How can I view the 95% CIs calculated from a ggcompetingrisks model that uses a cuminc object?-CodePudding

I created a cuminc object and can print the estimates and variances directly in my console. I then used ggcompetingrisks to create a plot, and used the option conf.int = TRUE to plot the confidence intervals. However, I can't find the values of the confidence intervals.

CodePudding user response：

The confidence bounds are not saved in the output. Fortunately it's easy to calculate them manually. Just extract est and std from plot$data and calculate est 1.96*std and est-1.96*std, replacing 1.96 with other z-values for other confidence limits.

library(survminer)
library(cmprsk)
set.seed(2)
ss <- rexp(100)
gg <- factor(sample(1:3,100,replace=TRUE),1:3,c('BRCA','LUNG','OV'))
cc <- factor(sample(0:2,100,replace=TRUE),0:2,c('no event', 'death', 'progression'))
strt <- sample(1:2,100,replace=TRUE)
fit <- cmprsk::cuminc(ss,cc,gg,strt)
plot <- ggcompetingrisks(fit, conf.int = TRUE)
plot

# Calculate confidence bounds manually
dat <- plot$data
dat$ymax <- dat$est   1.96*dat$std
dat$ymin <- dat$est - 1.96*dat$std
head(dat)
#>        time        est         var          name event group        std
#> 1 0.0000000 0.00000000 0.000000000 BRCA no event    no  BRCA 0.00000000
#> 2 0.2731330 0.00000000 0.000000000 BRCA no event    no  BRCA 0.00000000
#> 3 0.2731330 0.03448276 0.001202639 BRCA no event    no  BRCA 0.03467908
#> 4 0.3312168 0.03448276 0.001202639 BRCA no event    no  BRCA 0.03467908
#> 5 0.3312168 0.06896552 0.002321886 BRCA no event    no  BRCA 0.04818595
#> 6 0.7388502 0.06896552 0.002321886 BRCA no event    no  BRCA 0.04818595
#>        ymax        ymin
#> 1 0.0000000  0.00000000
#> 2 0.0000000  0.00000000
#> 3 0.1024538 -0.03348824
#> 4 0.1024538 -0.03348824
#> 5 0.1634100 -0.02547895
#> 6 0.1634100 -0.02547895

# Plot to confirm they match
ggcompetingrisks(fit, conf.int = FALSE)  
  geom_ribbon(data = dat, 
              mapping = aes(ymin = ymin, ymax = ymax, x = time,
                            fill = event),
              alpha = 0.2)

^{Created on 2022-09-13 with reprex v2.0.2}