ddsjoberg

library(gtsummary)
library(broom.helpers)
library(tidyverse)
packageVersion("gtsummary")
#> [1] '1.6.0'

# build model
mod <- biglm::bigglm(response ~ age + trt, data = trial, family = binomial()) 

# build a fancy tidy data frame one step at a time

ddsjoberg

library(gtsummary)
library(tidyverse)

trial %>%
  tbl_strata(
    strata = trt,
    ~ .x %>%
      nest(data = -grade) %>%
      arrange(grade) %>%
      rowwise() %>%

ddsjoberg

gt_rainbow_stripes <- function(x, rep_n = 5) {
  x <-
    gt::cols_width(x, label ~ gt::px(130)) %>%
    gt::tab_header(
      title = x[["_heading"]]$title,
      subtitle = "But Make it G - A - Y"
    ) %>%
    gt::tab_options(heading.subtitle.font.size = 20,
                    heading.subtitle.font.weight = "bolder")

ddsjoberg

set.seed(20210511)
library(gtsummary)
library(magrittr)

multinom_pivot_wider <- function(x) {
  # check inputs match expectatations
  if (!inherits(x, "tbl_regression") || !inherits(x$model_obj, "multinom")) {
    stop("`x=` must be class 'tbl_regression' summary of a `nnet::multinom()` model.")
  }
  

ddsjoberg

# ANALYSIS Rmd FILE

```{r}
# create a gtsummary table
tbl <- trial %>% tbl_summary()

# save it to file
saveRDS(tbl, file = "my_tbl_summary1.Rds")
```

ddsjoberg

split_gtsummary_tbl <- function(x, .split_after) {
  # get row index where splits occur
  df_index <-
    .split_after %>% 
    purrr::map_int(~x$table_body$variable %in% .x %>%
                 which() %>%
                 max()) %>%
    {union(., nrow(x$table_body))} %>%
    sort() %>%

ddsjoberg

use_significance_stars <- function(x) {
  if (!"estimate" %in% names(x$table_body)) return(x)
  # extracting old estimate fun
  old_est_fun <- switch(
    !is.null(x$table_header),
    x$table_header %>%
      dplyr::filter(column == "estimate") %>%
      purrr::pluck("fmt_fun", 1)
  )
  

ddsjoberg

library(gtsummary)

trial %>%
  select(marker, trt) %>%
  tbl_summary(
    by = trt,
    missing = "no",
    statistic = everything() ~ "{mean} ({sd})",
    # use a function to style output, rather than specify number of decimal places
    digits = everything() ~ style_sigfig

ddsjoberg

# Original stack overflow post
# https://stackoverflow.com/questions/65673290/merging-tbl-svysummary-and-stacked-tbl-regression-tables-with-different-variable

library(gtsummary)
packageVersion("gtsummary")
#> '1.3.6'

# 1. build reg models for varying outcomes, 
# 2. show covariate for age in table for all outcomes, 
# 3. adjust all models for marker level

ddsjoberg

#' @import dplyr %>%
#' @import purrr %||%
tbl_cmh <- function(data, case, exposure, strata, 
                    label = NULL, 
                    estimate_fun = gtsummary::style_ratio, 
                    overall_or = TRUE,
                    overall_label = "Crude") {
  # converting selectors to character names ------------------------------------
  case <- dplyr::select(data, {{ case }}) %>% names()
  exposure <- dplyr::select(data, {{ exposure }}) %>% names()