The tbl_regression()
function takes a regression model object in R and returns a formatted table of regression model results that is publication-ready. It is a simple way to summarize and present your analysis results using R! Like tbl_summary()
, tbl_regression()
creates highly customizable analytic tables with sensible defaults.
This vignette will walk a reader through the tbl_regression()
function, and the various functions available to modify and make additions to an existing formatted regression table.
Behind the scenes: tbl_regression()
uses broom::tidy()
to perform the initial model formatting, and can accommodate many different model types (e.g. lm()
, glm()
, survival::coxph()
, survival::survreg()
and other are vetted models known to work with {gtsummary}). It is also possible to specify your own function to tidy the model results if needed.
Before going through the tutorial, install and load {gtsummary}.
# install.packages("gtsummary")
library(gtsummary)
In this vignette we’ll be using the trial
data set which is included in the {gtsummary package}.
This data set contains information from 200 patients who received one of two types of chemotherapy (Drug A or Drug B).
The outcomes are tumor response and death.
Each variable in the data frame has been assigned an attribute label (i.e. attr(trial$trt, "label") == "Chemotherapy Treatment")
with the labelled package, which we highly recommend using. These labels are displayed in the {gtsummary} output table by default. Using {gtsummary} on a data frame without labels will simply print variable names, or there is an option to add labels later.
Variable | Class | Label |
---|---|---|
|
character | Chemotherapy Treatment |
|
numeric | Age |
|
numeric | Marker Level (ng/mL) |
|
factor | T Stage |
|
factor | Grade |
|
integer | Tumor Response |
|
integer | Patient Died |
|
numeric | Months to Death/Censor |
Includes mix of continuous, dichotomous, and categorical variables |
The default output from tbl_regression()
is meant to be publication ready.
trial
data set.# build logistic regression model
<- glm(response ~ age + stage, trial, family = binomial)
m1
# view raw model results
summary(m1)$coefficients
#> Estimate Std. Error z value Pr(>|z|)
#> (Intercept) -1.48622424 0.62022844 -2.3962530 0.01656365
#> age 0.01939109 0.01146813 1.6908683 0.09086195
#> stageT2 -0.54142643 0.44000267 -1.2305071 0.21850725
#> stageT3 -0.05953479 0.45042027 -0.1321761 0.89484501
#> stageT4 -0.23108633 0.44822835 -0.5155549 0.60616530
tbl_regression(m1, exponentiate = TRUE)
Characteristic | OR1 | 95% CI1 | p-value |
---|---|---|---|
Age | 1.02 | 1.00, 1.04 | 0.091 |
T Stage | |||
T1 | — | — | |
T2 | 0.58 | 0.24, 1.37 | 0.2 |
T3 | 0.94 | 0.39, 2.28 | 0.9 |
T4 | 0.79 | 0.33, 1.90 | 0.6 |
1
OR = Odds Ratio, CI = Confidence Interval
|
Note the sensible defaults with this basic usage (that can be customized later):
The model was recognized as logistic regression with coefficients exponentiated, so the header displayed “OR” for odds ratio.
Variable types are automatically detected and reference rows are added for categorical variables.
Model estimates and confidence intervals are rounded and formatted.
Because the variables in the data set were labelled, the labels were carried through into the {gtsummary} output table. Had the data not been labelled, the default is to display the variable name.
Variable levels are indented and footnotes added.
There are four primary ways to customize the output of the regression model table.
tbl_regression()
function input argumentsadd_*()
functionsThe tbl_regression()
function includes many arguments for modifying the appearance.
Argument | Description |
---|---|
|
modify variable labels in table |
|
exponentiate model coefficients |
|
names of variables to include in output. Default is all variables |
|
By default, categorical variables are printed on multiple rows. If a variable is dichotomous and you wish to print the regression coefficient on a single row, include the variable name(s) here. |
|
confidence level of confidence interval |
|
indicates whether to include the intercept |
|
function to round and format coefficient estimates |
|
function to round and format p-values |
|
function to specify/customize tidier function |
The {gtsummary} package has built-in functions for adding to results from tbl_regression()
. The following functions add columns and/or information to the regression table.
Function | Description |
---|---|
|
adds the global p-value for a categorical variables |
|
adds statistics from `broom::glance()` as source note |
|
adds column of the variance inflation factors (VIF) |
|
add a column of q values to control for multiple comparisons |
The {gtsummary} package comes with functions specifically made to modify and format summary tables.
