n_sim=100)

# proportional odds model
dag <- empty_dag() +
  node(c("A", "B", "C"), type="rnorm", mean=0, sd=1) +
  node("Y", type="poreg", formula= ~ -2 + A*0.2 + B*0.1,
       baseline="weibull", shape=1, scale=2)
data <- sim_from_dag(dag, n_sim=100)

# Young and Prentice model
dag <- empty_dag() +
  node(c("A", "B", "C"), type="rnorm", mean=0, sd=1) +
  node("Y", type="ypreg", formula= ~ -2 + A*0.2 + B*0.1,
       baseline="weibull", shape=1, scale=2,
       phi=c(-1, 1))
data <- sim_from_dag(dag, n_sim=100)
}
#> Loading required namespace: rsurv