Round the weighted survival estimates in the risk table

[josephuses]

I was trying to reproduce the plot in p. 311 of Berglund, Heeringa, and West's Applied Survey Data Analysis. The data is available online at their website, which I linked here. How can I round off the survival estimates in the risk table? Can we have something like risk.table.decimal.place which defaults to a whole number?

library(tidyverse)
library(survival)
library(survminer)
ncsr <- read_sas("~/ncsr_analysis_ex_c1_c10_2011.sas7bdat")
colnames(ncsr) <- tolower(names(ncsr))
fit <- survfit(Surv(ageonsetmde,mde)~racecat, data = ncsr, weights = ncsrwtsh)
ggsurvplot(fit, data = ncsr, 
           ggtheme = theme_bw(), 
           linetype = "strata", 
           risk.table = TRUE, 
           risk.table.col = "strata", 
           legend.labs = c("Other", "Hispanic", "Black", "White"), 
           risk.table.fontsize = 2.75, 
           risk.table.height = 0.275, 
           break.time.by = 20, risk.table.y.text.col = T, 
           risk.table.y.text = FALSE) 

[kassambara]

Would you please give the direct link to the "ncsr_analysis_ex_c1_c10_2011.sas7bdat" file, so that we can reproduce your example...

Thanks!

[kassambara]

Ok, I can found the data in a different post: #215

[kassambara]

Thank you or pointing this out.

Now, the default behaviour of ggsurvplot() is to round the number at risk using the option digits = 0

[josephuses]

Thank you! This solves my issue for this particular case. However, note that setting decimal.place to any other whole number does not give the desired number of decimal places.

[kassambara]

This option is not externaly available. The default is to systematically round the number at risk.