Creating a tbl_continuous with weight data using gtsummary

user_14708587 0
Viewed 125

Problem:

Can't find a way to create a tbl_continuous of a weighted numerical variable. I'm using tbl_svysummary to create my categorical variable tables but it's not useful when i try to do the same with numerical variable tables. Note: i'm not looking for a general mean of my numeric variable but separated by or groupped by a categorical variable.

Attempt:

For example, i've created this table with the help of the function tbl_continuous which does exactly what i want: the mean of my numeric variable but by the levels of my categorical variable. The only problem is that i can't insert a weight variable into this function.

```{r}
base2 %>% 
  as_label() %>% 
  select(ing_cap, ano, nacional, dominio) %>% 
  tbl_continuous(variable = ing_cap, 
                 by = ano,
                 statistic = list(everything() ~ "{median}"))

```

table

Also, i have been creating weighted data with the srvyr package in the following way:

base2 %>% 
  labelled::drop_unused_value_labels() %>% 
  as_label() %>% 
  as_survey_design(weight = fac500a)

Could add up to the solution.

Request:

Create this same table (shown in the image) but with a weight variable. My weight variable in my data is called fac500a.

My data:

My data can be dowloaded from my github repo and has the following dimensions:

> dim(base2)
[1] 108103     44

https://github.com/aito123/quarto_blog/raw/master/posts/tablas_tesis/base2.sav

(dput output is long)

My current packages:

I'm using this r packages so far: tidyverse, srvyr, gtsummary, sjlabelled, haven

Conclusion:

Let me know if it's neccesary to provide more information.

gtsummary r
1 Answers

The gtsummary package does not export an analogous function of tbl_continuous() for survey data. But you can construct the table. Example below!

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

svy_trial <- survey::svydesign(ids = ~1, data = trial, weights = ~1)

tbl <- 
  svy_trial %>%
  tbl_strata2(
    strata = grade,
    ~ .x %>%
      tbl_svysummary(
        by = trt, 
        include = age, 
        missing = "no",
        label = list(age = .y)
      ) %>%
      modify_header(all_stat_cols() ~ "**{level}**"),
    .combine_with = "tbl_stack",
    .combine_args = list(group_header = NULL)
  ) %>%
  modify_table_body(
    ~ .x %>%
      mutate(variable = "grade", row_type = "level") %>%
      tibble::add_row(
        row_type = "label",
        variable = "grade",
        label = "Grade",
        .before = 1L
      ) 
  ) %>%
  modify_column_indent(columns = label, rows = row_type == "level") %>%
  bold_labels() %>%
  modify_spanning_header(all_stat_cols() ~ "**Treatment**") %>%
  modify_footnote(all_stat_cols() ~ "Age: Median (IQR)")

enter image description here Created on 2022-05-16 by the reprex package (v2.0.1)

user_87112760
Related