National Health Interview Survey (NHIS)

Build Status Build status

The National Health Interview Survey (NHIS) is America’s most detailed household survey of health status and medical experience.

  • A main table with one row for each person within each sampled household, mergeable other tables like the sample child table with a more detailed questionnaire for only one child (when available) within each sampled household.

  • A complex sample survey designed to generalize to the civilian non-institutionalized population of the United States.

  • Released annually since 1963, the most recent major re-design in 1997.

  • Administered by the Centers for Disease Control and Prevention.

Simplified Download and Importation

The R lodown package easily downloads and imports all available NHIS microdata by simply specifying "nhis" with an output_dir = parameter in the lodown() function. Depending on your internet connection and computer processing speed, you might prefer to run this step overnight.

library(lodown)
lodown( "nhis" , output_dir = file.path( path.expand( "~" ) , "NHIS" ) )

lodown also provides a catalog of available microdata extracts with the get_catalog() function. After requesting the NHIS catalog, you could pass a subsetted catalog through the lodown() function in order to download and import specific extracts (rather than all available extracts).

library(lodown)
# examine all available NHIS microdata files
nhis_cat <-
    get_catalog( "nhis" ,
        output_dir = file.path( path.expand( "~" ) , "NHIS" ) )

# 2016 only
nhis_cat <- subset( nhis_cat , year == 2016 )
# download the microdata to your local computer
lodown( "nhis" , nhis_cat )

Analysis Examples with the survey library

Construct a multiply-imputed, complex sample survey design:

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

library(survey)
library(mitools)

nhis_personsx_df <- 
    readRDS( file.path( path.expand( "~" ) , "NHIS" , "2016/personsx.rds" ) )

nhis_income_list <- 
    readRDS( file.path( path.expand( "~" ) , "NHIS" , "2016/incmimp.rds" ) )

merge_variables <- c( "hhx" , "fmx" , "fpx" )

nhis_personsx_df[ merge_variables ] <- 
    sapply( nhis_personsx_df[ merge_variables ] , as.numeric )

inc_vars_to_keep <- 
    c( 
        merge_variables , 
        setdiff( 
            names( nhis_income_list[[ 1 ]] ) , 
            names( nhis_personsx_df )
        )
    )

# personsx variables to keep
vars_to_keep <- 
    c( merge_variables , "ppsu" , "pstrat" , "wtfa" ,
        'phstat' , 'sex' , 'hospno' , 'age_p' , 'hinotmyr' , 'notcov' )

nhis_personsx_df <- nhis_personsx_df[ vars_to_keep ]
    
nhis_personsx_list <-
    lapply( nhis_income_list ,
        function( w ){
            w <- w[ inc_vars_to_keep ]
            w[ merge_variables ] <- sapply( w[ merge_variables ] , as.numeric )
            result <- merge( nhis_personsx_df , w )
            stopifnot( nrow( result ) == nrow( nhis_personsx_df ) )
            result
        } )

# personsx design       
nhis_design <- 
    svydesign( 
        id = ~ppsu , 
        strata = ~pstrat ,
        nest = TRUE ,
        weights = ~wtfa ,
        data = imputationList( nhis_personsx_list )
    )

rm( nhis_personsx_list ) ; gc()

nhis_samadult_df <- 
    readRDS( file.path( path.expand( "~" ) , "NHIS" , "2016/samadult.rds" ) )

nhis_samadult_df[ merge_variables ] <- 
    sapply( nhis_samadult_df[ merge_variables ] , as.numeric )

samadult_vars_to_keep <- 
    c( 
        merge_variables , 
        setdiff( 
            names( nhis_samadult_df ) , 
            names( nhis_personsx_df ) 
        ) 
    )

nhis_personsx_samadult_df <-
    merge( nhis_personsx_df , nhis_samadult_df[ samadult_vars_to_keep ] )

stopifnot( nrow( nhis_personsx_samadult_df ) == nrow( nhis_samadult_df ) )

rm( nhis_personsx_df , nhis_samadult_df ) ; gc()

nhis_samadult_list <-
    lapply( nhis_income_list ,
        function( w ){
            w <- w[ inc_vars_to_keep ]
            w[ merge_variables ] <- sapply( w[ merge_variables ] , as.numeric )
            result <- merge( nhis_personsx_samadult_df , w )
            stopifnot( nrow( result ) == nrow( nhis_personsx_samadult_df ) )
            result
        } )

rm( nhis_income_list , nhis_personsx_samadult_df ) ; gc()

# sample adult design (commented out)
# nhis_samadult_design <- 
    # svydesign( 
        # id = ~ppsu , 
        # strata = ~pstrat ,
        # nest = TRUE ,
        # weights = ~wtfa_sa ,
        # data = imputationList( nhis_samadult_list )
    # )
    
rm( nhis_samadult_list ) ; gc()

