stratified cox model with gtSUMMARY

Question

I am trying to create multiple univariate cox regression analyses using GTsummary. However, I need to add a stratification variable into the model and not sure how to go about this. Could any of you kindly help?

I would need to stratify my current working code by a variable "job_site"

#CURRENT CODE: 
uvlm_table <- tbl_uvregression(
  tbl1 %>%
    select(right_censor, n_result, age, exp_case, job_cat, n_facilities)
  type = list(
    age ~ "continuous"
  ),
  method = coxph,
  y = Surv(right_censor, n_result),
  exponentiate = TRUE,
  pvalue_fun = ~style_pvalue(.x, digits=3)
)

uvlm_table    

I tried to write the formula out manually but it did not work.

Answer 1

If you are looking to stratify by job site you can modify the formula input to incorporate this. I added job site as a strata below.

library(survival)
library(gtsummary)

uvlm_table <- tbl_uvregression(
  tbl1 %>%
    select(right_censor, n_result, age, exp_case, job_cat, n_facilities, job_site),
  type = list(
    age ~ "continuous"
  ),
  method = coxph,
  y = Surv(right_censor, n_result),
  formula = "{y} ~ {x} + strata(job_site)",
  exponentiate = TRUE,
  pvalue_fun = ~style_pvalue(.x, digits=3)
)

Here is a reproducible example with the trial dataset from gtsummary.

uvlm_table <- tbl_uvregression(
  trial %>%
    select(ttdeath, death, age,marker, grade),
  type = list(
    age ~ "continuous"
  ),
  method = coxph,
  y = Surv(ttdeath, death),
  formula = "{y} ~ {x} + strata(grade)",
  exponentiate = TRUE,
  pvalue_fun = ~style_pvalue(.x, digits=3)
)