How to use tidycoRe for CorEvitas table and report formatting • tidycoRe

Welcome to tidycoRe!

A formatting package to create standardized CorEvitas R query projects, tables, and reports.

Get started

To get started, let’s load the package into our working session:

library(tidycoRe)
library(tidyverse)
library(arsenal)

If you have not downloaded it yet from our private github repo, please see the Readme on the package home page.

Create standard-format summary tables

tidycoRe comes equipped with functions to format arsenal summary tableby objects and regular dataframes into the standardized CorEvitas style flextable objects for word, html, and pdf documents.

Tableby objects are an important part of tidycoRe and additional documentation can be found here: https://cran.r-project.org/web/packages/arsenal/vignettes/tableby.html.

The formatted tableby object

The format_tableby function converts tableby objects into data frames. The format_tableby_flextable function styles data frames created by format_tableby into summary flextables that can exported via RMarkdown or Quarto into html, pdf, docx, pptx, etc, documents.

Features include:

rounding numeric values to 1 significant digit
condensing rows where numeric variables are summarized by N (if specified (one of the most important benefits)
summarizing categorical variables using their specified summary statistics in the tableby command
if specified, borders are applied

First we will create the tableby object.

demo_tableby <- arsenal::tableby(
  female ~
    age +
    white +
    hispanic +
    insurance_private +
    insurance_medicare +
    insurance_medicaid +
    insurance_none,
  cat.stats = c("countpct"),
  numeric.stats = c("meansd"),
  test = TRUE,
  data = demo_data
)

Which looks like this unformatted:

demo_tableby %>% 
  summary()

	Male (N=110)	Female (N=90)	Total (N=200)	p value
Age (years)				0.721
Mean (SD)	51.236 (12.971)	50.556 (13.869)	50.930 (13.353)
White				0.064
Other race	29 (26.4%)	14 (15.6%)	43 (21.5%)
White	81 (73.6%)	76 (84.4%)	157 (78.5%)
Hispanic Ethnicity				0.300
No	106 (97.2%)	83 (94.3%)	189 (95.9%)
Yes	3 (2.8%)	5 (5.7%)	8 (4.1%)
Private Insurance				0.874
Not reported	23 (20.9%)	18 (20.0%)	41 (20.5%)
Yes	87 (79.1%)	72 (80.0%)	159 (79.5%)
Medicare Insurance				0.767
Not reported	91 (82.7%)	73 (81.1%)	164 (82.0%)
Yes	19 (17.3%)	17 (18.9%)	36 (18.0%)
Medicaid Insurance				0.917
Not reported	101 (91.8%)	83 (92.2%)	184 (92.0%)
Yes	9 (8.2%)	7 (7.8%)	16 (8.0%)
No Insurance				0.839
Not reported	108 (98.2%)	88 (97.8%)	196 (98.0%)
Yes	2 (1.8%)	2 (2.2%)	4 (2.0%)

Versus using format_tableby() below:

format_tableby(
  tableby_data = demo_tableby,
  condense = TRUE
) %>% 
  format_tableby_flextable()

	Male (N=110)	Female (N=90)	Total (N=200)	p value
Age (years)	51.2 (13.0)	50.6 (13.9)	50.9 (13.4)	0.721
White				0.064
Other race	29 (26.4%)	14 (15.6%)	43 (21.5%)
White	81 (73.6%)	76 (84.4%)	157 (78.5%)
Hispanic Ethnicity	3 (2.8%)	5 (5.7%)	8 (4.1%)	0.300
Private Insurance	87 (79.1%)	72 (80.0%)	159 (79.5%)	0.874
Medicare Insurance	19 (17.3%)	17 (18.9%)	36 (18.0%)	0.767
Medicaid Insurance	9 (8.2%)	7 (7.8%)	16 (8.0%)	0.917
No Insurance	2 (1.8%)	2 (2.2%)	4 (2.0%)	0.839

Important notes about condense = TRUE

If more than one summary statistic is specified for categorical data, the rows will not be condensed for categorical data
The variable rows will only be condensed if it has a ‘Yes’ value and the one other value is: ‘No’, ‘Not reported’, ‘Unknown’, ‘Unk’. The variable must only have two levels where Yes is one of the levels and the other is from the list previously mentioned.
Numeric summary statistics will pull the first numeric statistic and place it on the variable row

The formatted dataframe

This function styles dataframes into flextables applicable for html, pdf, or word documents.

Features include:

if specified, borders are applied

format_basictable(
  data = tidycoRe::demo_data %>%
    select(
      biohxgroup,
      age,
      female,
      white,
      hispanic,
      insurance_private,
      insurance_medicare,
      insurance_medicaid,
      insurance_none) %>% 
    head(10),
  borders = TRUE
)

biohxgroup	age	female	white	hispanic	insurance_private	insurance_medicare	insurance_medicaid	insurance_none
Prior other Bio Non-Failure	35	Female	Other race	No	Yes	Not reported	Not reported	Not reported
Prior other Bio Non-Failure	72	Male	White	No	Not reported	Yes	Not reported	Not reported
Prior SEC Failure	50	Female	White	No	Yes	Not reported	Not reported	Not reported
Prior SEC Failure	49	Male	White	Yes	Yes	Not reported	Not reported	Not reported
Prior other Bio Failure	61	Female	White	No	Yes	Not reported	Not reported	Not reported
Prior SEC Failure	50	Male	Other race	No	Yes	Not reported	Not reported	Not reported
Prior other Bio Failure	45	Female	White	No	Yes	Not reported	Not reported	Not reported
Prior other Bio Failure	66	Male	White	No	Not reported	Yes	Not reported	Not reported
Prior other Bio Failure	69	Male	White	No	Not reported	Yes	Not reported	Not reported
Prior SEC Failure	39	Female	White	No	Yes	Not reported	Not reported	Not reported

Tips on formatting flextables

In general, other than headers and footers, there is no easy way to add additional rows to a flextable object or flextables together. This needs to be done when the objects are still data frames, i.e., before using ‘format_basictable()’, or after ‘format_tableby()’ and before ‘format_tableby_flextable()’.

You can pull a dataframe object out of a flextable using the following code:

table_name$body$dataset

to extract and manipulate before reapplying the flextable formatting with the above tidycoRe functions.