add_covariates {mapbayr}

Add covariate columns to a dataset

Description

The add_covariates() function adds an one or several covariate columns to a dataset provided as a proper data.frame or within a 'mrgsolve' model. Used in combination with adm_rows() and obs_rows(), it helps the creation of datasets in the proper format for simulations with 'mrgsolve' or parameter estimation with 'mapbayr', as explained in data_helpers.

Usage

add_covariates(x, ...)

## S3 method for class 'data.frame'
add_covariates(x, ..., covariates = list(), AOLA = FALSE, TOLA = FALSE)

## S3 method for class 'mrgmod'
add_covariates(x, ..., covariates = list(), AOLA = NULL, TOLA = NULL)

Arguments

x

either a data.frame or a 'mrgsolve' model object

...

covariates values to add to the data. For each variable, supply a vector of length 1 or with the same number of rows. Ignored if covariates argument is used.

covariates

Covariates passed as a single list of variables. Overrides ....

AOLA, TOLA

a logical. Should the "Amount Of Last Administration" and "Time Of Last Administration" variables be added into the dataset? Default if FALSE if x is a dataset, TRUE if x is a 'mrgsolve' model where AOLA and TOLA are defined as covariates

Value

a data.frame, or a 'mrgsolve' model with a dataset in the ⁠@args$data⁠ slot (accessible with get_data()).

See Also

data_helpers

Examples

# Cannot start from scratch
## Not run: 
add_covariates(BW = 90, SEX = 0)

## End(Not run)

library(magrittr)
adm_rows(time = c(0, 24, 48), cmt = 1, amt = c(100, 200, 300)) %>%
  add_covariates(BW = c(90, 85, 80), SEX = 0)

# If covariates are stored in a list, use `covariates = `
adm_rows(time = c(0, 24, 48), cmt = 1, amt = c(100, 200, 300)) %>%
  add_covariates(covariates = list(BW = c(90, 85, 80), SEX = 0))

# Missing values are filled with the "next observation carried backward" rule
adm_rows(time = c(0, 24, 48), cmt = 1, amt = c(100, 200, 300)) %>%
  add_covariates(BW = c(90, 85, 80), SEX = 0) %>%
  obs_rows(time = 36, DV = .0123, cmt = 2)
# Always verify the output in case of time-varying covariates

# Possibility to add Time and Amount of last administration as covariates
adm_rows(time = c(0, 24, 48), amt = c(100, 200, 300), cmt = 1) %>%
  obs_rows(time = c(8, 16, 32, 40), cmt = 2, DV = runif(4)) %>%
  add_covariates(TOLA = TRUE, AOLA = TRUE) %>%
  obs_rows(time = 72, cmt = 2, DV = .123) # AOLA/TOLA re-updated afterwards

# Automatic inclusion of `TOLA`/`AOLA` if they are covariates of the model
library(mrgsolve)
model <- mcode("model", "
$PARAM @annotated @covariates
TOLA : 0 : Time Last Adm
AOLA : 0 : Amount Last Adm
", compile = FALSE)

model %>%
  adm_rows(time = c(0, 24, 48), amt = c(100, 200, 300), cmt = 1) %>%
  add_covariates() %>%
  get_data()