gtsummary

Using gtsummary package to create publication ready tables

Authors

Hannah Bowley

Micaela Lembo

What is gtsummary?

  • R tool that creates publication ready tables

  • Summarizes data sets, regression models and other analysis

  • Highly customizable

gtsummary

Install gtsummary package

# install.packages("gtsummary")

Data Source: gapminder

# install.packages("gapminder")

Load in libraries

library(gtsummary)
library(gapminder)
library(tidyverse)

About gapminder

Code 
head(gapminder)
# A tibble: 6 × 6
  country     continent  year lifeExp      pop gdpPercap
  <fct>       <fct>     <int>   <dbl>    <int>     <dbl>
1 Afghanistan Asia       1952    28.8  8425333      779.
2 Afghanistan Asia       1957    30.3  9240934      821.
3 Afghanistan Asia       1962    32.0 10267083      853.
4 Afghanistan Asia       1967    34.0 11537966      836.
5 Afghanistan Asia       1972    36.1 13079460      740.
6 Afghanistan Asia       1977    38.4 14880372      786.
Code 
str(gapminder)
tibble [1,704 × 6] (S3: tbl_df/tbl/data.frame)
 $ country  : Factor w/ 142 levels "Afghanistan",..: 1 1 1 1 1 1 1 1 1 1 ...
 $ continent: Factor w/ 5 levels "Africa","Americas",..: 3 3 3 3 3 3 3 3 3 3 ...
 $ year     : int [1:1704] 1952 1957 1962 1967 1972 1977 1982 1987 1992 1997 ...
 $ lifeExp  : num [1:1704] 28.8 30.3 32 34 36.1 ...
 $ pop      : int [1:1704] 8425333 9240934 10267083 11537966 13079460 14880372 12881816 13867957 16317921 22227415 ...
 $ gdpPercap: num [1:1704] 779 821 853 836 740 ...

Create basic gtsummary table

table1 <- 
  gapminder %>%
  # Year we are interested in
  filter(year == 2007) %>% 
  # Columns we want to look at
  select( gdpPercap, lifeExp, pop ) %>% 
  # function to create the summary table
  tbl_summary()

table1
Characteristic N = 1421
gdpPercap 6,124 (1,625, 18,009)
lifeExp 72 (57, 76)
pop 10,517,531 (4,508,034, 31,210,042)
1 Median (IQR)

Customization option: by()

  • by() : allows users to select which variable they would like to use for group comparison.

    • In the example below, we chose continents, but another option could have been countries.
table1 <- 
  gapminder %>%
  filter(year == 2007) %>%  
  select( gdpPercap, lifeExp, pop ,continent ) %>% 
  tbl_summary(
    # group by continent
    by = continent  
  )

table1
Characteristic Africa, N = 521 Americas, N = 251 Asia, N = 331 Europe, N = 301 Oceania, N = 21
gdpPercap 1,452 (863, 3,994) 8,948 (5,728, 11,978) 4,471 (2,452, 22,316) 28,054 (14,812, 33,818) 29,810 (27,498, 32,123)
lifeExp 53 (48, 59) 73 (72, 76) 72 (65, 76) 79 (75, 80) 81 (80, 81)
pop 10,093,311 (2,909,227, 19,363,655) 9,319,622 (5,675,356, 28,674,757) 24,821,286 (6,426,679, 69,453,570) 9,493,598 (4,780,560, 20,849,695) 12,274,974 (8,195,372, 16,354,575)
1 Median (IQR)

Customization option: label(), modify_header(), bold_labels()

  • label(): Allows you to rename the variables from their default names within dataset
  • modify_header(): Change the default header label within your chart
    • in the example below we changed it to the word “variables” bolded
  • bold_labels(): Makes the rows and column names bold
table1 <- 
  gapminder %>%
  filter(year == 2007) %>% 
  select( gdpPercap, lifeExp, pop ,continent ) %>% 
  tbl_summary(
    by = continent, 
    # Changing the name of our rows 
    label = list( 
      gdpPercap ~ "GDP per capita",
      lifeExp ~ "Life expectancy",
      pop ~ "Population"
      ) 
    ) %>% 
  # Change header name
  modify_header(label = "**Variables**")  %>% 
  # Bold row and column names
  bold_labels()
 

table1
Variables Africa, N = 521 Americas, N = 251 Asia, N = 331 Europe, N = 301 Oceania, N = 21
GDP per capita 1,452 (863, 3,994) 8,948 (5,728, 11,978) 4,471 (2,452, 22,316) 28,054 (14,812, 33,818) 29,810 (27,498, 32,123)
Life expectancy 53 (48, 59) 73 (72, 76) 72 (65, 76) 79 (75, 80) 81 (80, 81)
Population 10,093,311 (2,909,227, 19,363,655) 9,319,622 (5,675,356, 28,674,757) 24,821,286 (6,426,679, 69,453,570) 9,493,598 (4,780,560, 20,849,695) 12,274,974 (8,195,372, 16,354,575)
1 Median (IQR)

Customization change default statistics: statistics (), add_p()

