nis.Rmd

# National Immunization Survey (NIS) {-}

[![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/nis/actions"><img src="https://github.com/asdfree/nis/actions/workflows/r.yml/badge.svg" alt="Github Actions Badge"></a> 

The vaccination coverage rate tracker for national, state, and selected local areas.

* One table with one row per sampled toddler.

* A complex sample survey designed to generalize to children aged 19-35 months in the United States.

* Released annually since 1995, plus an adolescent (13-17 years) sample since 2008.

* Administered by the [Centers for Disease Control and Prevention](http://www.cdc.gov/).

---

Please skim before you begin:

1. [About NIS](https://www.cdc.gov/vaccines/imz-managers/nis/about.html)

2. [National Immunization Survey-Child: A User's Guide for the 2021 Public-Use Data File](https://www.cdc.gov/vaccines/imz-managers/nis/downloads/NIS-PUF21-DUG.pdf)

3. A haiku regarding this microdata:

```{r}
# i hear babies cry
# protesting lungs of iron
# a wonderful world
```

---

## Download, Import, Preparation {-}

Download the fixed-width file:
```{r eval = FALSE , results = "hide" }
dat_tf <- tempfile()

dat_url <- "https://ftp.cdc.gov/pub/Vaccines_NIS/NISPUF21.DAT"

download.file( dat_url , dat_tf , mode = 'wb' )
```

Edit then execute the import script provided by the CDC:
```{r eval = FALSE , results = "hide" }
library(Hmisc)

r_tf <- tempfile()

r_script_url <- "https://ftp.cdc.gov/pub/Vaccines_NIS/NISPUF21.R"

r_input_lines <- readLines( r_script_url )

# do not let the script do the save()
r_input_lines <- gsub( "^save\\(" , "# save(" , r_input_lines )

# redirect the path to the flat file to the local save location of `dat_tf`
r_input_lines <- gsub( '\\"path\\-to\\-file\\/(.*)\\.DAT\\"' , "dat_tf" , r_input_lines )

# save the edited script locally
writeLines( r_input_lines , r_tf )

# run the edited script
source( r_tf , echo = TRUE )

# rename the resultant data.frame object
nis_df <- NISPUF21

names( nis_df ) <- tolower( names( nis_df ) )

nis_df[ , 'one' ] <- 1
```

### Save Locally \ {-}

Save the object at any point:

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

Load the same object:

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

### Survey Design Definition {-}
Construct a complex sample survey design:

```{r eval = FALSE , results = "hide" }
library(survey)

options( survey.lonely.psu = "adjust" )

nis_design <- 
	svydesign(
		id = ~ seqnumhh , 
		strata = ~ stratum , 
		weights = ~ provwt_c , 
		data = subset( nis_df , provwt_c > 0 ) 
	)
```

### Variable Recoding {-}

Add new columns to the data set:
```{r eval = FALSE , results = "hide" }
nis_design <- 
	
	update( 
		
		nis_design , 
		
		first_fed_formula =
			ifelse( bf_formr20 %in% 888 , NA , bf_formr20 ) ,
		
		dtap_3p =

			as.numeric(

				( p_numdah >= 3 ) |
				( p_numdhi >= 3 ) |
				( p_numdih >= 3 ) |
				( p_numdta >= 3 ) |
				( p_numdtp >= 3 )

			) ,
		
		dtap_4p =

			as.numeric(

				( p_numdah >= 4 ) |
				( p_numdhi >= 4 ) |
				( p_numdih >= 4 ) |
				( p_numdta >= 4 ) |
				( p_numdtp >= 4 )

			)
			
	)
```

---

## Analysis Examples with the `survey` library \ {-}

### Unweighted Counts {-}

Count the unweighted number of records in the survey sample, overall and by groups:
```{r eval = FALSE , results = "hide" }
sum( weights( nis_design , "sampling" ) != 0 )

svyby( ~ one , ~ state , nis_design , unwtd.count )
```

### Weighted Counts {-}
Count the weighted size of the generalizable population, overall and by groups:
```{r eval = FALSE , results = "hide" }
svytotal( ~ one , nis_design )

svyby( ~ one , ~ state , nis_design , svytotal )
```

### Descriptive Statistics {-}

Calculate the mean (average) of a linear variable, overall and by groups:
```{r eval = FALSE , results = "hide" }
svymean( ~ first_fed_formula , nis_design , na.rm = TRUE )

svyby( ~ first_fed_formula , ~ state , nis_design , svymean , na.rm = TRUE )
```

Calculate the distribution of a categorical variable, overall and by groups:
```{r eval = FALSE , results = "hide" }
svymean( ~ sex , nis_design , na.rm = TRUE )

svyby( ~ sex , ~ state , nis_design , svymean , na.rm = TRUE )
```

Calculate the sum of a linear variable, overall and by groups:
```{r eval = FALSE , results = "hide" }
svytotal( ~ first_fed_formula , nis_design , na.rm = TRUE )

svyby( ~ first_fed_formula , ~ state , nis_design , svytotal , na.rm = TRUE )
```

Calculate the weighted sum of a categorical variable, overall and by groups:
```{r eval = FALSE , results = "hide" }
svytotal( ~ sex , nis_design , na.rm = TRUE )

svyby( ~ sex , ~ state , nis_design , svytotal , na.rm = TRUE )
```

