Skip to contents

Perform a grid search for a particular target by simulating a grid of doses

dose_grid_search.Rd

Set refine = TRUE if the model is nonlinear so that the grid search happens iteratively.

Usage

dose_grid_search(
  est_model = NULL,
  regimen,
  target_design = create_target_design(targettype = "conc", targetvalue = 10, time = 24),
  auc_comp = NULL,
  pta = NULL,
  omega = NULL,
  ruv = NULL,
  dose_update = 1,
  grid = seq(1, 6000, by = 10),
  grid_type = "dose",
  dose_resolution = 1,
  refine = NULL,
  refine_range = c(0.7, 1.4),
  check_boundaries = TRUE,
  max_dose = NULL,
  min_dose = NULL,
  md = list(),
  parameters = NULL,
  covariates = NULL,
  verbose = FALSE,
  ...
)

Arguments

est_model

model used for estimation ("clinician facing")

regimen

PKPDsim regimen object

target_design

object specifying target design, created using function create_target_design()

auc_comp

auc compartment (starting from 1, R-style not C-style!)

pta

probability of target attainment, list with arguments type and value, also requires omega if non-NULL. If NULL, will just aim for specific conc or auc.

omega

IIV matrix, for estimation model, for probability of target attainment target types.

ruv

list specifying residual error for estimation model: list(prop = 0.1, add = 1.5), for probability of target attainment target types.

dose_update

update dose from which dose?

grid

vector specifying doses or intervals to use as test grid, Example: seq(from = 50, to = 500, by = (500 - 50) / 10)

grid_type

either "dose" or "interval"

dose_resolution

to which precision should the output be rounded (e.g. 50), useful when in practice only a specific set of dose units. Can of course also be controlled by altering the grid.

refine

should the found optimal dose be refined more? If not specified, will refine if the model linearity (in attr(model, "misc")) is not described as "linear"

refine_range

after initial optimization, should a second refinement step be implemented? If refine_range is specified e.g. as c(0.9, 1.1) then it will implement a second optimization using a grid spanning from 90% to 110% of the initial optimal dose. Useful only for non-linear models.

check_boundaries

if optimal dose is at lower/upper boundary of grid, should grid be expanded?

max_dose

maximum dose cap

min_dose

minimum dose cap

md

metadata object (only needed if we have to use get_quantity_from_variable() to generate target value)

parameters

list of model parameters

covariates

covariates object

verbose

verbose output?

...

passed on to PKPDsim function

Value

A numeric value indicating the recommended dose