How do you calculate a success rate for a column in tbl_summary using a different column value?

Question

My question can be conceptualized using the same sample data used for the tbl_summary trial dataset from the tutorial here: https://www.danieldsjoberg.com/gtsummary/articles/tbl_summary.html.

How would I calculate the mortality rate based on grade of tumor instead of calculating the incidence of grade I-III tumors? So far as I can find, there is not a way to calculate a custom statistic based on a third column, which is essentially what I am trying to do.

--- Edit ---

My input data has 3 columns - stent_type, success, stent_30. stent_type is either a category 0 or 1, success is either a 0 or 1, and stent_30 is either a 0 or 1 (basically was the stent placed before or after 30 days). I would like to create a 2x3 table with the percent/number of successfully placed stents for category 0 or 1 stents, along with a p-value comparing the two columns for stent_30. Something like this:

	Stent < 30	Stent > 30	p-value
stent 0	75%(10)	70% (9)	0.1
stent 1	50% (6)	40% (2)	0.05

I'm basically having trouble because I'm not sure how to summarize a category based on 2 different categories (if that makes sense?)

Jay Bee · Accepted Answer · 2024-02-27 03:59:30Z

0

You can do this with general tidyverse operations (group_by and summarise).

# Load the required libraries.

library(gtsummary)
library(tidyverse)

# Get the mortality counts.

mortality_counts <- trial %>%
  group_by(grade, death) %>%
  summarise(count = n()) %>%
  ungroup()

# Get the mortality proportions.

mortality_proportions <- mortality_counts %>%
  group_by(grade) %>%
  mutate(total = sum(count),
         proportion = count / total)

With the output of mortality_proportions as:

grade	death	count	total	proportion
I	0	35	68	0.5147059
I	1	33	68	0.4852941
II	0	32	68	0.4705882
II	1	36	68	0.5294118
III	0	21	64	0.3281250
III	1	43	64	0.6718750

If you wanted to make that clearer just to show the grade and proportions, you could do:

mortality_proportions %>%
  filter(death == 1) %>%
  select(grade, proportion) %>%
  mutate(proportion = round(proportion, 2))

For output:

grade	proportion
I	0.49
II	0.53
III	0.67

edited Feb 27, 2024 at 3:59

answered Feb 27, 2024 at 3:50

Jay Bee

5706 silver badges22 bronze badges

Sign up to request clarification or add additional context in comments.

3 Comments

jon Over a year ago

Sorry, I guess I should have made it more clear in the initial question. This is very helpful but the comparison that I would like to make is to then compare death rates across drug A and drug B. With this approach it doesn't seem like that would be possible right?

Jay Bee Over a year ago

Can you edit your question to include your desired output? So a table showing what the end result should look like, even if it has fake numbers.

Jay Bee Over a year ago

Okay, I think you should ask this as a new question. The original question was about data from gtsummary but the edit is about a different dataset and now includes p-values. I feel like it's definitely possible to achieve this, but better as a new question after reading these links: (1) stackoverflow.com/help/minimal-reproducible-example and (2) stackoverflow.com/questions/5963269/…

Collectives™ on Stack Overflow

How do you calculate a success rate for a column in tbl_summary using a different column value?

1 Answer 1

3 Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

3 Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related