Function | Description |
---|---|
|
update column headers |
|
update column footnote |
|
update spanning headers |
|
update table caption/title |
|
bold variable labels |
|
bold variable levels |
|
italicize variable labels |
|
italicize variable levels |
|
bold significant p-values |
The {gt} package is packed with many great functions for modifying table output—too many to list here. Review the package’s website for a full listing.
To use the {gt} package functions with {gtsummary} tables, the regression table must first be converted into a {gt} object. To this end, use the as_gt()
function after modifications have been completed with {gtsummary} functions.
%>%
m1 tbl_regression(exponentiate = TRUE) %>%
as_gt() %>%
::tab_source_note(gt::md("*This data is simulated*")) gt
Characteristic | OR1 | 95% CI1 | p-value |
---|---|---|---|
Age | 1.02 | 1.00, 1.04 | 0.091 |
T Stage | |||
T1 | — | — | |
T2 | 0.58 | 0.24, 1.37 | 0.2 |
T3 | 0.94 | 0.39, 2.28 | 0.9 |
T4 | 0.79 | 0.33, 1.90 | 0.6 |
This data is simulated | |||
1
OR = Odds Ratio, CI = Confidence Interval
|
There are formatting options available, such as adding bold and italics to text. In the example below,
- Coefficients are exponentiated to give odds ratios
- Global p-values for Stage are reported - Large p-values are rounded to two decimal places
- P-values less than 0.10 are bold - Variable labels are bold
- Variable levels are italicized
# format results into data frame with global p-values
%>%
m1 tbl_regression(
exponentiate = TRUE,
pvalue_fun = ~style_pvalue(.x, digits = 2),
%>%
) add_global_p() %>%
bold_p(t = 0.10) %>%
bold_labels() %>%
italicize_levels()
#> add_global_p: Global p-values for variable(s) `add_global_p(include = c("age",
#> "stage"))` were calculated with
#> `car::Anova(x$model_obj, type = "III")`
Characteristic | OR1 | 95% CI1 | p-value |
---|---|---|---|
Age | 1.02 | 1.00, 1.04 | 0.087 |
T Stage | 0.62 | ||
T1 | — | — | |
T2 | 0.58 | 0.24, 1.37 | |
T3 | 0.94 | 0.39, 2.28 | |
T4 | 0.79 | 0.33, 1.90 | |
1
OR = Odds Ratio, CI = Confidence Interval
|
The tbl_uvregression()
function produces a table of univariate regression models. The function is a wrapper for tbl_regression()
, and as a result, accepts nearly identical function arguments. The function’s results can be modified in similar ways to tbl_regression()
.
%>%
trial select(response, age, grade) %>%
tbl_uvregression(
method = glm,
y = response,
method.args = list(family = binomial),
exponentiate = TRUE,
pvalue_fun = ~style_pvalue(.x, digits = 2)
%>%
) add_global_p() %>% # add global p-value
add_nevent() %>% # add number of events of the outcome
add_q() %>% # adjusts global p-values for multiple testing
bold_p() %>% # bold p-values under a given threshold (default 0.05)
bold_p(t = 0.10, q = TRUE) %>% # now bold q-values under the threshold of 0.10
bold_labels()
#> add_global_p: Global p-values for variable(s) `add_global_p(include = c("age",
#> "grade"))` were calculated with
#> `car::Anova(mod = x$model_obj, type = "III")`
#> add_q: Adjusting p-values with
#> `stats::p.adjust(x$table_body$p.value, method = "fdr")`
Characteristic | N | Event N | OR1 | 95% CI1 | p-value | q-value2 |
---|---|---|---|---|---|---|
Age | 183 | 58 | 1.02 | 1.00, 1.04 | 0.091 | 0.18 |
Grade | 193 | 61 | 0.93 | 0.93 | ||
I | — | — | ||||
II | 0.95 | 0.45, 2.00 | ||||
III | 1.10 | 0.52, 2.29 | ||||
1
OR = Odds Ratio, CI = Confidence Interval
2
False discovery rate correction for multiple testing
|
The {gtsummary} regression functions and their related functions have sensible defaults for rounding and formatting results. If you, however, would like to change the defaults there are a few options. The default options can be changed using the {gtsummary} themes function set_gtsummary_theme()
. The package includes pre-specified themes, and you can also create your own. Themes can control baseline behavior, for example, how p-values are rounded, coefficients are rounded, default headers, confidence levels, etc. For details on creating a theme and setting personal defaults, visit the themes vignette.