Sample covariates using multivariate imputation using chained equations (mice)

Usage

sample_covariates_mice(
  data,
  cat_covs = NULL,
  conditional = NULL,
  n_subjects = nrow(data),
  cont_method = "pmm",
  replicates = 1,
  ...
)

data: data.frame (n x p) containing the original, observed, time-invariant covariates (ID should not be included) that will be used to inform the imputation.
cat_covs: character vector containing the names of the categorical covariates in orgCovs.
conditional: list with conditional limits for sampled population, e.g. list("WT" = c(40, 60), "BMI" = c(15, 25)).
n_subjects: number of simulated subjects, default is the number of subjects in the data.
cont_method: method used to predict continuous covariates within mice, default is pmm.
replicates: number of multiple imputations replicates to sample. Default is 1.
...: additional arguments passed to mice::mice() function

data.frame with the simulated covariates, with n_subjects * m rows and p columns

missing values in data must be coded as NA