Overview

What regions in the United States are best equipped to deliver intensive care treatment to COVID-19 patients? This dataset compares the number of people at high-risk for requiring ICU treatment due to COVID-19 with the number of ICU beds in that region. In regions where the number of high-risk residents is large relative to ICU beds, there is greater danger that hospital resources will fail to meet the need for medical care. In regions where the number of high-risk residents is low relative to ICU beds, it is more likely that hospital resources will be adequate to the demand for ICU care during the pandemic. Click here for the original article at FiveThirtyEight.

Importing the data

library(tidyverse)
## -- Attaching packages --------------------------------------- tidyverse 1.3.0 --
## v ggplot2 3.3.3     v purrr   0.3.4
## v tibble  3.0.4     v dplyr   1.0.2
## v tidyr   1.1.2     v stringr 1.4.0
## v readr   1.4.0     v forcats 0.5.0
## -- Conflicts ------------------------------------------ tidyverse_conflicts() --
## x dplyr::filter() masks stats::filter()
## x dplyr::lag()    masks stats::lag()
data_on_github = "https://raw.githubusercontent.com/fivethirtyeight/data/master/covid-geography/mmsa-icu-beds.csv"

covid_geography <- read_csv(url(data_on_github))
## 
## -- Column specification --------------------------------------------------------
## cols(
##   MMSA = col_character(),
##   total_percent_at_risk = col_character(),
##   high_risk_per_ICU_bed = col_double(),
##   high_risk_per_hospital = col_double(),
##   icu_beds = col_double(),
##   hospitals = col_double(),
##   total_at_risk = col_double()
## )

Selecting and renaming columns

covid_geography <- covid_geography[, c(1, 2, 3, 5, 6, 7)]

covid_geography <- rename(covid_geography, c("region" = "MMSA", "percent_residents_at_risk" = "total_percent_at_risk", "residents_at_risk_per_ICU_bed" = "high_risk_per_ICU_bed", "ICU_beds" = "icu_beds", "residents_at_risk" = "total_at_risk"))

names(covid_geography)
## [1] "region"                        "percent_residents_at_risk"    
## [3] "residents_at_risk_per_ICU_bed" "ICU_beds"                     
## [5] "hospitals"                     "residents_at_risk"

Visualizing residents at risk per ICU bed

ggplot(data = covid_geography, mapping = aes(x = residents_at_risk_per_ICU_bed)) +
  geom_histogram()
## `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
## Warning: Removed 1 rows containing non-finite values (stat_bin).

Conclusions

As noted in the FiveThirtyEight article, Hilton Head, SC is one of the regions in greatest danger. Only Manhattan, KS has a greater ratio of residents at risk to ICU beds. The authors may have chosen to focus on Hilton Head rather than Manhattan, KS because Hilton Head has roughly three times the number of residents at risk.

There is also significant variability across regions in the ratio of residents at risk to ICU beds, as can be seen in the histogram above.

Identifying which regions are most likely to face shortages of health care due to COVID-19 can help decision-makers move resources to where they’re most needed.

To further investigate which regions of the US are at greatest risk of overwhelming their hospitals, it would be useful to have some measure of how flexible hospital resources are. The article described how Coastal Carolina Hospital has added ICU capacity, and studying how to train a more flexible workforce. Hospitals that are able to repurpose their space and staff are less likely to be overwhelmed, regardless of their initial ICU bed capacity.

Questions