I'm using ggsurvfit to create a Kaplan–Meier estimator for survival probability. I know that I can add quantiles for different time points by using e.g. add_quantile(x_value = 5). However, I would like to show the survival probability for those quantiles also as a number on the y-axis in percent. Is there any possibility to do this? I have looked up into the ggsurvfit examples and also on stack overflow but have not found anything.
survfit2(Surv(time, status) ~ surg, data = df_colon) %>%
ggsurvfit(linewidth = 0.8) +
add_censor_mark(size = 2, alpha = 0.2) +
add_quantile(x_value = 5, color = "grey30", linewidth = 0.8) +
scale_ggsurvfit()
The data that you're looking for is in the 3rd layer, so we can extract that and add it to the breaks of a scale_*_continous()
library(ggsurvfit)
s <- survfit2(Surv(time, status) ~ surg, data = df_colon) %>%
ggsurvfit(linewidth = 0.8) +
add_censor_mark(size = 2, alpha = 0.2) +
add_quantile(x_value = 2.718282,
color = "grey30",
linewidth = 0.8)
s +
scale_y_continuous(
limits = c(0, 1),
labels = scales::label_percent(),
breaks = c(s[["layers"]][[3]][["data"]][["y"]], seq(0, 1, 0.2))
) +
scale_x_continuous(
labels = scales::label_number(accuracy = 1),
breaks = c(
s[["layers"]][[3]][["data"]][["x"]][[1]],
seq(floor(min( df_colon$time)),
ceiling(max(df_colon$time )), 2)))