CASCR: Semiparametric Causal Mediation Modeling of Semi-Competing...

In eric40065/CausalAnalysisforSemiCompRisks: Semi-parametric causal mediation modeling of semicompeting risks

Description

This function analyzes semicompeting risks data and gives the estimators of direct and indirect effects, along with their variances.

Usage

CASCR(
  df,
  effect = c("DE", "IE"),
  intervention = c(1, 0),
  cal_level = "median",
  myunit = "raw",
  downsample = 1,
  sen_ana = FALSE,
  get_variance = c("asymptotic"),
  boot_times = 1000,
  timer = TRUE,
  num_of_cores = 1,
  plot_result = FALSE,
  variance_method = "new",
  threshold = 1e-10
)

Arguments

df	a data frame with designated column order: T1, T2, d1, d2, Z, X where X maybe empty or a matrix.
effect	the causal effect to be estimated. Choices are "DE" or "IE". Default is c("DE", "IE")
intervention	numeric vector with two elements. Default is c(1, 0).
cal_level	the level X should be evaluated at. It can be a numeric vector with length(cal_level) = dim(X)[2] or "median", "mean", etc. Default is "median".
myunit	the length to be considered as one unit. Default is "raw".
downsample	indicates how many consecutive event should be considered as one event. Default is 1.
sen_ana	doing sensitivity analysis or not. Default is FALSE.
get_variance	the method to compute the variance. Choices are "a" or "b". Default is "a".
boot_times	the times of bootstrap. Default is 1000.
timer	will show the progress. Default is TRUE.
num_of_cores	the number of cores assigned. Default is 1.
plot_result	will show some primary result. Default is FALSE.
threshold	specifies if a logistic regression converge or not. Default is 1e-10.

Value

CASCR returns a list with components specified by effect.

eric40065/CausalAnalysisforSemiCompRisks documentation built on Dec. 20, 2021, 5:28 a.m.