Search code examples
gtsummary

Creating stratified rows with gtsummary

I want to create a table stratified three ways: I'd like the columns to represent one variable (location), and the rows to display other variables (age group and sex), stratified by a second variable (gcs). I also want to include the missing values in the table.

Example dataframe here:

dat <- data.frame(id = 1:20,
                  age_group = as.factor(c("66-70", "71-75", "61-65", "76-80", "61-65", "81-85", "76-80","66-70","81-85","66-70",
                                "66-70", "81-85","81-85","76-80","61-65","76-80","66-70","81-85","66-70","71-75" )),
                  location = as.factor(c("indoor", NA, "indoor", "indoor", "outdoor","indoor", "outdoor", "indoor", "indoor", "outdoor",
                               "indoor", "outdoor","indoor", NA,"outdoor", "outdoor","indoor", "outdoor","indoor", "outdoor")),
                  sex = as.factor(c("M", "F", "M", "F", NA, "M", "F", "M", "F", "M","F", "M", "F", "M", "F", "F", "M", "F", "F", "M")),
                  gcs = as.factor(c("moderate", "severe", "minor", "minor", "minor", "severe", "minor", "moderate", NA, "minor",
                          "moderate", "minor", NA, "minor", "severe", "moderate", "minor", "severe", "minor", "severe")))

The closest I've gotten is here:

table <- tbl_strata(data = dat,
                    strata = location,
                    .tbl_fun = 
                      ~.x %>%
                      tbl_summary( include = ! id,
                                   by = gcs, missing = "always"))

But I'd like the rows to be stratified by gcs rather than the columns like this

Thanks!

Solution

  • Is this what you're after?

    library(gtsummary)

dat <- data.frame(id = 1:20,
                  age_group = as.factor(c("66-70", "71-75", "61-65", "76-80", "61-65", "81-85", "76-80","66-70","81-85","66-70",
                                          "66-70", "81-85","81-85","76-80","61-65","76-80","66-70","81-85","66-70","71-75" )),
                  location = as.factor(c("indoor", NA, "indoor", "indoor", "outdoor","indoor", "outdoor", "indoor", "indoor", "outdoor",
                                         "indoor", "outdoor","indoor", NA,"outdoor", "outdoor","indoor", "outdoor","indoor", "outdoor")),
                  sex = as.factor(c("M", "F", "M", "F", NA, "M", "F", "M", "F", "M","F", "M", "F", "M", "F", "F", "M", "F", "F", "M")),
                  gcs = as.factor(c("moderate", "severe", "minor", "minor", "minor", "severe", "minor", "moderate", NA, "minor",
                                    "moderate", "minor", NA, "minor", "severe", "moderate", "minor", "severe", "minor", "severe")))

table <- 
  tbl_strata(
    data = dat,
    strata = age_group,
    .tbl_fun = 
      ~.x %>%
      tbl_summary(
        include = gcs,
        by = location, 
        missing = "always"
      ),
    .combine_with = "tbl_stack"
  )
#> 1 observations missing `location` have been removed. To include these observations, use `forcats::fct_na_value_to_level()` on `location` column before passing to `tbl_summary()`.
#> 1 observations missing `location` have been removed. To include these observations, use `forcats::fct_na_value_to_level()` on `location` column before passing to `tbl_summary()`.
#> ℹ Column headers among stacked tables differ. Headers from the first table are
#> used. Use `quiet = TRUE` to supress this message.

    enter image description here Created on 2023-08-24 with reprex v2.0.2