CODEWITHSUNDEEP
rsurvival-analysis
Shan Cheung
Shan Cheung

Reputation: 75

Plot two curves only (Not all) in survfit output in R

I have a set of survival data with a covariate (SubType) in 5 groups. I managed to plot the 5 groups using:

km.type <- survfit(SurvObj ~ SubType, data = data, conf.type = "log-log")
plot(km.type, mark.time=FALSE)

I would like to display only 2 groups instead of all 5 groups. Is there any way to do this?

Upvotes: 0

Views: 1439

Answers (1)

rawr
rawr

Reputation: 20811

You can subset the data before, during, or after fitting the survival model

library('survival')

summary(aml)
# time            status                   x     
# Min.   :  5.00   Min.   :0.0000   Maintained   :11  
# 1st Qu.: 12.50   1st Qu.:1.0000   Nonmaintained:12  
# Median : 23.00   Median :1.0000                     
# Mean   : 29.48   Mean   :0.7826                     
# 3rd Qu.: 33.50   3rd Qu.:1.0000                     
# Max.   :161.00   Max.   :1.0000                    

## before
dd <- aml[aml$x == 'Maintained', ]
fit <- survfit(Surv(time, status) ~ x, data = dd)
plot(fit, conf.int = FALSE)

## during
fit <- survfit(Surv(time, status) ~ x, data = aml, subset = x == 'Maintained')
## or
fit <- survfit(Surv(time, status) ~ x, data = aml[aml$x == 'Maintained', ])
plot(fit, conf.int = FALSE)

## after
fit <- survfit(Surv(time, status) ~ x, data = aml)
plot(fit, col = c('black','transparent'), conf.int = FALSE)

where col is a vector of colors for each of the groups in order.

All are identical:

enter image description here

Upvotes: 2

Related Questions