Analysis of healthcare data

README.md

The US government released medicare data in early 2013. This is an analysis of that data across the US. It compares the cost of treatments per hospital and state vs. total medicare coverage and shows where there is greatest discrepancy.

healthcare.r

## Ilan Man
## 05/17/2013
## U.S Medicare Provider Charge data
 
library(ggplot2)
setwd('D:/Users/iman/Desktop')
hospital<-read.csv('healthcaredata.csv')
hospital[,3]<-sub(" HOSPITAL",'',as.character(hospital[,3]))
hospital[,3]<-sub(" CENTER",'',hospital[,3])
 
## narrow data
# Select 4 states in the northeast - PA, NY, NJ, MA
unique(hospital[,5])
data<-hospital[hospital[,6]%in%c('NY'),]
conditions<-unique(data[,1])
data<-data[data[,1]==conditions[52],] ## diabetes
 
### plot maximum payment
 
sort.payment<-data[order(data[,11]),]
 
ggplot(data=sort.payment,aes(y=sort.payment[,11],
                             x=reorder(sort.payment[,3],sort.payment[,11])))+
  geom_bar(colour="black", fill="#DE1887", width=.5, stat="identity")+
  xlab("Hospital")+ylab("Average cost")+
  scale_x_discrete(breaks=NULL)+
  ggtitle("Average cost for treating patients diagnosed with diabetes")
 
### plot maximum covered charge
 
sort.charges<-data[order(data[,10]),]
 
ggplot(data=sort.charges,aes(y=sort.charges[,10],
                             x=reorder(sort.charges[,3],sort.charges[,10])))+
  geom_bar(colour="black", fill="#DE1887", width=.5, stat="identity")+
  xlab("Hospital")+ylab("Average covered charge")+
  scale_x_discrete(breaks=NULL)+
  ggtitle("Average covered charge for treating patients diagnosed with diabetes")
 
### find maximum difference 
 
diff<-data.frame("City"=data[,5],"Hospital"=data[,3], "Diff"=as.numeric(data[,10]-data[,11]))
diff<-diff[order(diff$Diff,decreasing=TRUE),][1:20,] ## top 20 hospitals
 
ggplot(data=diff,aes(y=Diff,x=reorder(diff$Hospital,diff$Diff)))+
  geom_bar(colour="black", fill="#DE1887", width=.5, stat="identity")+
  xlab("Top 20 Hospitals")+ylab("Average Difference")+
  scale_x_discrete(breaks=NULL)+
  ggtitle("Average difference between covered charges and total\n payments for treating patients diagnosed with diabetes")
 
 
library(maps)
 
hospital<-read.csv('healthcaredata.csv')
conditions<-unique(hospital[,1])
rm(avCost)
data<-hospital[hospital[,1]==conditions[52],]
avCost<-data.frame(Payment=unlist(tapply(data[,11],factor(data[,6]),mean)),
                   CovCharge=unlist(tapply(data[,10],factor(data[,6]),mean)))
avCost$State<-row.names(avCost$Payment)
avCost$PayCat <- as.numeric(cut(avCost$Payment,8))
avCost$CovCat <- as.numeric(cut(avCost$CovCharge,8))
cols <- c("#800000", "#8B0000", "#A52A2A", "#B22222",
          "#FF7F50", "#DC143C","#FF6347", "#FF0000")
par(mfrow=c(2,1))
map("state", col=cols[avCost$PayCat],fill=TRUE)
title("$ amount of Total Payments")
map("state", col=cols[avCost$CovCat],fill=TRUE)
title("$ amount of Covered Charge")