getKMcurve: Extract data describing the Kaplan-Meier curve from a...
In RichardBirnie/survtools: A collection of useful functions for survival analysis

Description Usage Arguments Value Examples

View source: R/utilities.R

Extract data describing the Kaplan-Meier curve from a survfit object

1	getKMcurve(km, time.col, event.col, group.col, data)

`km`	A survfit object containing the Kaplan-Meier estimate of the survival function as returned by `survfit`

A data frame with columns:

Time: The time of the events
Survival: The probability of survival at each event time
stderr: The standard error of the survival estimate
lower: The lower limit of the confidence interval for the survival curve
upper: The upper limit of the confidence interval for the survival curve
Nrisk: The number of patients at risk at each event time
group: (Optional) The group that each observation belongs to if there are multiple patient groups. This will be absent if there is only one group

library(survival)
library(survminer)
library(flexsurv)
library(dplyr)
library(survtools)

#load some example data from the flexsurv package
#Survival times of 686 patients with primary node positive breast cancer
#censrec = censoring, 1 = dead, 0 = censored
#rectime = time of death or censoring in days
#recyrs = time of death or censoring in years
#group = prognostic group
data(bc)

#create a survival object
surv.obj <- Surv(recyrs, censrec) ~ group

#simple Kaplan-Meier estimate
surv.km <- survfit(surv.obj, data = bc)

#Get the Kaplan-Meier data
#.col variables should correspond to column names in the data set
km.curve <- getKMcurve(km = surv.km, time.col = 'recyrs', event.col = 'censrec', group.col = 'group', data = bc)

#plot KM
plot(km.curve$Time[km.curve$group == 'Good'], km.curve$Survival [km.curve$group == 'Good'], type="l", ylim=c(0,1),
     xlab = 'Time', ylab = 'Survival', col = 'red')
lines(km.curve$Time[km.curve$group == 'Medium'], km.curve$Survival [km.curve$group == 'Medium'], col='green')
lines(km.curve$Time[km.curve$group == 'Poor'], km.curve$Survival [km.curve$group == 'Poor'], col='blue')