ahrf2016r: R-compatible HHS HRSA Area Health Resource File datasets (2016)

About the Data

Source United States Department of Health and Human Services Health Resources and Services Administration

HRSA Data Warehouse: https://datawarehouse.hrsa.gov/data/datadownload.aspx

Summary
The Area Health Resource File (ARF) is a health resource information database containing more than 6,000 variables for each of the nation's counties. ARF contains information on health facilities, health professions, measures of resource scarcity, health status, economic activity, health training programs, and socioeconomic and environmental characteristics.

Detailed information about the AHRF dataset See the files in the metadata subdirectory:

• AHRF User Guide: http://bit.ly/2ngOWBw
• AHRF Technical Documentation (Data Dictionary): http://bit.ly/2mY3pzS
• Health Care Disparities in Rural Areas Fact Sheet: http://bit.ly/2nrsPZw
• National Health Care Quality Report: http://bit.ly/2nMRTvf

Installation

This repository is designed to download and install the datasets as an R package. The size of ahrf_county.rda is 24.2 MB, so it might take a while to install and load into memory.

# install.packages("devtools")
devtools::install_github("olyerickson/ahrf2016r")

# To uninstall the package, use:
# remove.packages("ahrf2016r")

Notes for Machine Learning Applications

The data directory contains the full county data, plus training (75% of observations) and test (25%) subsets:

ahrf_county.rda
ahrf_county_train.rda
ahrf_county_test.rda

NOTE:

• Only ahrf_county_train.rda will be available before and during the 2017 RPI Datathon (25-26 Mar 2017)
• ahrf_county.rda and ahrf_count_test.rda will become available after 26 Mar 2017

Example Usage

There are 3230 rows and 6921 columns in the full county file (wide format):

NOTE: Substitute ahrf_county_train as needed...

library(dplyr, warn.conflicts = FALSE)
dim(ahrf2016r::ahrf_county)
#> [1]  3230 6921

What columns mention "beds"?

> beds_rows <- ahrf_county_layout[grep("beds", ahrf_county_layout$label_2, ignore.case=T),]
> View(beds_rows)

What columns mention "population"?

> pop_rows <- ahrf_county_layout[grep("population", ahrf_county_layout$label_2, ignore.case=T),]
> View(pop_rows)

County-level hospital beds in 2016

> ahrf2016r::ahrf_county %>% 
        select(county = F04437, 
               fips = F00002, 
               beds_2016 = F0892113,
               pop_2016 = F1198415) %>% 
        mutate(beds_2016 = as.integer(beds_2016),
               pop_2016 = as.integer(pop_2016),
               beds_2016_p10k = beds_2016 / pop_2016 * 10000) -> beds
> beds
# A tibble: 3,230 × 5
         county  fips beds_2016 pop_2016 beds_2016_p10k
          <chr> <chr>     <int>    <int>          <dbl>
1   Autauga, AL 01001        50    55347       9.033913
2   Baldwin, AL 01003       364   203709      17.868626
3   Barbour, AL 01005        47    26489      17.743214
4      Bibb, AL 01007        20    22583       8.856219
5    Blount, AL 01009        40    57673       6.935654
6   Bullock, AL 01011        54    10696      50.486163
7    Butler, AL 01013        83    20154      41.182892
8   Calhoun, AL 01015       458   115620      39.612524
9  Chambers, AL 01017       188    34123      55.094804
10 Cherokee, AL 01019        45    25859      17.402065
# ... with 3,220 more rows
> summary(beds$beds_2016)
   Min. 1st Qu.  Median    Mean 3rd Qu.    Max.    NA's 
    0.0    17.0    50.0   294.6   181.5 25310.0       7 
> summary(beds$pop_2016)
    Min.  1st Qu.   Median     Mean  3rd Qu.     Max.     NA's 
      89    11280    25960   100900    66750 10170000        9 
> summary(beds$beds_2016_p10k)
   Min. 1st Qu.  Median    Mean 3rd Qu.    Max.    NA's 
  0.000   7.367  20.090  30.680  37.250 797.000      14 
> quantile(beds$beds_2016_p10k, na.rm = TRUE)
        0%        25%        50%        75%       100% 
  0.000000   7.366996  20.093600  37.248540 797.016672 

Geographic Distribution of Hospital Beds

beds$beds_2016_dist = Hmisc::cut2(beds$beds_2016_p10k, cuts = c(7.63, 20.50, 38.09))

# devtools::install_github("jjchern/usmapdata")
usmapdata::county %>% 
  left_join(beds, by = c("id" = "fips")) %>% 
        mutate(region = id) -> beds_map

usmapdata::state %>% 
        mutate(region = id) -> state_map

library(ggplot2)
ggplot() +
  geom_map(data = beds_map, map = beds_map,
           aes(x = long, y = lat, map_id = id, fill = beds_2016_dist),
           colour = alpha("white", 0.1), size=0.2) +
  geom_map(data = state_map, map = state_map,
           aes(x = long, y = lat, map_id = region),
           colour = "white", fill = "NA") +
  coord_map("albers", lat0 = 30, lat1 = 40) +
  viridis::scale_fill_viridis(discrete=TRUE, option = "D") +
  ggtitle("Hospital Beds per 10,000 Population in 2016") +
  ggthemes::theme_map() +
  theme(legend.position = c(.85, .3),
        legend.title=element_blank())