Survival strata usage may cause some issues

[kieranjmartin]

In this page and a few others the code to use strata is given as follows:

fit.km.str <- survival::survfit(
  survival::Surv(LENFOLY, FSTAT) ~ AFB + survival::strata(GENDER),
  data = dat
)

survminer::surv_pvalue(fit.km.str, data = dat)

Unfortunately, for anyone using survival 3.8.0 or earlier this code will not work! As per the news up until that point doing survival::strata in a formula won't create a strata as you might expect.

I ran into this issue today and spent a few hours trying to work out what on earth was wrong before stumbling across the solution by using a later version of survival. As many pharma users may have a fixed version of survival (It looks like from CRAN that 3.8.3 became available in December of last year, the other versions either weren't released on cran or have been removed from the archive)

[DrLynTaylor]

This is a great catch. Would you have anyone who could update the pages, to specify something like: do option A if <3.8.0 and do B if >=3.8.0 to make the results correct? Possibly cross reference the News file too to point people to this important change.