如何通过检查多列中的多个值来对新变量进行有条件的编码

Question

我想搜索一组变量(DX1-DX5)中的一系列值(ICD9代码列表或ICD10代码列表)，条件是DXVER=9搜索ICD9列表，DXVER=10搜索ICD10列表。这是可以手工写出来的，但是因为变量和代码列表可能会更长，所以我想通过按名称引用组来学习这一点。

我已经尝试了if else语句、rowsum和编码每个步骤。这些是可以工作的，尽管我总是遗漏了三个要求中的一个。1.以DXVER为条件2. DX1-DX5列表参考DXS 3.代码列表参考ICD9和ICD10

ICD9 <- c('042', '7953', '79571', 'V08')
ICD10 <- c('888', 'a10', 'b10', '987')

DXS <- c(paste('DX', seq(1:5), sep = '')) 

MergedData<-data.frame(DXVER=c("9", "9", "0", "0", "9"),
                   DX1 = c('042',"1","1","2","1"),
                   DX2 = c("4","3",'2',"4","3"),
                   DX3 = c("2","2","4","2","2"),
                   DX4 = c("3",'101',"2","3","2"),
                   DX5 = c('79571',"3",'888',"2","2"),
                   ID  = c(345, 123, 246, 432, 121))

MergedData$HIV_claim<NA
MergedData$HIV_claim[MergedData$DX1 %in% ICD9 & MergedData$DXVER=="9"] <- 
1 
MergedData$HIV_claim[MergedData$DX2 %in% ICD9 & MergedData$DXVER=="9"] <- 
1 
MergedData$HIV_claim[MergedData$DX3 %in% ICD9 & MergedData$DXVER=="9"] <- 
1 
MergedData$HIV_claim[MergedData$DX4 %in% ICD9 & MergedData$DXVER=="9"] <- 
1 
MergedData$HIV_claim[MergedData$DX5 %in% ICD9 & MergedData$DXVER=="9"] <- 
1 

MergedData$HIV_claim[MergedData$DX1 %in% ICD10 & MergedData$DXVER=="0"] <- 
1 
MergedData$HIV_claim[MergedData$DX2 %in% ICD10 & MergedData$DXVER=="0"] <- 
1 
MergedData$HIV_claim[MergedData$DX3 %in% ICD10 & MergedData$DXVER=="0"] <- 
1    
MergedData$HIV_claim[MergedData$DX4 %in% ICD10 & MergedData$DXVER=="0"] <- 
1 
MergedData$HIV_claim[MergedData$DX5 %in% ICD10 & MergedData$DXVER=="0"] <- 
1 


 MergedData2 <- MergedData

 MergedData2$HIV_claim9<NA
MergedData2$HIV_claim10<NA
MergedData2$HIV_claim9 <-  ((rowSums(MergedData2[, DXS] == ICD9,  na.rm=T) 
> 0) * 1)
MergedData2$HIV_claim10 <- ((rowSums(MergedData2[, DXS] == ICD10, na.rm=T) 
> 0) * 1)

MergedData2$HIV_claim  <-  ifelse(MergedData2$DXVER == 9, 
ifelse(MergedData2$DX1 %in% ICD9 |  MergedData2$DX2 %in% ICD9
 |  MergedData2$DX3 %in% ICD9 |  MergedData2$DX4 %in% ICD9
|  MergedData2$DX5 %in% ICD9,1,0),0)

Answer 1

library(dplyr)

md %>% gather(dx,v,-DXVER,-ID) %>% 
  mutate(hiv = ifelse((v %in% ICD9 & DXVER=='9') | (v %in% ICD10 & DXVER=='0'),1,NA)) %>% 
  filter(hiv==1) %>% group_by(ID) %>% slice(1) %>% 
  select(ID,hiv) %>% left_join(md,.)

Joining, by = "ID"
  DXVER DX1 DX2 DX3 DX4   DX5  ID hiv
1     9 042   4   2   3 79571 345   1
2     9   1   3   2 101     3 123  NA
3     0   1   2   4   2   888 246   1
4     0   2   4   2   3     2 432  NA
5     9   1   3   2   2     2 121  NA