Implementing New Methods in `impart`

impart has built-in methods for covariate adjustment that allow for continuous and binary outcomes, but is designed to be able to use methods from other R packages. This vignette demonstrates how to use new methods with impart, allowing new software to be used in information monitoring and group sequential designs.

library(speff2trial)
#> Error in library(speff2trial): there is no package called 'speff2trial'
# Extract two treatment arms: Lev+5FU (Chemotherapy) and Obs (Observation)
colon_cancer_5fu_vs_obs <-
  subset(
    x = colon_cancer,
    arm %in% c("Lev+5FU", "Obs")
  ) |>
  droplevels()
#> Error: object 'colon_cancer' not found
  

sp_unadjusted <-
  survival::coxph(
    formula =
      survival::Surv(time = years_to_death, event = event_death) ~ arm,
    data = colon_cancer_5fu_vs_obs
  )
#> Error: object 'colon_cancer_5fu_vs_obs' not found

sp_full <-
  speff2trial::speffSurv(
    formula =
      survival::Surv(time = years_to_death, event = event_death) ~
      age + sex + obstruction + perforation + organ_adherence + positive_nodes +
      differentiation +local_spread,
    data = colon_cancer_5fu_vs_obs,
    trt.id = "arm",
    fixed = TRUE
  )
#> Error in loadNamespace(x): there is no package called 'speff2trial'