Calculate the median (50th percentile) of a linear variable, overall and by groups:
```{r eval = FALSE , results = "hide" }
svyquantile( ~ first_fed_formula , nis_design , 0.5 , na.rm = TRUE )

svyby( 
	~ first_fed_formula , 
	~ state , 
	nis_design , 
	svyquantile , 
	0.5 ,
	ci = TRUE , na.rm = TRUE
)
```

Estimate a ratio:
```{r eval = FALSE , results = "hide" }
svyratio( 
	numerator = ~ bf_exclr06 , 
	denominator = ~ bf_endr06 , 
	nis_design ,
	na.rm = TRUE
)
```

### Subsetting {-}

Restrict the survey design to toddlers up to date on polio shots:
```{r eval = FALSE , results = "hide" }
sub_nis_design <- subset( nis_design , p_utdpol == 1 )
```
Calculate the mean (average) of this subset:
```{r eval = FALSE , results = "hide" }
svymean( ~ first_fed_formula , sub_nis_design , na.rm = TRUE )
```

### Measures of Uncertainty {-}

Extract the coefficient, standard error, confidence interval, and coefficient of variation from any descriptive statistics function result, overall and by groups:
```{r eval = FALSE , results = "hide" }
this_result <- svymean( ~ first_fed_formula , nis_design , na.rm = TRUE )

coef( this_result )
SE( this_result )
confint( this_result )
cv( this_result )

grouped_result <-
	svyby( 
		~ first_fed_formula , 
		~ state , 
		nis_design , 
		svymean ,
		na.rm = TRUE 
	)
	
coef( grouped_result )
SE( grouped_result )
confint( grouped_result )
cv( grouped_result )
```

Calculate the degrees of freedom of any survey design object:
```{r eval = FALSE , results = "hide" }
degf( nis_design )
```

Calculate the complex sample survey-adjusted variance of any statistic:
```{r eval = FALSE , results = "hide" }
svyvar( ~ first_fed_formula , nis_design , na.rm = TRUE )
```

Include the complex sample design effect in the result for a specific statistic:
```{r eval = FALSE , results = "hide" }
# SRS without replacement
svymean( ~ first_fed_formula , nis_design , na.rm = TRUE , deff = TRUE )

# SRS with replacement
svymean( ~ first_fed_formula , nis_design , na.rm = TRUE , deff = "replace" )
```

Compute confidence intervals for proportions using methods that may be more accurate near 0 and 1. See `?svyciprop` for alternatives:
```{r eval = FALSE , results = "hide" }
svyciprop( ~ dtap_3p , nis_design ,
	method = "likelihood" )
```

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

Perform a design-based t-test:
```{r eval = FALSE , results = "hide" }
svyttest( first_fed_formula ~ dtap_3p , nis_design )
```

Perform a chi-squared test of association for survey data:
```{r eval = FALSE , results = "hide" }
svychisq( 
	~ dtap_3p + sex , 
	nis_design 
)
```

Perform a survey-weighted generalized linear model:
```{r eval = FALSE , results = "hide" }
glm_result <- 
	svyglm( 
		first_fed_formula ~ dtap_3p + sex , 
		nis_design 
	)

summary( glm_result )
```

---

## Replication Example {-}

This example matches the statistics and standard errors from [Data User's Guide Table 4](https://www.cdc.gov/vaccines/imz-managers/nis/downloads/NIS-PUF21-DUG.pdf#page=35):

```{r eval = FALSE , results = "hide" }

results <-
	svyby( 
		~ p_utd431h314_rout_s , 
		~ raceethk , 
		nis_design , 
		svymean
	)

coefficients <- results[ , "p_utd431h314_rout_sUTD" , drop = FALSE ]

standard_errors <- results[ , "se.p_utd431h314_rout_sUTD" , drop = FALSE ]

stopifnot( round( coefficients[ "HISPANIC" , ] , 3 ) == .711 )
stopifnot( round( coefficients[ "NON-HISPANIC WHITE ONLY" , ] , 3 ) == .742 )
stopifnot( round( coefficients[ "NON-HISPANIC BLACK ONLY" , ] , 3 ) == .647 )
stopifnot( round( standard_errors[ "HISPANIC" , ] , 3 ) == .015 )
stopifnot( round( standard_errors[ "NON-HISPANIC WHITE ONLY" , ] , 3 ) == .009 )
stopifnot( round( standard_errors[ "NON-HISPANIC BLACK ONLY" , ] , 3 ) == .022 )
```

---

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

The R `srvyr` library calculates summary statistics from survey data, such as the mean, total or quantile using [dplyr](https://github.com/tidyverse/dplyr/)-like syntax. [srvyr](https://github.com/gergness/srvyr) allows for the use of many verbs, such as `summarize`, `group_by`, and `mutate`, the convenience of pipe-able functions, the `tidyverse` style of non-standard evaluation and more consistent return types than the `survey` package. [This vignette](https://cran.r-project.org/web/packages/srvyr/vignettes/srvyr-vs-survey.html) details the available features. As a starting point for NIS users, this code replicates previously-presented examples:

```{r eval = FALSE , results = "hide" }
library(srvyr)
nis_srvyr_design <- as_survey( nis_design )
```
Calculate the mean (average) of a linear variable, overall and by groups:
```{r eval = FALSE , results = "hide" }
nis_srvyr_design %>%
	summarize( mean = survey_mean( first_fed_formula , na.rm = TRUE ) )

nis_srvyr_design %>%
	group_by( state ) %>%
	summarize( mean = survey_mean( first_fed_formula , na.rm = TRUE ) )
```