nppes.Rmd

# National Plan and Provider Enumeration System (NPPES) {-}

[![License: GPL v3](https://img.shields.io/badge/License-GPLv3-blue.svg)](https://www.gnu.org/licenses/gpl-3.0) <a href="https://github.com/asdfree/nppes/actions"><img src="https://github.com/asdfree/nppes/actions/workflows/r.yml/badge.svg" alt="Github Actions Badge"></a> 

The registry of every medical practitioner actively operating in the United States healthcare industry.

* A single large table with one row per enumerated health care provider.

* A census of individuals and organizations that bill for medical services in the United States.

* Updated weekly with new providers.

* Maintained by the United States [Centers for Medicare & Medicaid Services (CMS)](http://www.cms.gov/)

---

## Recommended Reading {-}

Two Methodology Documents:

> [NPI: What You Need To Know](https://www.cms.gov/Outreach-and-Education/Medicare-Learning-Network-MLN/MLNProducts/Downloads/NPI-What-You-Need-To-Know.pdf)

> [Wikipedia Entry](https://en.wikipedia.org/wiki/National_Provider_Identifier)

<br>

One Haiku:

```{r}
# how many doctors
# ranked sergeant, last name pepper
# practice in the states?
```

---

## Download, Import, Preparation {-}

Download and import the national file:
```{r eval = FALSE , results = "hide" }
library(readr)

tf <- tempfile()

npi_datapage <-
	readLines( "http://download.cms.gov/nppes/NPI_Files.html" )

latest_files <- grep( 'NPPES_Data_Dissemination_' , npi_datapage , value = TRUE )

latest_files <- latest_files[ !grepl( 'Weekly Update' , latest_files ) ]

this_url <-
	paste0(
		"http://download.cms.gov/nppes/",
		gsub( "(.*)(NPPES_Data_Dissemination_.*\\.zip)(.*)$", "\\2", latest_files )
	)

download.file( this_url , tf , mode = 'wb' )

npi_files <- unzip( tf , exdir = tempdir() )

npi_filepath <-
	grep(
		"npidata_pfile_20050523-([0-9]+)\\.csv" ,
		npi_files ,
		value = TRUE
	)

column_names <-
	names( 
		read.csv( 
			npi_filepath , 
			nrow = 1 )[ FALSE , , ] 
	)

column_names <- gsub( "\\." , "_" , tolower( column_names ) )

column_types <-
	ifelse( 
		grepl( "code" , column_names ) & 
		!grepl( "country|state|gender|taxonomy|postal" , column_names ) , 
		'n' , 'c' 
	)

columns_to_import <-
	c( "entity_type_code" , "provider_gender_code" , "provider_enumeration_date" ,
	"is_sole_proprietor" , "provider_business_practice_location_address_state_name" )

stopifnot( all( columns_to_import %in% column_names ) )

# readr::read_csv() columns must match their order in the csv file
columns_to_import <-
	columns_to_import[ order( match( columns_to_import , column_names ) ) ]

nppes_tbl <-
	readr::read_csv( 
		npi_filepath , 
		col_names = columns_to_import , 
		col_types = 
			paste0( 
				ifelse( column_names %in% columns_to_import , column_types , '_' ) , 
				collapse = "" 
			) ,
		skip = 1
	) 

nppes_df <- 
	data.frame( nppes_tbl )
```

### Save Locally \ {-}

Save the object at any point:

```{r eval = FALSE , results = "hide" }
# nppes_fn <- file.path( path.expand( "~" ) , "NPPES" , "this_file.rds" )
# saveRDS( nppes_df , file = nppes_fn , compress = FALSE )
```

Load the same object:

```{r eval = FALSE , results = "hide" }
# nppes_df <- readRDS( nppes_fn )
```

### Variable Recoding {-}

Add new columns to the data set:
```{r eval = FALSE , results = "hide" }
nppes_df <- 
	transform( 
		nppes_df , 
		
		individual = as.numeric( entity_type_code ) ,
		
		provider_enumeration_year =
			as.numeric( substr( provider_enumeration_date , 7 , 10 ) ) ,
		
		state_name = provider_business_practice_location_address_state_name
		
	)
```

---

## Analysis Examples with base R \ {-}

### Unweighted Counts {-}

Count the unweighted number of records in the table, overall and by groups:
```{r eval = FALSE , results = "hide" }
nrow( nppes_df )

table( nppes_df[ , "provider_gender_code" ] , useNA = "always" )
```

### Descriptive Statistics {-}

Calculate the mean (average) of a linear variable, overall and by groups:
```{r eval = FALSE , results = "hide" }
mean( nppes_df[ , "provider_enumeration_year" ] , na.rm = TRUE )

tapply(
	nppes_df[ , "provider_enumeration_year" ] ,
	nppes_df[ , "provider_gender_code" ] ,
	mean ,
	na.rm = TRUE 
)
```

Calculate the distribution of a categorical variable, overall and by groups:
```{r eval = FALSE , results = "hide" }
prop.table( table( nppes_df[ , "is_sole_proprietor" ] ) )

prop.table(
	table( nppes_df[ , c( "is_sole_proprietor" , "provider_gender_code" ) ] ) ,
	margin = 2
)
```

