I have survival data and I'm comparing two groups. The groups are optimal matched prior to analysis, so I need to include the pairings in the KM for the log rank test.
I completed the KM with the code below:
library(survival)
library(survminer)
library(tibble)
mydata<-tribble(
~id, ~ptime, ~pstatus, ~group, ~subclass,
1, 4010, 0, 1, 1,
2, 3, 1, 0, 1,
3, 3672, 0, 1, 2,
4, 980, 1, 0, 2,
5, 3571, 0, 0, 3,
6, 1606, 1, 1, 3,
7, 622, 1, 0, 4,
8, 1169, 1, 1, 4,
9, 1589, 0, 1, 5,
10, 816, 0, 0, 5,
11, 551, 1, 1, 6,
12, 1398, 0, 0, 6,
13, 1578, 0, 1, 7,
14, 431, 0, 0, 7,
15, 1883, 0, 0, 8,
16, 1549, 0, 1, 8
)
object<-Surv(mydata$ptime, mydata$pstatus)
fit_km<-survfit(object ~ group + strata(subclass), data=mydata)
surv_pvalue(fit_km, data=mydata)
The next step is I'd like to get a KM plot. However, when I try it the usual way, I get an error - it seems like it's looking for something for every case of the data?
km_figure<-ggsurvplot(fit_km, data=mydata,
pval=TRUE,
risk.table=TRUE, fontsize=4, tables.height=0.15,
conf.int=TRUE,
ylim=c(0,1),
legend.labs=c("A", "B"), legend.title="Group", legend=c(0.7,0.8),
tables.theme=theme_cleantable(),
ylab="Survival Probability",
break.x.by=365.25,
xscale=c("d_y"),
xlab="Time Since Transplant (years)",
xlim=c(0,2922),
palette = "lancet",
linetype = c("dotted", "solid"),
font.legend=12, pval.coord=c(0.2, 0.2), surv.scale="percent") +
guides(colour=guide_legend(nrow=2))
#Error in ggsurvplot_df(d, fun = fun, color = color, palette = palette, :
#The length of legend.labs should be 16
Is there an argument I need to be adding to the ggsurvplot to let it know that I've paired the data? Or another method I'm missing?