Variable Recoding

Add new columns to the data set:

nhis_design <- 
    update( 
        nhis_design , 
        
        one = 1 ,
        
        poverty_category =
            factor( 
                findInterval( povrati3 , 1:4 ) ,
                labels = 
                    c( "below poverty" , "100-199%" , "200-299%" , "300-399%" , "400%+" )
            ) ,
            
        fair_or_poor_reported_health = 
            ifelse( phstat %in% 1:5 , as.numeric( phstat >= 4 ) , NA ) ,
            
        sex = factor( sex , labels = c( "male" , "female" ) ) ,
        
        hospno = ifelse( hospno > 366 , NA , hospno )

    )

Unweighted Counts

Count the unweighted number of records in the survey sample, overall and by groups:

MIcombine( with( nhis_design , svyby( ~ one , ~ one , unwtd.count ) ) )

MIcombine( with( nhis_design , svyby( ~ one , ~ poverty_category , unwtd.count ) ) )

Weighted Counts

Count the weighted size of the generalizable population, overall and by groups:

MIcombine( with( nhis_design , svytotal( ~ one ) ) )

MIcombine( with( nhis_design ,
    svyby( ~ one , ~ poverty_category , svytotal )
) )

Descriptive Statistics

Calculate the mean (average) of a linear variable, overall and by groups:

MIcombine( with( nhis_design , svymean( ~ age_p ) ) )

MIcombine( with( nhis_design ,
    svyby( ~ age_p , ~ poverty_category , svymean )
) )

Calculate the distribution of a categorical variable, overall and by groups:

MIcombine( with( nhis_design , svymean( ~ sex ) ) )

MIcombine( with( nhis_design ,
    svyby( ~ sex , ~ poverty_category , svymean )
) )

Calculate the sum of a linear variable, overall and by groups:

MIcombine( with( nhis_design , svytotal( ~ age_p ) ) )

MIcombine( with( nhis_design ,
    svyby( ~ age_p , ~ poverty_category , svytotal )
) )

Calculate the weighted sum of a categorical variable, overall and by groups:

MIcombine( with( nhis_design , svytotal( ~ sex ) ) )

MIcombine( with( nhis_design ,
    svyby( ~ sex , ~ poverty_category , svytotal )
) )

Calculate the median (50th percentile) of a linear variable, overall and by groups:

MIcombine( with( nhis_design , svyquantile( ~ age_p , 0.5 , se = TRUE ) ) )

MIcombine( with( nhis_design ,
    svyby( 
        ~ age_p , ~ poverty_category , svyquantile , 0.5 ,
        se = TRUE , keep.var = TRUE , ci = TRUE 
) ) )

Estimate a ratio:

MIcombine( with( nhis_design ,
    svyratio( numerator = ~ hinotmyr , denominator = ~ hospno , na.rm = TRUE )
) )

Subsetting

Restrict the survey design to uninsured:

sub_nhis_design <- subset( nhis_design , notcov == 1 )

Calculate the mean (average) of this subset:

MIcombine( with( sub_nhis_design , svymean( ~ age_p ) ) )

Measures of Uncertainty

Extract the coefficient, standard error, confidence interval, and coefficient of variation from any descriptive statistics function result, overall and by groups:

this_result <-
    MIcombine( with( nhis_design ,
        svymean( ~ age_p )
    ) )

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

grouped_result <-
    MIcombine( with( nhis_design ,
        svyby( ~ age_p , ~ poverty_category , svymean )
    ) )

coef( grouped_result )
SE( grouped_result )
confint( grouped_result )
cv( grouped_result )

Calculate the degrees of freedom of any survey design object:

degf( nhis_design$designs[[1]] )

Calculate the complex sample survey-adjusted variance of any statistic:

MIcombine( with( nhis_design , svyvar( ~ age_p ) ) )

Include the complex sample design effect in the result for a specific statistic:

# SRS without replacement
MIcombine( with( nhis_design ,
    svymean( ~ age_p , deff = TRUE )
) )

# SRS with replacement
MIcombine( with( nhis_design ,
    svymean( ~ age_p , deff = "replace" )
) )

Compute confidence intervals for proportions using methods that may be more accurate near 0 and 1. See ?svyciprop for alternatives:

lodown:::MIsvyciprop( ~ fair_or_poor_reported_health , nhis_design ,
    method = "likelihood" , na.rm = TRUE )

Regression Models and Tests of Association

Perform a design-based t-test:

lodown:::MIsvyttest( age_p ~ fair_or_poor_reported_health , nhis_design )

Perform a chi-squared test of association for survey data:

lodown:::MIsvychisq( ~ fair_or_poor_reported_health + sex , nhis_design )

Perform a survey-weighted generalized linear model:

glm_result <- 
    MIcombine( with( nhis_design ,
        svyglm( age_p ~ fair_or_poor_reported_health + sex )
    ) )
    
summary( glm_result )

Replication Example