我们在既往的文章《R语言绘制复杂抽样设计数据cox回归生存曲线(Kaplan-Meier)》中介绍了怎么使用jskm包的svykm函数绘制复杂抽样设计数据cox回归生存曲线(Kaplan-Meier)，但是有粉丝觉得讲得不够详细，希望讲得详细一点，今天我们继续来介绍一下。
先导入我们的R包和数据

library(jskm)
library(survey)
library(jskm)
pbc<-read.csv("E:/r/test/pbc.csv",sep=',',header=TRUE) 

这是一个原发性胆道胆管炎数据，公众号回复：胆管炎数据，可以获得数据，
数据我们解释几个等下要用到的变量，age：年龄，trt：治疗方案：1D-青霉烯，2安慰剂，edema：水肿， status: 结局变量0/1/2表示审查、移植、死亡。
咱们先来一波小操作，生成一个预测值，等下好操作，不喜欢可以跳过这部分，对后面的操作没影响.

pbc$randomized <- with(pbc, !is.na(trt) & trt>0)
biasmodel <- glm(randomized~age*edema,data=pbc)
pbc$randprob <- fitted(biasmodel)

生成预测值randprob后我们就可以正式分析了，我们先生成一个调查数据

dpbc <- svydesign(id=~1, prob=~randprob, strata=~edema, data=subset(pbc,randomized))

直接使用svykm函数生成预测值,这里也要生存可信区间

s1 <- svykm(Surv(time,status>0)~sex, design=dpbc,se=T)

现在我们直接使用svyjskm函数来进行绘图

svyjskm(s1)

更改X轴，Y轴标签和标题

svyjskm(s1,xlabs = "时间",ylabs = "生存率",main = "时间与生存率关系图")

对图例进行修改

svyjskm(s1,xlabs = "时间",ylabs = "生存率",ystrataname = "性别",ystratalabs=c("男","女"))

对X轴的范围进行限制，对Y轴也是一样的

svyjskm(s1,xlabs = "时间",ylabs = "生存率",ystrataname = "性别",
        ystratalabs=c("男","女"),xlims=c(0,3000))

把Y轴改成以百分比显示

svyjskm(s1,xlabs = "时间",ylabs = "生存率",ystrataname = "性别",
        ystratalabs=c("男","女"),surv.scale="percent")

更改显示的时间轴点数，默认是7个点

svyjskm(s1,xlabs = "时间",ylabs = "生存率",ystrataname = "性别",
        ystratalabs=c("男","女"),timeby=700)

在图中生成P值

svyjskm(s1,xlabs = "时间",ylabs = "生存率",ystrataname = "性别",
        ystratalabs=c("男","女"),timeby=700,pval=T)

调整文中P值的位置和字体的大小

svyjskm(s1,xlabs = "时间",ylabs = "生存率",ystrataname = "性别",
        ystratalabs=c("男","女"),timeby=700,pval=T,pval.coord=c(1000,0.4),pval.size=8)

添加Log-rank进P值后面

svyjskm(s1,xlabs = "时间",ylabs = "生存率",ystrataname = "性别",
        ystratalabs=c("男","女"),timeby=700,pval=T,
        pval.coord=c(1000,0.4),pval.testname=T)

调整图例位置

svyjskm(s1,xlabs = "时间",ylabs = "生存率",ystrataname = "性别",
        ystratalabs=c("男","女"),timeby=700,pval=T,
        pval.coord=c(1000,0.4),pval.testname=T,legendposition=c(0.6,0.8))

不显示图例

svyjskm(s1,xlabs = "时间",ylabs = "生存率",ystrataname = "性别",
        ystratalabs=c("男","女"),timeby=700,pval=T,
        pval.coord=c(1000,0.4),pval.testname=T,legendposition=c(0.6,0.8),legend=F)

更换色彩风格，默认是Set1

svyjskm(s1,xlabs = "时间",ylabs = "生存率",ystrataname = "性别",
        ystratalabs=c("男","女"),timeby=700,pval=T,
        pval.coord=c(1000,0.4),pval.testname=T,linecols="Set2")

dashed设置虚线，好像只能设置一条

svyjskm(s1,xlabs = "时间",ylabs = "生存率",ystrataname = "性别",
        ystratalabs=c("男","女"),timeby=700,pval=T,
        pval.coord=c(1000,0.4),pval.testname=T,linecols="Set2",dashed=T)

显示累积发生率函数

svyjskm(s1,xlabs = "时间",ylabs = "生存率",ystrataname = "性别",
        ystratalabs=c("男","女"),timeby=700,pval=T,
        pval.coord=c(1000,0.4),pval.testname=T,linecols="Set2",cumhaz=T)

调用ggpubr::ggarrange函数来形成表格，给了我不少启发

svyjskm(s1,xlabs = "时间",ylabs = "生存率",ystrataname = "性别",
        ystratalabs=c("男","女"),timeby=700,pval=T,
        pval.coord=c(1000,0.4),pval.testname=T,linecols="Set2",table=T)

修改表格下面的副标题名字

svyjskm(s1,xlabs = "时间",ylabs = "生存率",ystrataname = "性别",
        ystratalabs=c("男","女"),timeby=700,pval=T,
        pval.coord=c(1000,0.4),pval.testname=,
        linecols="Set2",table=T,label.nrisk="生存人数")

设置Numbers at risk字体大小，默认是10

svyjskm(s1,xlabs = "时间",ylabs = "生存率",ystrataname = "性别",
        ystratalabs=c("男","女"),timeby=700,pval=T,
        pval.coord=c(1000,0.4),pval.testname=T,
        linecols="Set2",table=T,label.nrisk="生存人数",size.label.nrisk=12)

给曲线加上百分比显示

svyjskm(s1,xlabs = "时间",ylabs = "生存率",ystrataname = "性别",
        ystratalabs=c("男","女"),timeby=700,pval=T,
        pval.coord=c(1000,0.4),pval.testname=T,
        linecols="Set2",table=T,label.nrisk="生存人数",size.label.nrisk=12,showpercent = T)

cut.landmark是设置一个新的地标

svyjskm(s1,xlabs = "时间",ylabs = "生存率",ystrataname = "性别",
        ystratalabs=c("男","女"),timeby=700,pval=T,
        pval.coord=c(1000,0.4),pval.testname=T,
        linecols="Set2",table=T,label.nrisk="生存人数",size.label.nrisk=12,showpercent = T,cut.landmark=2100)