California Health Interview Survey (CHIS)

California’s National Health Interview Survey (NHIS), a healthcare survey for the nation’s largest state.

One adult, one teenage (12-17), and one child table, each with one row per sampled respondent.
A complex survey designed to generalize to the civilian non-institutionalized population of California.
Released annually since 2011, and biennially since 2001.
Administered by the UCLA Center for Health Policy Research.

Please skim before you begin:

CHIS 2021-2022 Methodology Report Series, Report 1: Sample Design DESIGN
CHIS 2021-2022 Methodology Report Series, Report 5: Weighting and Variance Estimation
A haiku regarding this microdata:

# strike gold, movie star
# play, third wish cali genie
# statewide health survey

Function Definitions

Define a function to unzip and import each Stata file:

library(haven)

chis_import <-
    function( this_fn ){
        
        these_files <- unzip( this_fn , exdir = tempdir() )

        stata_fn <- grep( "ADULT\\.|CHILD\\.|TEEN\\." , these_files , value = TRUE )

        this_tbl <- read_stata( stata_fn )

        this_df <- data.frame( this_tbl )

        names( this_df ) <- tolower( names( this_df ) )

        # remove labelled classes
        labelled_cols <- 
            sapply( this_df , function( w ) class( w )[1] == 'haven_labelled' )

        this_df[ labelled_cols ] <-
            sapply( this_df[ labelled_cols ] , as.numeric )

        this_df
    }

Download, Import, Preparation

Register at the UCLA Center for Health Policy Research at https://healthpolicy.ucla.edu/user/register.
Choose Year: 2022, Age Group: Adult and Teen and Child, File Type: Stata.
Download the 2022 Adult, Teen, and Child Stata files (version Oct 2023).

Import the adult, teen, and child stata tables into data.frame objects:

chis_adult_df <- 
    chis_import( file.path( path.expand( "~" ) , "adult_stata_2022.zip" ) )

chis_teen_df <- 
    chis_import( file.path( path.expand( "~" ) , "teen_stata_2022.zip" ) )

chis_child_df <- 
    chis_import( file.path( path.expand( "~" ) , "child_stata_2022.zip" ) )

Harmonize the general health condition variable across the three data.frame objects:

chis_adult_df[ , 'general_health' ] <-
    c( 1 , 2 , 3 , 4 , 4 )[ chis_adult_df[ , 'ab1' ] ]

chis_teen_df[ , 'general_health' ] <- chis_teen_df[ , 'tb1_p1' ]

chis_child_df[ , 'general_health' ] <-
    c( 1 , 2 , 3 , 4 , 4 )[ chis_child_df[ , 'ca6' ] ]

Add four age categories across the three data.frame objects:

chis_adult_df[ , 'age_categories' ] <-
    ifelse( chis_adult_df[ , 'srage_p1' ] >= 65 , 4 , 3 )

chis_teen_df[ , 'age_categories' ] <- 2

chis_child_df[ , 'age_categories' ] <- 1

Harmonize the usual source of care variable across the three data.frame objects:

chis_adult_df[ , 'no_usual_source_of_care' ] <-
    as.numeric( chis_adult_df[ , 'ah1v2' ] == 2 )

chis_teen_df[ , 'no_usual_source_of_care' ] <-
    as.numeric( chis_teen_df[ , 'tf1v2' ] == 2 )

chis_child_df[ , 'no_usual_source_of_care' ] <-
    as.numeric( chis_child_df[ , 'cd1v2' ] == 2 )

Add monthly fruit and vegetable counts to the adult data.frame object, blanking the other two:

chis_adult_df[ , 'adult_fruits_past_month' ] <- chis_adult_df[ , 'ae2' ]

chis_adult_df[ , 'adult_veggies_past_month' ] <- chis_adult_df[ , 'ae7' ]

chis_teen_df[ , c( 'adult_fruits_past_month' , 'adult_veggies_past_month' ) ] <- NA

chis_child_df[ , c( 'adult_fruits_past_month' , 'adult_veggies_past_month' ) ] <- NA

Specify which variables to keep in each of the data.frame objects, then stack them:

variables_to_keep <-
    c(
        grep( '^rakedw' , names( chis_adult_df ) , value = TRUE ) ,
        'general_health' , 'age_categories' , 'adult_fruits_past_month' , 
        'adult_veggies_past_month' , 'srsex' , 'povll2_p1v2' , 'no_usual_source_of_care'
    )

chis_df <- 
    rbind( 
        chis_child_df[ variables_to_keep ] , 
        chis_teen_df[ variables_to_keep ] , 
        chis_adult_df[ variables_to_keep ] 
    )

Save Locally

Save the object at any point:

# chis_fn <- file.path( path.expand( "~" ) , "CHIS" , "this_file.rds" )
# saveRDS( chis_df , file = chis_fn , compress = FALSE )

Load the same object:

# chis_df <- readRDS( chis_fn )

Survey Design Definition

Construct a complex sample survey design:

library(survey)

chis_design <-
    svrepdesign(
        data = chis_df , 
        weights = ~ rakedw0 , 
        repweights = "rakedw[1-9]" ,
        type = "other" , 
        scale = 1 ,
        rscales = 1 , 
        mse = TRUE
    )

