如何基于合并纵向研究数据集构建并列对比分析表格

需求说明 #

我有一个合并数据集，包含两项纵向研究（study1与study2）的测量记录，示例数据如下：

data <- data.frame(
  weight_study1 = rnorm(100, mean = 60, sd = 5),  # 数值型变量
  weight_study2 = rnorm(100, mean = 62, sd = 6),  # 数值型变量
  gender_study1 = sample(c("Male", "Female"), 100, replace = TRUE),  # 分类变量
  gender_study2 = sample(c("Male", "Female"), 100, replace = TRUE)  # 分类变量
)

需要创建一个并列表格，将study1与study2的分析结果分别置于表格两侧，期望输出样式如下：

Characteristics Pilot study 1 Pilot study 2 
Age             34(3)         67(8)
Gender          Male (40%)    Male(30%)
                Female (60%)  Female (70%)

解决方案 #

下面提供两种在R中实现的方法，按需选择：

方式一：用tableone包快速生成（推荐） #

tableone是专门生成基线特征对比表的工具，步骤如下：

1. 安装并加载依赖包

install.packages(c("tableone", "dplyr"))
library(tableone)
library(dplyr)

2. 把宽格式数据转成适合分组的长格式

# 拆分并标记两组数据
data_study1 <- data %>%
  select(weight = weight_study1, gender = gender_study1) %>%
  mutate(study = "Pilot study 1")

data_study2 <- data %>%
  select(weight = weight_study2, gender = gender_study2) %>%
  mutate(study = "Pilot study 2")

# 合并成长格式
data_long <- bind_rows(data_study1, data_study2)

3. 生成并输出Markdown格式表格

# 指定变量类型
vars <- c("weight", "gender")
catVars <- c("gender")

# 创建特征表
table <- CreateTableOne(vars = vars, 
                        factorVars = catVars, 
                        data = data_long, 
                        strata = "study")

# 输出为Markdown格式
print(table, printToggle = FALSE, format = "markdown")

输出示例：

方式二：手动计算后构建表格 #

如果不想用第三方包，可手动统计后生成表格：

# 计算study1的统计值
weight1_stats <- sprintf("%.1f(%.1f)", mean(data$weight_study1), sd(data$weight_study1))
gender1_male <- sprintf("Male (%.0f%%)", mean(data$gender_study1 == "Male")*100)
gender1_female <- sprintf("Female (%.0f%%)", mean(data$gender_study1 == "Female")*100)

# 计算study2的统计值
weight2_stats <- sprintf("%.1f(%.1f)", mean(data$weight_study2), sd(data$weight_study2))
gender2_male <- sprintf("Male (%.0f%%)", mean(data$gender_study2 == "Male")*100)
gender2_female <- sprintf("Female (%.0f%%)", mean(data$gender_study2 == "Female")*100)

# 构建表格数据框
table_df <- data.frame(
  Characteristics = c("Weight", "Gender", ""),
  `Pilot study 1` = c(weight1_stats, gender1_male, gender1_female),
  `Pilot study 2` = c(weight2_stats, gender2_male, gender2_female)
)

# 输出Markdown表格
knitr::kable(table_df, align = "lll")

输出示例：

内容的提问来源于stack exchange，提问作者user19807620