Produce dataset for covariate forest plotting

The main workhorse for computing uncertainty quantiles of covariate effects in different subpopulations.

cov_forest_data(m, covariate_scenarios)

Arguments

m

An nm object.

covariate_scenarios

A data.frame. Need columns cov, value and (optional) text. See details for more information.

Value

dplyr::tibble with quantile information suitable for cov_forest_plot().

Details

The column cov in covariate_scenarios refers to covariate variables in the dataset. The column value refers to covariate values of importance. Typically these will be quantiles of continuous variables and categories (for categorical covariates). The column text is option but is a labelling column for cov_forest_plot() to adjust how the covariate scenarios are printed on the axis

Examples

## requires NONMEM to be installed
if (FALSE) {

dcov <- input_data(m1, filter = TRUE)
dcov <- dcov[!duplicated(dcov$ID), ]

covariate_scenarios <- dplyr::bind_rows(
  dplyr::tibble(cov = "HEALTHGP", value = c(0, 1)),
  dplyr::tibble(cov = "HEPATIC", value = unique(dcov$HEPATIC[dcov$HEPATIC > -99])),
  dplyr::tibble(cov = "BWTIMP", value = c(50, 80, 120)),
  dplyr::tibble(cov = "ECOG", value = c(0, 1, 2, 3)),
  dplyr::tibble(cov = "BEGFRIMP", value = quantile(dcov$BEGFR[dcov$BEGFR > -99])),
  dplyr::tibble(cov = "RACE", value = c(1, 2), text = c("white", "black")),
  dplyr::tibble(cov = "PPI", value = c(0, 1)),
  dplyr::tibble(cov = "H2RA", value = c(0, 1))
)

dplot <- cov_forest_data(m1, covariate_scenarios = covariate_scenarios)
cov_forest_plot(dplot)
}