Error using {N_miss} and the like for continuous variables

[haziqj]

Hi, it used to be possible to request statistics on the number of missing and non-missing observations and their proportions for continuous variables. This now returns an error. Is this a bug or the feature has been removed?

It works fine for categorical variables.

Thanks

library(gtsummary)
trial |>
  select(age) |>
  tbl_summary(
    statistic = all_continuous() ~ "{N_miss}",
    missing = "no"
  )
#> Error in `x[c("variable", "stat_name")]`:
#> ! Can't subset columns that don't exist.
#> ✖ Columns `variable` and `stat_name` don't exist.

^{Created on 2024-12-06 with reprex v2.1.1}

[ddsjoberg]

You can still use them for continuous variables, but the function is expecting at least one continuous summary statistic, like the mean in the example below.

library(gtsummary)

trial |>
  select(age) |>
  tbl_summary(
    statistic = all_continuous() ~ "{mean}; N missing={N_miss}",
    missing = "no"
  ) |> 
  as_kable()

Characteristic	N = 200
Age	47; N missing=11

^{Created on 2024-12-05 with reprex v2.1.1}

I never expected someone to only summarize the number of non-missing variables, but I guess it's something we should support again.

[haziqj]

My use case is for a missing data analysis, where I would like to show the severity of missing data for each variable by a certain categorical variable. tbl_summary() was a neat way of doing this, especially with add_overall().

In the mean time, I've done a hacky thing to get what I want:

library(gtsummary)
blank_fn <- function(x) NA
tab <- trial |>
  select(age) |>
  tbl_summary(
    statistic = all_continuous() ~ "{blank_fn} {N_miss} ({p_miss}%)",
    missing = "no"
  )
tab$table_body <-
  tab$table_body |>
  mutate(across(everything(), \(x) gsub("NA ", "", x)))
as_kable(tab)

Characteristic	N = 200
Age	11 (5.5%)

^{Created on 2024-12-06 with reprex v2.1.1}

Thanks

[ddsjoberg]

FWIW, this is how I imaged this type of operation would be run:

library(gtsummary)

trial |>
  select(age) |>
  dplyr::mutate(dplyr::across(everything(), is.na)) |> 
  tbl_summary() |> 
  as_kable() # convert to kable to display on GH

Characteristic	N = 200
age	11 (5.5%)

^{Created on 2024-12-17 with reprex v2.1.1}