Assignment 1
Cassandra Boylan
08/29/2021
Introduction
In this analysis, I took a look at a dataset that examined the “at risk population” within various metro areas around the United States. The dataset also captured the number of hospitals and icu beds available in each metro area as a measure of how prepared each may be in meeting the health demands of the current pandemic.
library(dplyr)
library(tidyverse)
library(stringr)Importing Dataset
urlfile=("https://raw.githubusercontent.com/fivethirtyeight/data/master/covid-geography/mmsa-icu-beds.csv")
state_icu_beds<-read_csv(url(urlfile),show_col_types = FALSE)Renaming Columns
names(state_icu_beds)[1] <- "city_state"
names(state_icu_beds)[7] <- "high_risk_population"
names(state_icu_beds)[3] <- "high_risk_pop_per_icu_bed"
names(state_icu_beds)[4] <- "high_risk_pop_per_hospital"state_icu_beds_clean <- separate(state_icu_beds, col=city_state, into=c('metro_area', 'state'), sep=', ')state_icu_beds_clean$high_risk_pop_per_icu_bed <- round(state_icu_beds_clean$high_risk_pop_per_icu_bed,digits=0)
state_icu_beds_clean$high_risk_pop_per_hospital <- round(state_icu_beds_clean$high_risk_pop_per_hospital,digits=0)state_icu_beds_clean <- na.omit(state_icu_beds_clean)Summary Statistics of Hospitals per Metro Area
hospitals <- state_icu_beds_clean$hospitals
fivenum(hospitals)## [1] 1 5 9 18 100hist(hospitals,
main="Hospitals per US metro area",
xlim=c(0,60),
col="blue")
Summary Statistics of ICU beds per Metro Area
icu_beds <- state_icu_beds_clean$icu_beds
fivenum(icu_beds)## [1] 8.0 89.5 221.0 489.5 2777.0Metro Area with fewest count of ICU beds
state_icu_beds_clean[which.min(state_icu_beds_clean$icu_beds),0:2]## # A tibble: 1 x 2
## metro_area state
## <chr> <chr>
## 1 Manhattan KSMetro Area with greatest count of ICU beds
state_icu_beds_clean[which.max(state_icu_beds_clean$icu_beds),0:2]## # A tibble: 1 x 2
## metro_area state
## <chr> <chr>
## 1 Los Angeles-Long Beach-Anaheim CANumber of urban areas analyzed per state
metro_by_state <- state_icu_beds_clean %>% count(state, sort=TRUE)
head(metro_by_state)## # A tibble: 6 x 2
## state n
## <chr> <int>
## 1 FL 12
## 2 TX 9
## 3 SC 6
## 4 CA 4
## 5 KS 4
## 6 NE 4Subset to Metro Areas in Florida
by_state_FL <- subset(state_icu_beds_clean, state == 'FL', select = c("metro_area", "icu_beds", "hospitals", "high_risk_population", "high_risk_pop_per_icu_bed","high_risk_pop_per_hospital"))
icu_beds_FL <- by_state_FL$icu_beds
hospitals_FL <- by_state_FL$hospitals
at_risk_FL <- by_state_FL$high_risk_populationDistribution of Hospitals per Metro Area in FL
fivenum(hospitals_FL)## [1] 2 3 6 12 43hist(hospitals_FL,
main="Hospitals per FL metro area",
xlim=c(0,30),
col="blue")
FL Metro Area with highest count of Hospitals
by_state_FL[which.max(by_state_FL$hospitals),]## # A tibble: 1 x 6
## metro_area icu_beds hospitals high_risk_popula~ high_risk_pop_per_~
## <chr> <dbl> <dbl> <dbl> <dbl>
## 1 Miami-Fort Lauderdal~ 1581 43 2749081. 1739
## # ... with 1 more variable: high_risk_pop_per_hospital <dbl>Range of ICU Beds in FL by Metro Area
select(by_state_FL, metro_area, icu_beds,hospitals,high_risk_pop_per_icu_bed) %>% arrange(icu_beds)## # A tibble: 12 x 4
## metro_area icu_beds hospitals high_risk_pop_per_i~
## <chr> <dbl> <dbl> <dbl>
## 1 Tallahassee 60 3 2737
## 2 Panama City 77 3 1390
## 3 Crestview-Fort Walton Beach-Destin 85 5 1639
## 4 Pensacola-Ferry Pass-Brent 136 6 1811
## 5 Port St. Lucie 143 3 1713
## 6 Deltona-Daytona Beach-Ormond Beach 166 6 2304
## 7 North Port-Sarasota-Bradenton 208 7 2183
## 8 Gainesville 210 2 621
## 9 Jacksonville 487 12 1579
## 10 Orlando-Kissimmee-Sanford 650 12 1694
## 11 Tampa-St. Petersburg-Clearwater 921 26 1717
## 12 Miami-Fort Lauderdale-West Palm Beach 1581 43 1739fivenum(icu_beds_FL)## [1] 60.0 110.5 187.0 568.5 1581.0Florida Ratio of ICU beds available per 10k of high risk population
by_state_FL$icu_beds_per_10k <- round(icu_beds_FL/(at_risk_FL/10000),digits=2)
icu_per_10k <- by_state_FL$icu_beds_per_10k
fivenum(icu_per_10k)## [1] 3.650 5.050 5.830 6.215 16.090hist(icu_per_10k,
main="ICU beds per 10k of high risk pop",
xlab="beds per every 10k",
xlim=c(0,10),
col="chartreuse4")
Avg ICU Beds per Hospital in FL Metro Areas
by_state_FL$icu_beds_per_hospital <- round(icu_beds_FL/hospitals_FL,digits=0)
beds_per_hospital <- by_state_FL$icu_beds_per_hospital
fivenum(beds_per_hospital)## [1] 17.0 24.5 32.5 44.5 105.0hist(beds_per_hospital,
main="Average ICU beds per hospital",
xlab="beds per hospital per FL metro area",
xlim=c(0,60),
col="blue")
Conclusion
I wanted to see the number of hospitals available on average per each metro area and the distribution frequency of hospital count per metro area. As we can see, the number skews heavily to less than 10 for the majority of metro areas surveyed, with an average of 9 and IQR of 5-18.
It was striking to see the range of ICU beds available from less than 10 to over 2500. Unsurprisingly, the metro area with the largest number of beds serves the greater Los Angeles area while the metro area with the fewest is located in the Midwest.
I chose to subset the dataset to focus on the state of Florida which likely is burdened with some of the largest at-risk population density.
In the 12 metro areas measured, 8 have less than 10 hospitals, with an average of 6 and a minimum of 2.
Tallahasee has the least ICU beds available at any given time, and ranks as one of the most overburdened metro areas with a single bed available for every ~2700 high risk individuals
Miami Fort Lauderdale area ranks as having the highest number of hospitals with a grand count of 43 - due to its population burden however, it only has a bed available for every ~1700 high risk residents.
On average in Florida, there is a ratio of roughly 2-6 ICU beds available for every 10k people deemed high risk.
```