In metrumresearchgroup/pmtables: Tables for Pharmacometrics

knitr::opts_chunk$set(
collapse = TRUE,
comment = "."
)

pmtables is a package for R that automates the building and formatting of some tables that we commonly create when reporting modeling and simulation results in pharmacometrics.

Tables

pmtable currently supports three main types of tables

1. continuous covariate summary (see continuous-table.html)
2. categorical / discrete covariate summary (see discrete-table.html)
3. data set inventory summary (number of observations, number of individuals, etc) (see data-inventory.html)

pmtables receives your data set along with some options, summarizes the data in a very specific way and creates a very specifically formatted table. In this sense, pmtables is like pmplots: it is not a general grammar of tables (replacing gt or mrggt or kable etc) but the goal is to make a very specific table. We designed pmtables to make some very common tables in a very standard way. If pmtables won't / can't make the table that you want, then the user should use tidyverse + mrggt to create a custom data summary and formatted table.

Vocabulary

cols / by

For most tables, you will specify column names in two of four groups:

• cols is a character vector (or quosure) naming columns in the data set that you would like summarized
• by is one grouping variable for the summary; by should be just a single value
• panel is another grouping variable for the summary in a long format
• span identifies stratification variable for discrete data in long format (see below)

wide / long

You will also see that some tables are "wide" and some are "long": wide indicates that the summarized covariates / values run across the table (west to east) and long indicates that the covariates run along the left hand side of the table (north to south).

Example - wide

data <- pmtables:::data("id")

library(pmtables)
library(dplyr)

For example, we can use cols and by to create a summary of continuous covariates (weight, serum creatinine, and albumin) by study ID in wide format:

pt_cont_wide(
  data, 
  cols = vars(WT,SCR,ALB), 
  by = vars("Study ID" = STUDYf)
)

There is also a discrete data table in wide format

pt_cat_wide(
  data, 
  cols = vars(Sex = SEXf, Formulation = FORMf), 
  by = vars("Study ID" = STUDYf)
)

Example - long

We can also make these tables in long format. In the long format, the by option is used up by the variable names; but you can use the panel option to group a continuous summary and the span option to group a categorical summary

The continuous summary would be:

pt_cont_long(
  data, 
  cols = vars(WT,SCR,ALB), 
  panel = vars("Study ID" = STUDYf)
)

and the catetgorical summary would be:

pt_cat_long(
  data, 
  cols = vars(Sex = SEXf, Formulation = FORMf), 
  span = vars("Study ID" = STUDYf)
)

metrumresearchgroup/pmtables documentation built on Oct. 27, 2024, 5:16 p.m.