  • statistics(): customize to calculate specified summary statistic; default option for continuous variables is median{IQR}
    • statistics options continuous variables:
      • median, mean, sd, var, min, max, sum, p##: any integer percentage, foo: any function of the form foo
    • statistics options categorical variables:
      • n: frequency, N: denominator or sample size, p: percentage
  • add_p(): Add in p values as a new column; default adds p-value as last column in table
table1 <- 
  gapminder %>%
  filter(year == 2007) %>% 
  select( gdpPercap, lifeExp, pop ,continent ) %>% 
  tbl_summary(
    by = continent, 
    label = list(
      gdpPercap ~ "GDP per capita",
      lifeExp ~ "Life expectancy",
      pop ~ "Population"
      ) ,
    # Calculate mean for gdpPercap variable
    statistic = list(gdpPercap ~ "{mean}")
    ) %>% 
  modify_header(label = "**Variables**") %>% 
  add_p() 

table1
Variables Africa, N = 521 Americas, N = 251 Asia, N = 331 Europe, N = 301 Oceania, N = 21 p-value2
GDP per capita 3,089 11,003 12,473 25,054 29,810 <0.001
Life expectancy 53 (48, 59) 73 (72, 76) 72 (65, 76) 79 (75, 80) 81 (80, 81) <0.001
Population 10,093,311 (2,909,227, 19,363,655) 9,319,622 (5,675,356, 28,674,757) 24,821,286 (6,426,679, 69,453,570) 9,493,598 (4,780,560, 20,849,695) 12,274,974 (8,195,372, 16,354,575) 0.044
1 Mean; Median (IQR)
2 Kruskal-Wallis rank sum test

Customization to add a column with overall statistics: add_overall()

  • add_overall(): Gives an additional column with the summary of all observations without grouping established with by() function
table1 <- 
  gapminder %>%
  filter(year == 2007) %>% 
  select( gdpPercap, lifeExp, pop ,continent ) %>% 
  tbl_summary(
    by = continent, 
    label = list (
      gdpPercap ~ "GDP per capita",
      lifeExp ~ "Life expectancy",
      pop ~ "Population"
      ) ,
    statistic = list( gdpPercap ~ "{mean}")
    ) %>% 
  modify_header(label = "**Variables**") %>% 
  add_overall(last = TRUE) %>% 
  add_p()

table1
Variables Africa, N = 521 Americas, N = 251 Asia, N = 331 Europe, N = 301 Oceania, N = 21 Overall, N = 1421 p-value2
GDP per capita 3,089 11,003 12,473 25,054 29,810 11,680 <0.001
Life expectancy 53 (48, 59) 73 (72, 76) 72 (65, 76) 79 (75, 80) 81 (80, 81) 72 (57, 76) <0.001
Population 10,093,311 (2,909,227, 19,363,655) 9,319,622 (5,675,356, 28,674,757) 24,821,286 (6,426,679, 69,453,570) 9,493,598 (4,780,560, 20,849,695) 12,274,974 (8,195,372, 16,354,575) 10,517,531 (4,508,034, 31,210,042) 0.044
1 Mean; Median (IQR)
2 Kruskal-Wallis rank sum test

Using gtsummary for regression models

  1. First a linear regression model needs to be created
  • the standard R summary output (shown below) is not publication ready, but the gtsummary package makes it looked more polished
mod1 <- lm(lifeExp ~ year + continent + continent*year, data = gapminder)

summary(mod1)

Call:
lm(formula = lifeExp ~ year + continent + continent * year, data = gapminder)

Residuals:
     Min       1Q   Median       3Q      Max 
-28.8854  -4.2696   0.3298   3.9835  21.1306 

Coefficients:
                         Estimate Std. Error t value Pr(>|t|)    
(Intercept)            -524.25785   32.96343 -15.904  < 2e-16 ***
year                      0.28953    0.01665  17.387  < 2e-16 ***
continentAmericas      -138.84845   57.85058  -2.400  0.01650 *  
continentAsia          -312.63305   52.90355  -5.909 4.14e-09 ***
continentEurope         156.84685   54.49776   2.878  0.00405 ** 
continentOceania        182.34988  171.28299   1.065  0.28720    
year:continentAmericas    0.07812    0.02922   2.673  0.00758 ** 
year:continentAsia        0.16359    0.02672   6.121 1.15e-09 ***
year:continentEurope     -0.06760    0.02753  -2.455  0.01417 *  
year:continentOceania    -0.07926    0.08653  -0.916  0.35980    
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Residual standard error: 7.18 on 1694 degrees of freedom
Multiple R-squared:  0.6927,    Adjusted R-squared:  0.6911 
F-statistic: 424.3 on 9 and 1694 DF,  p-value: < 2.2e-16

tbl_regression() creates a polished summary table for the linear regression model

  • modify_caption(): adds a title to the table
  • bold_p(): makes all significant p-values bold
table2 <- tbl_regression(mod1) %>% 
  bold_labels() %>% 
  bold_p() %>% 
  # includes caption on top of table
  modify_caption("Table 2: Regression results for Life Expectancy")

table2
Table 2: Regression results for Life Expectancy
Characteristic Beta 95% CI1 p-value
year 0.29 0.26, 0.32 <0.001
continent
    Africa — —
    Americas -139 -252, -25 0.016
    Asia -313 -416, -209 <0.001
    Europe 157 50, 264 0.004
    Oceania 182 -154, 518 0.3
year * continent
    year * Americas 0.08 0.02, 0.14 0.008
    year * Asia 0.16 0.11, 0.22 <0.001
    year * Europe -0.07 -0.12, -0.01 0.014
    year * Oceania -0.08 -0.25, 0.09 0.4
1 CI = Confidence Interval

(Sjoberg et al. 2023)

Exporting GT Summary tables to Word Documents

  • gtsummary is now compatible with Word Documents.
  • These are the steps to follow in order to upload your table to a word document
tbl %>%
  as_gt() %>%
  # If you are using word, use extension .docx 
  # alternative options: .html, .png, .pdg, .tex, .rtf
  gt::gtsave(filename = ".")

References

Sjoberg, Daniel D., Joseph Larmarange, Michael Curry, Jessica Lavery, Karissa Whiting, Emily C. Zabor, Xing Bai, et al. 2023. “Gtsummary: Presentation-Ready Data Summary and Analytic Result Tables.” https://cran.r-project.org/web/packages/gtsummary/index.html.