Add a custom statistic column

Adds one or more columns of statistics computed by functions you supply, one call per variable, to a table created with tbl_summary(). Each function returns a single value, or a vector or data frame with one element or row per row the statistic belongs on (see location). A single value goes into a column named add_stat_1, add_stat_2, and so on; the columns of a data frame keep their names, and a numeric column named p.value or q.value is formatted as a p-value.

Usage

add_stat(x, fns, location = everything() ~ "label")

Arguments

x

(tbl_summary)
A table created with tbl_summary().

fns

(formula-list)
The function to run for each variable, e.g. list(all_continuous() ~ my_stat). Each function is called as fn(data = , variable = , by = , tbl = ): the data frame of the table, the variable name, the by variable name (or NULL) and the table itself. A function that accepts ... receives all of them; otherwise only the arguments it declares are passed.

location

(formula-list)
The rows the statistic is placed on, per variable: "label" (the default), "level" or "missing". With "level", the function returns one value per level.

Value

A table of class c("tbl_summary", "ltsummary") with the new columns shown; numeric columns are formatted with label_style_sigfig(digits = 3).

See also

modify_header() to name the new column, modify_fmt_fun() to change its formatting.

Other add statistics: add_ci(), add_difference(), add_difference_row(), add_n(), add_overall(), add_p(), add_p_continuous, add_q(), add_stat_label(), tests

Examples

# Example 1 ----------------------------------
# a p-value from a Kruskal-Wallis test, one row per variable
my_kruskal <- function(data, variable, by, ...) {
  kruskal.test(data[[variable]] ~ as.factor(data[[by]]))$p.value
}
trial |>
  tbl_summary(by = trt, include = c(age, marker), missing = "no") |>
  add_stat(fns = everything() ~ my_kruskal) |>
  modify_header(add_stat_1 = "**p-value**")

# Example 2 ----------------------------------
# a data frame of statistics, placed on the level rows
level_counts <- function(data, variable, by, ...) {
  counts <- table(data[[variable]])
  data.frame(n_level = as.vector(counts), p_level = as.vector(prop.table(counts)))
}
trial |>
  tbl_summary(by = trt, include = grade, missing = "no") |>
  add_stat(fns = grade ~ level_counts, location = grade ~ "level") |>
  modify_header(n_level = "**N**", p_level = "**Fraction**")