`tbl_summary` should pass to custom `add_stat` functiond the variable type as gtsummary treats that variable, not the real one

[raffaelemancuso]

Consider the following example:

N <- 1000
toydf <- data.frame(
  children = rbinom(N, size = 3, prob = 0.5),
  age = extraDistr::rdunif(N, 10, 100),
  gender = sample(c("M", "F"), N, replace = TRUE)
)
toydf$age %<>% as.integer()
toydf$gender %<>% as.factor()

my_mean_diff <- function(data, variable, by, tbl, ...) {
  x <- data[[variable]]
  g <- data[[by]]
  lvls <- levels(g)
  vartype <- class(x)[1]
  
  # <UNCOMMENT THIS TO MAKE IT WORK>
  #if(
  #  (vartype=="character" | vartype=="factor" | vartype=="numeric" | vartype=="integer") &
  #  length(unique(x)) <= 10
  #) 
  #{
  #  vartype <- "categorical"
  #}
  
  print(paste0("Variable: ", variable, ", Type: ", vartype))
  
  switch(
    vartype,
    categorical = {
      prop <- prop.table(table(x, g), margin = 2)
      return((prop[, lvls[2]] - prop[, lvls[1]]) * 100)
    },
    integer = {
      return(diff(tapply(x, g, mean, na.rm = TRUE)))
    },
    {
      stop(glue("ERROR: Unrecognized type {vartype}"))
    }
  )
}

toydf %>% 
  tbl_summary(
    statistic = list(
      all_continuous() ~ "{mean} ({sd})",
      all_categorical() ~ "{n} ({p}%)"
    ),
    missing = "no",
    by = "gender",
    percent = "column"
  ) %>% add_stat(
  fns = everything() ~ my_mean_diff,
  location = list(
    all_continuous() ~ "label",
    all_categorical() ~ "level",
    all_dichotomous() ~ "label"
  )
) 

It will return:

Error in `add_stat()`:
! Dimension of "children" and the added statistic do not match.
ℹ Expecting statistic/data frame to be length/no. rows 4.
Run `rlang::last_trace()` to see where the error occurred.

This is because, having children fewer than 10 levels, gtsummary is treating it as a factor variable.

But at this point we should pass the variable to custom add_stat function as factor, not as numeric.

[raffaelemancuso]

It's a real mess because the variable type can be modified by the type argument, but no matter what gtsummary pass the original type.

At least, pass a parameter containing the new type.