使用PEC包计算C指数的时候报错

Question

使用PEC包计算C指数的时候报错

f<-cph(Surv(survival.time,death)~Age+stroke+Angiotensin.II.receptor.blocker+hematocrit+albumin+creatinine+alanine.aminotransferase+total_KtV,data=tdata,x=T,Y=T,surv = T)
set.seed(123)
c_index<-cindex(list("f"=f),
          eval.times=seq(0,3650,365),
          cens.model = "cox",
          keep.pvalues = T,
          confInt = T,
          confLevel = 0.95,
          splitMethod = "bootcv",
          B=1000)
报错内容
Error in UseMethod("predictEventProb", object) : 
  no applicable method for 'predictEventProb' applied to an object of class "c('cph', 'rms', 'coxph')"
In addition: Warning messages:
1: In cindex(list(`model f` = f), eval.times = seq(0, 3650, 365), cens.model = "cox",  :
  Formula missing. Using formula from first model
2: In cindex(list(`model f` = f), eval.times = seq(0, 3650, 365), cens.model = "cox",  :
  Argument data is missing. I use the data from the call to the first model instead.

Answer 1

两个问题，第一个问题cindex函数你没指定生存分析模型公式，这样修改：

c_index <- cindex(list("f"=f),
                  formula = Surv(survival.time, death) ~ Age + stroke + Angiotensin.II.receptor.blocker + hematocrit + albumin + creatinine + alanine.aminotransferase + total_KtV,
                  eval.times = seq(0, 3650, 365),
                  cens.model = "cox",
                  keep.pvalues = T,
                  confInt = T,
                  confLevel = 0.95,
                  splitMethod = "bootcv",
                  B = 1000)

第二个f对象类型不对

library(rms)
model <- cph(Surv(survival.time, death) ~ Age + stroke + Angiotensin.II.receptor.blocker + hematocrit + albumin + creatinine + alanine.aminotransferase + total_KtV, data = tdata, x = T, Y = T, surv = T)
c_index <- cindex(list("model" = model),
                  eval.times = seq(0, 3650, 365),
                  cens.model = "cox",
                  keep.pvalues = T,
                  confInt = T,
                  confLevel = 0.95,
                  splitMethod = "bootcv",
                  B = 1000)