Creating a KM curve function

Question

I am trying to make a KM curve function, however, I am not sure where I am going wrong.

Edit: I have changed the code to see if it will work with the publically available ovarian data set https://stat.ethz.ch/R-manual/R-devel/library/survival/html/ovarian.html

library(tidyverse)
library(ggplot)
library(survival)
library(survminer)



km_curve <- function(df, tm, ev, predictor) {
            surv_object <- Surv(time = df$tm, event = df$ev)
            fit1 <- survfit(surv_object ~ predictor, data = df)
            ggsurvplot(fit1, data = df, pval = TRUE, xlab = "Months")
            }

The above code works perfectly well If I replace all df, tm, ev, and predictor with the actual names.

I also tried this

km_curve <- function(df, tm, ev, predictor) {
            surv_object <- Surv(time = df[,tm], event = df[,ev])
            fit1 <- survfit(surv_object ~ predictor, data = df)
            ggsurvplot(fit1, data = df, pval = TRUE, xlab = "Months")
            }

I have also tried

   km_curve <- function(df, tm, ev, predictor) {
               tim <- df %>% 
                      select(tm)
               even <- df %>% 
                      select(ev)
               surv_object <- Surv(time = tim, event = even)
               fit1 <- survfit(surv_object ~ predictor, data = df)
               ggsurvplot(fit1, data = df, pval = TRUE, xlab = "Months")
             }

km_curve(df = ovarian, 
         tm = futime, 
         ev = fustat, 
         predictor = rx)

It keeps giving me error messages that pf_time_months not found. Even though it is a column in my data frame I specify.

This does not work either

library(tidyverse)
library(survival)
library(survminer)
ovarian
km_graph<-function(t, e, var, df, xaxis = "Months"){
  surv_object <- Surv(time = t, event =e)
  fit1<- survfit(surv_object ~ var, data = df)

  kmgraph<-ggsurvplot(fit1, data = df, pval = TRUE, xaxis = "Months", risk.table = T)
  kmgraph$table <- ggrisktable(fit1,
                               data = df,
                               color = "strata",
                               y.text = T,   ylab = "",  xlab = "",
                               tables.theme = theme_cleantable(),
      )
       return(kmgraph)
    }

    km_graph(t = ovarian$futime, e =ovarian$fustat, var = rx, df = ovarian)

Error in as.data.frame.default(x[[i]], optional = TRUE) : cannot coerce class ‘"function"’ to a data.frame

Answer 1

My guess (I have to guess since you don't provide a reproducible code) is that pfs_time_months and others are the names of some columns of df. That's why that does not work. You can pass them as character strings if you use this function:

km_curve <- function(df, tm, ev, predictor) {
  surv_object <- Surv(time = df[, tm], event = df[, ev])
  f <- as.formula(sprintf("surv_object ~ %s", predictor))
  fit1 <- survfit(f, data = df)
  ggsurvplot(fit1, data = df, pval = TRUE, xlab = "Months")
}

km_curve(df = first_primaries, 
         tm = "pfs_time_months", 
         ev = "pfs_fail.x", 
         predictor = "risk"
)