如何用R语言table1()函数实现供体与受体变量分组表格布局？

问题描述 #

我正在尝试用R的table1()函数生成表格，想要模仿目标格式（Desired layout of Table1），但当前生成的表格（Generated Table1 from table1() function）无法实现供体（Donor）和受体（Recipient）变量的分组行布局（对应图中标注g1、g2的行）。我试过用键分离变量、用strata参数按行分层都没成功，当前代码如下：

library(table1)

label(t1data$`Recipient Age`) <- "Age, y"
label(t1data$`Recipient Sex`) <- "Sex"
label(t1data$`Recipient Race`) <- "Race"
label(t1data$`Recipient Ethnicity`) <- "Ethnicity"
 
label(t1data$`Donor Age`) <- "Age, y"
label(t1data$`Donor Gender`) <- "Sex"
label(t1data$`Donor Race`) <- "Race"
label(t1data$`Donor Ethnicity`) <- "Ethnicity"

table1( ~ `Recipient Age` + `Recipient Sex` + `Recipient Ethnicity` + `Recipient Race` +
          `Donor Age` + `Donor Gender` + `Donor Ethnicity` + `Donor Race` | Status, data = t1data, overall = FALSE, rowlabelhead = "Demographics")

我想知道是需要先预处理数据，还是table1()本身有参数可以实现这个需求？

解决方法 #

不需要额外预处理数据，table1()函数本身支持通过公式分组或groups参数实现行分组布局，具体有两种可行方式：

方式1：公式内括号分组变量 #

直接在table1()的公式中，将受体、供体变量分别放在括号内，每组会自动生成分组标题行，还可通过groupLabels自定义标题：

library(table1)

# 保留原变量标签设置
label(t1data$`Recipient Age`) <- "Age, y"
label(t1data$`Recipient Sex`) <- "Sex"
label(t1data$`Recipient Race`) <- "Race"
label(t1data$`Recipient Ethnicity`) <- "Ethnicity"
 
label(t1data$`Donor Age`) <- "Age, y"
label(t1data$`Donor Gender`) <- "Sex"
label(t1data$`Donor Race`) <- "Race"
label(t1data$`Donor Ethnicity`) <- "Ethnicity"

# 修改公式，按受体/供体分组变量
table1( ~ (Recipient Age + Recipient Sex + Recipient Ethnicity + Recipient Race) + 
          (Donor Age + Donor Gender + Donor Ethnicity + Donor Race) | Status, 
        data = t1data, 
        overall = FALSE, 
        rowlabelhead = "Demographics",
        groupLabels = c("Recipient", "Donor")) # 自定义分组标题

方式2：用groups参数指定分组 #

如果变量命名有统一前缀（如Recipient/Donor），可先定义分组映射向量，再传入groups参数：

library(table1)

# 保留原变量标签设置
label(t1data$`Recipient Age`) <- "Age, y"
label(t1data$`Recipient Sex`) <- "Sex"
label(t1data$`Recipient Race`) <- "Race"
label(t1data$`Recipient Ethnicity`) <- "Ethnicity"
 
label(t1data$`Donor Age`) <- "Age, y"
label(t1data$`Donor Gender`) <- "Sex"
label(t1data$`Donor Race`) <- "Race"
label(t1data$`Donor Ethnicity`) <- "Ethnicity"

# 定义变量分组映射
var_groups <- ifelse(grepl("Recipient", names(t1data)), "Recipient", 
                     ifelse(grepl("Donor", names(t1data)), "Donor", NA))

table1( ~ `Recipient Age` + `Recipient Sex` + `Recipient Ethnicity` + `Recipient Race` +
          `Donor Age` + `Donor Gender` + `Donor Ethnicity` + `Donor Race` | Status, 
        data = t1data, 
        overall = FALSE, 
        rowlabelhead = "Demographics",
        groups = var_groups)

补充说明 #

• 两种方式都会生成带有Recipient/Donor分组标题的行布局，匹配你需要的g1、g2行效果；
• 若要调整分组标题样式，可搭配render.grouplabel参数自定义（如加粗、调整字体等）；
• 全程无需转换数据格式，直接使用原宽格式数据即可。

内容的提问来源于stack exchange，提问作者user22770348