Variable Recoding

Add new columns to the data set:

chis_design <- 
    update( 
        chis_design , 
        
        one = 1 ,
        
        gender = factor( srsex , levels = 1:2 , labels = c( 'male' , 'female' ) ) ,
        
        age_categories =
            factor( 
                age_categories , 
                levels = 1:4 , 
                labels = 
                    c( 'children under 12' , 'teens age 12-17' , 'adults age 18-64' , 'seniors' )
            ) ,
        
        general_health =
            factor(
                general_health ,
                levels = 1:4 ,
                labels = c( 'Excellent' , 'Very good' , 'Good' , 'Fair/Poor' )
            )
    )

Analysis Examples with the `survey` library

Unweighted Counts

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

sum( weights( chis_design , "sampling" ) != 0 )

svyby( ~ one , ~ general_health , chis_design , unwtd.count )

Weighted Counts

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

svytotal( ~ one , chis_design )

svyby( ~ one , ~ general_health , chis_design , svytotal )

Descriptive Statistics

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

svymean( ~ povll2_p1v2 , chis_design )

svyby( ~ povll2_p1v2 , ~ general_health , chis_design , svymean )

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

svymean( ~ gender , chis_design )

svyby( ~ gender , ~ general_health , chis_design , svymean )

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

svytotal( ~ povll2_p1v2 , chis_design )

svyby( ~ povll2_p1v2 , ~ general_health , chis_design , svytotal )

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

svytotal( ~ gender , chis_design )

svyby( ~ gender , ~ general_health , chis_design , svytotal )

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

svyquantile( ~ povll2_p1v2 , chis_design , 0.5 )

svyby( 
    ~ povll2_p1v2 , 
    ~ general_health , 
    chis_design , 
    svyquantile , 
    0.5 ,
    ci = TRUE 
)

Estimate a ratio:

svyratio( 
    numerator = ~ adult_fruits_past_month , 
    denominator = ~ adult_veggies_past_month , 
    chis_design ,
    na.rm = TRUE
)

Subsetting

Restrict the survey design to seniors:

sub_chis_design <- subset( chis_design , age_categories == 'seniors' )

Calculate the mean (average) of this subset:

svymean( ~ povll2_p1v2 , sub_chis_design )

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 <- svymean( ~ povll2_p1v2 , chis_design )

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

grouped_result <-
    svyby( 
        ~ povll2_p1v2 , 
        ~ general_health , 
        chis_design , 
        svymean 
    )
    
coef( grouped_result )
SE( grouped_result )
confint( grouped_result )
cv( grouped_result )

Calculate the degrees of freedom of any survey design object:

degf( chis_design )

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

svyvar( ~ povll2_p1v2 , chis_design )

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

# SRS without replacement
svymean( ~ povll2_p1v2 , chis_design , deff = TRUE )

# SRS with replacement
svymean( ~ povll2_p1v2 , chis_design , deff = "replace" )

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

svyciprop( ~ no_usual_source_of_care , chis_design ,
    method = "likelihood" )

Regression Models and Tests of Association

Perform a design-based t-test:

svyttest( povll2_p1v2 ~ no_usual_source_of_care , chis_design )

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

svychisq( 
    ~ no_usual_source_of_care + gender , 
    chis_design 
)

Perform a survey-weighted generalized linear model:

glm_result <- 
    svyglm( 
        povll2_p1v2 ~ no_usual_source_of_care + gender , 
        chis_design 
    )

summary( glm_result )

Replication Example

This matches the proportions and counts from AskCHIS. The standard errors do not match precisely, but the team at UCLA confirmed this survey design definition to be correct, and that the minor standard error and confidence interval differences should not impact any analyses from a statistical perspective:

chis_adult_design <-
    svrepdesign(
        data = chis_adult_df , 
        weights = ~ rakedw0 , 
        repweights = "rakedw[1-9]" ,
        type = "other" , 
        scale = 1 ,
        rscales = 1 , 
        mse = TRUE
    )
    
chis_adult_design <-
    update(
        chis_adult_design ,
        ab1 = 
            factor( 
                ab1 , 
                levels = 1:5 , 
                labels = c( 'Excellent' , 'Very good' , 'Good' , 'Fair' , 'Poor' )
            )
    )
    
this_proportion <- svymean( ~ ab1 , chis_adult_design )

stopifnot( round( coef( this_proportion ) , 3 ) == c( 0.183 , 0.340 , 0.309 , 0.139 , 0.029 ) )

this_count <- svytotal( ~ ab1 , chis_adult_design )

stopifnot( 
    round( coef( this_count ) , -3 ) == c( 5414000 , 10047000 , 9138000 , 4106000 , 855000 )
)

Analysis Examples with `srvyr`

The R srvyr library calculates summary statistics from survey data, such as the mean, total or quantile using dplyr-like syntax. 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 details the available features. As a starting point for CHIS users, this code replicates previously-presented examples:

library(srvyr)
chis_srvyr_design <- as_survey( chis_design )