如何在单变量分析中有效地循环变量？-解网

问：

我写了下面的代码来运行 cox 回归并将它们输出到表中。我想对许多变量运行单变量。你能帮我使用一个循环来让它更快吗？变量有 Grade、DCIS、LVE、Tsize 等......

我写的代码如下：

CoxphToDF <- function(y) {
    stopifnot(class(y) == "summary.coxph")
    cbind(y[["coefficients"]], 
          `lower 0.95` = y[["conf.int"]][, "lower .95"], 
          `upper 0.95` = y[["conf.int"]][, "upper .95"])
          }

Grade <- data[!is.na(data$Grade) & data$Grade != "", ]
df1 = CoxphToDF(summary(coxph(Surv(OS, OS.Event) ~ Grade, data = Grade)))

DCIS <- data[!is.na(data$DCIS) & data$DCIS != "", ]
df2 = CoxphToDF(summary(coxph(Surv(OS, OS.Event) ~ DCIS, data = DCIS)))

LVE <- data[!is.na(data$LVE) & data$LVE != "", ]
df3 = CoxphToDF(summary(coxph(Surv(OS, OS.Event) ~ LVE, data = LVE)))

Tsize <- data[!is.na(data$Tsize) & data$Tsize != "", ]
df4 = CoxphToDF(summary(coxph(Surv(OS, OS.Event) ~ Tsize, data = Tsize)))

write.table(rbind(df1,df2,df3,df4), "univariate_results.txt", sep="\t", col.names = NA)

我尝试了以下方法：

#vector of variables
variables_to_test <- c("Grade", "DCIS", "LVE", "Tsize")

#list to store the results
results_list <- list()

# Iterate
for (variable in variables_to_test) {
  subset_data <- data[!is.na(data[[variable]]) & data[[variable]] != "", ]
  result <- CoxphToDF(summary(coxph(Surv(OS, OS.Event) ~ ., data = subset_data)))
  results_list[[variable]] <- result
}

# single data frame
results_df <- do.call(rbind, results_list)

它给了我意想不到的结果，数据框中的所有变量都没有被选中，而且结果很奇怪

R 数据帧分析 Cox 回归生存

#vector of variables
variables_to_test <- c("Grade", "DCIS", "LVE", "Tsize")

do.call(what = "rbind", 
        args = lapply(X = variables_to_test, 
                      FUN = \(x) {
                        data <- data[!is.na(data[[x]]) 
                                     & data[[x]] != "", ]
                        CoxphToDF(summary(coxph(Surv(OS, OS.Event) ~ ., data = data)))
                      }
                      )
        )

请注意，每个模型拟合都会发出警告

1: In coxph.fit(X, Y, istrat, offset, init, control, weights = weights,  :
  Ran out of iterations and did not converge

我看不出有什么优势。考虑将结果存储在列表中，即 .do.call(what = "rbind", args = something)list_of_sth <- lapply(...

上一个：在术语图 R 中将 x 轴更改为对数刻度

下一个：如何将 cox 回归输出转换为数据帧？

如何在单变量分析中有效地循环变量？

How to effectively loop variables in univariate analysis?

评论