Plot Treatment Differences in the Linear Predictor

plot_lp_difference() evaluates a fitted ordinal Markov model at one patient profile and plots differences in the model linear predictor between counterfactual levels of one variable over time. The plot is conditional on the supplied profile and faceted by previous state.

Usage

plot_lp_difference(
  x,
  profile,
  variables,
  times,
  previous_states = NULL,
  y_levels = NULL,
  absorb = NULL,
  time_var = "time",
  p_var = "yprev",
  gap_var = NULL,
  time_covariates = NULL,
  line_width = 0.7
)

Arguments

x: A fitted orm, blrm, vglm, or robcov_vglm Markov model.
profile: One-row data frame containing the covariate profile.
variables: Named list with one variable and at least two values. The first value is the reference.
times: Time values at which to evaluate the linear predictor.
previous_states: Optional previous-state values to facet by. Defaults to all non-absorbing states in y_levels.
y_levels: Optional state levels. If omitted, levels are inferred from the model when possible.
absorb: Optional absorbing-state label excluded from the default previous-state grid.
time_var: Character name of the time column.
p_var: Character name of the previous-state column.
gap_var: Optional time-gap_var variable used by the fitted model.
time_covariates: Optional time-varying covariate lookup table used by the fitted model.
line_width: Line width for the plotted contrasts.

Value

A ggplot object.

Details

plot_lp_difference() is model-based only. At minimum, supply model, a one-row profile, variables, and times. The profile should contain the model covariates for the patient or scenario being plotted; times and previous_states define the plotting grid.

Examples

if (FALSE) { # \dontrun{
plot_lp_difference(
  fit,
  profile = data.frame(tx = 0, age = 59, sex = "Male", yprev = 3),
  variables = list(tx = c(0, 1)),
  times = 1:28,
  y_levels = 1:8,
  absorb = 8
)
} # }