使用R语言mstate包做多状态分析时msprep报错的解决求助

使用mstate包的msprep函数构建多状态分析数据框时，出现错误：Error in time[, -startings] : incorrect number of dimensions，已确认变量存在、拼写正确、无缺失值，转移矩阵定义符合预期，起始状态的time和status设为NA也符合官方文档要求。以下是可行的解决方案：

• 检查数据集类型，转换为普通data.frame：
  如果数据集是tidyverse的tibble格式，部分旧版本mstate包对其兼容性不佳，可转换为普通data.frame后重试：

  d <- as.data.frame(d)
  # 重新运行msprep
  dlong <- msprep(time = c(NA, "aki_1_time",  "rec_1_time", "aki_2_time",  "rec_2_time", 
                           "aki_3_time",  "rec_3_time", "aki_4_time",  "rec_4_time", "death_time"),
                  status = c(NA, "aki_1_status",  "rec_1_status", "aki_2_status",  "rec_2_status", 
                             "aki_3_status",  "rec_3_status", "aki_4_status",  "rec_4_status", "death_status"),
                  data = d, id = "subject", trans = tmat)
  

• 修改起始状态的time和status参数：
  将起始状态的time设为0（代表初始时间点），status设为1（代表所有个体均进入起始状态），替换原有的NA：

  dlong <- msprep(time = c(0, "aki_1_time",  "rec_1_time", "aki_2_time",  "rec_2_time", 
                           "aki_3_time",  "rec_3_time", "aki_4_time",  "rec_4_time", "death_time"),
                  status = c(1, "aki_1_status",  "rec_1_status", "aki_2_status",  "rec_2_status", 
                             "aki_3_status",  "rec_3_status", "aki_4_status",  "rec_4_status", "death_status"),
                  data = d, id = "subject", trans = tmat)
  

• 手动构建转移矩阵：
  避免transMat的list输入可能引发的索引问题，手动构建转移矩阵：

  # 创建10*10的空转移矩阵，命名状态
  tmat <- matrix(NA, nrow = 10, ncol = 10,
                 dimnames = list(
                   c("start", "aki_1", "rec_1", "aki_2", "rec_2", "aki_3", "rec_3", "aki_4", "rec_4", "death"),
                   c("start", "aki_1", "rec_1", "aki_2", "rec_2", "aki_3", "rec_3", "aki_4", "rec_4", "death")
                 ))
  # 定义各状态的允许转移及转移编号
  tmat["start", c("aki_1", "death")] <- c(1, 2)
  tmat["aki_1", c("rec_1", "death")] <- c(3, 4)
  tmat["rec_1", c("aki_2", "death")] <- c(5, 6)
  tmat["aki_2", c("rec_2", "death")] <- c(7, 8)
  tmat["rec_2", c("aki_3", "death")] <- c(9, 10)
  tmat["aki_3", c("rec_3", "death")] <- c(11, 12)
  tmat["rec_3", c("aki_4", "death")] <- c(13, 14)
  tmat["aki_4", c("rec_4", "death")] <- c(15, 16)
  tmat["rec_4", "death"] <- 17
  
  # 重新运行msprep
  dlong <- msprep(time = c(NA, "aki_1_time",  "rec_1_time", "aki_2_time",  "rec_2_time", 
                           "aki_3_time",  "rec_3_time", "aki_4_time",  "rec_4_time", "death_time"),
                  status = c(NA, "aki_1_status",  "rec_1_status", "aki_2_status",  "rec_2_status", 
                             "aki_3_status",  "rec_3_status", "aki_4_status",  "rec_4_status", "death_status"),
                  data = d, id = "subject", trans = tmat)
  

内容的提问来源于stack exchange，提问作者davidbuis