How can I make odds ratio and 95% CI for chi-square test in gtsummary with my own function?

You can add the odds ratio via a custom function passed to add_difference(). Instructions on how to construct a custom function are here https://www.danieldsjoberg.com/gtsummary/reference/tests.html#custom-functions-1.

I've written an example below. But please take careful note of how fisher.test() constructs the odds ratio from the 2x2 table. You may need to switch the order of the variables in table() or even reverse the order of your variable levels to get the odds ratio you need.

library(gtsummary)
packageVersion("gtsummary")
#> [1] '1.6.0'

# function to calculate OR from 2x2 table
or_fun <- function(data, variable, by, ...) {
  table(data[[by]], data[[variable]]) %>%
    fisher.test() %>%
    broom::tidy()
}
# testing function
or_fun(trial, "trt", "response")
#> # A tibble: 1 × 6
#>   estimate p.value conf.low conf.high method                         alternative
#>      <dbl>   <dbl>    <dbl>     <dbl> <chr>                          <chr>      
#> 1     1.21   0.540    0.633      2.34 Fisher's Exact Test for Count… two.sided

# add OR to a `tbl_summary()`
tbl <-
  trial %>%
  select(response, trt) %>%
  tbl_summary(by = trt, missing = "no", statistic = all_categorical() ~ "{n} / {N} ({p}%)") %>%
  add_difference(test = response ~ or_fun, estimate_fun = response ~ style_ratio) %>%
  modify_header(estimate ~ "**Odds Ratio**")

^{Created on 2022-05-08 by the reprex package (v2.0.1)}