Calculate the sum of a linear variable, overall and by groups:
```{r eval = FALSE , results = "hide" }
sum( nppes_df[ , "provider_enumeration_year" ] , na.rm = TRUE )

tapply(
	nppes_df[ , "provider_enumeration_year" ] ,
	nppes_df[ , "provider_gender_code" ] ,
	sum ,
	na.rm = TRUE 
)
```

Calculate the median (50th percentile) of a linear variable, overall and by groups:
```{r eval = FALSE , results = "hide" }
quantile( nppes_df[ , "provider_enumeration_year" ] , 0.5 , na.rm = TRUE )

tapply(
	nppes_df[ , "provider_enumeration_year" ] ,
	nppes_df[ , "provider_gender_code" ] ,
	quantile ,
	0.5 ,
	na.rm = TRUE 
)
```

### Subsetting {-}

Limit your `data.frame` to California:
```{r eval = FALSE , results = "hide" }
sub_nppes_df <- subset( nppes_df , state_name = 'CA' )
```
Calculate the mean (average) of this subset:
```{r eval = FALSE , results = "hide" }
mean( sub_nppes_df[ , "provider_enumeration_year" ] , na.rm = TRUE )
```

### Measures of Uncertainty {-}

Calculate the variance, overall and by groups:
```{r eval = FALSE , results = "hide" }
var( nppes_df[ , "provider_enumeration_year" ] , na.rm = TRUE )

tapply(
	nppes_df[ , "provider_enumeration_year" ] ,
	nppes_df[ , "provider_gender_code" ] ,
	var ,
	na.rm = TRUE 
)
```

### Regression Models and Tests of Association {-}

Perform a t-test:
```{r eval = FALSE , results = "hide" }
t.test( provider_enumeration_year ~ individual , nppes_df )
```

Perform a chi-squared test of association:
```{r eval = FALSE , results = "hide" }
this_table <- table( nppes_df[ , c( "individual" , "is_sole_proprietor" ) ] )

chisq.test( this_table )
```

Perform a generalized linear model:
```{r eval = FALSE , results = "hide" }
glm_result <- 
	glm( 
		provider_enumeration_year ~ individual + is_sole_proprietor , 
		data = nppes_df
	)

summary( glm_result )
```

---

## Analysis Examples with `dplyr` \ {-}

The R `dplyr` library offers an alternative grammar of data manipulation to base R and SQL syntax. [dplyr](https://github.com/tidyverse/dplyr/) offers many verbs, such as `summarize`, `group_by`, and `mutate`, the convenience of pipe-able functions, and the `tidyverse` style of non-standard evaluation. [This vignette](https://cran.r-project.org/web/packages/dplyr/vignettes/dplyr.html) details the available features. As a starting point for NPPES users, this code replicates previously-presented examples:

```{r eval = FALSE , results = "hide" }
library(dplyr)
nppes_tbl <- as_tibble( nppes_df )
```
Calculate the mean (average) of a linear variable, overall and by groups:
```{r eval = FALSE , results = "hide" }
nppes_tbl %>%
	summarize( mean = mean( provider_enumeration_year , na.rm = TRUE ) )

nppes_tbl %>%
	group_by( provider_gender_code ) %>%
	summarize( mean = mean( provider_enumeration_year , na.rm = TRUE ) )
```

---

## Analysis Examples with `data.table` \ {-}

The R `data.table` library provides a high-performance version of base R's data.frame with syntax and feature enhancements for ease of use, convenience and programming speed. [data.table](https://r-datatable.com) offers concise syntax: fast to type, fast to read, fast speed, memory efficiency, a careful API lifecycle management, an active community, and a rich set of features. [This vignette](https://cran.r-project.org/web/packages/data.table/vignettes/datatable-intro.html) details the available features. As a starting point for NPPES users, this code replicates previously-presented examples:

```{r eval = FALSE , results = 'hide' }
library(data.table)
nppes_dt <- data.table( nppes_df )
```
Calculate the mean (average) of a linear variable, overall and by groups:
```{r eval = FALSE , results = 'hide' }
nppes_dt[ , mean( provider_enumeration_year , na.rm = TRUE ) ]

nppes_dt[ , mean( provider_enumeration_year , na.rm = TRUE ) , by = provider_gender_code ]
```

---

## Analysis Examples with `duckdb` \ {-}

The R `duckdb` library provides an embedded analytical data management system with support for the Structured Query Language (SQL). [duckdb](https://duckdb.org) offers a simple, feature-rich, fast, and free SQL OLAP management system. [This vignette](https://duckdb.org/docs/api/r) details the available features. As a starting point for NPPES users, this code replicates previously-presented examples:

```{r eval = FALSE , results = 'hide' }
library(duckdb)
con <- dbConnect( duckdb::duckdb() , dbdir = 'my-db.duckdb' )
dbWriteTable( con , 'nppes' , nppes_df )
```
Calculate the mean (average) of a linear variable, overall and by groups:
```{r eval = FALSE , results = 'hide' }
dbGetQuery( con , 'SELECT AVG( provider_enumeration_year ) FROM nppes' )

dbGetQuery(
	con ,
	'SELECT
		provider_gender_code ,
		AVG( provider_enumeration_year )
	FROM
		nppes
	GROUP BY
		provider_gender_code'
)
```