Health Data from PNAD Continua with healthbR

Overview

The PNAD Continua (Pesquisa Nacional por Amostra de Domicilios Continua) is IBGE’s continuous household sample survey. While its core questionnaire focuses on labor and income, it includes supplementary health-related modules that are applied in specific quarters.

The healthbR package provides access to 4 health-related supplementary modules:

Module Description Years
deficiencia Persons with disabilities 2019, 2022, 2024
habitacao Housing characteristics (sanitation, water) 2012–2024
moradores General resident characteristics 2012–2024
aps Primary health care access 2022 (Q2)

Getting started

library(healthbR)
library(dplyr)

Module information

pnadc_info()

Available modules and years

pnadc_modules()
#> # A tibble: 4 x 4
#>   module       name_pt                            name_en                years
#>   <chr>        <chr>                              <chr>                  <list>
#> 1 deficiencia  Pessoas com deficiencia             Persons with disab... <int>
#> 2 habitacao    Caracteristicas dos domicilios      Housing character...  <int>
#> 3 moradores    Caracteristicas gerais dos morad... General character...  <int>
#> 4 aps          Atencao primaria a saude            Primary health care   <int>

# Years for a specific module
pnadc_years("deficiencia")
#> [1] 2019 2022 2024

Downloading data

Basic download

# Disability module, 2022
df <- pnadc_data(module = "deficiencia", year = 2022)

Select variables

df <- pnadc_data(
  module = "deficiencia",
  year = 2022,
  vars = c("UF", "V2007", "V2009", "V2010", "G001", "G003", "G006")
)

Multiple years

# Housing conditions across all available years
df <- pnadc_data(module = "habitacao")

Exploring variables

# List variable names
pnadc_variables(module = "deficiencia", year = 2022)

# Full dictionary with positions and widths
pnadc_dictionaries(module = "deficiencia", year = 2022)

Survey design

PNAD Continua uses a complex sample design. Survey design variables (UPA, Estrato, V1028) are always included in the output. Use as_survey = TRUE to create a survey design object:

# Requires srvyr package
svy <- pnadc_data(
  module = "deficiencia",
  year = 2022,
  as_survey = TRUE
)

# Use srvyr verbs for proper variance estimation
library(srvyr)
svy |>
  group_by(UF) |>
  survey_mean(G001 == "1", na.rm = TRUE)  # disability prevalence by state

Example: Disability prevalence

df <- pnadc_data(module = "deficiencia", year = 2022)

# G001: "Tem dificuldade permanente de enxergar" (vision difficulty)
# 1 = Sim, nao consegue de modo algum (cannot at all)
# 2 = Sim, muita dificuldade (great difficulty)
# 3 = Sim, alguma dificuldade (some difficulty)
# 4 = Nao, nenhuma dificuldade (no difficulty)
vision <- df |>
  filter(G001 %in% c("1", "2", "3", "4")) |>
  count(G001) |>
  mutate(pct = n / sum(n) * 100)

Example: Housing sanitation over time

df <- pnadc_data(module = "habitacao", year = c(2016, 2019, 2022))

# Analyze water supply and sanitation trends
# Variables vary by year -- check pnadc_variables() for each edition

Example: Primary health care access

# APS module only available for 2022 Q2
df <- pnadc_data(module = "aps", year = 2022)

Cache and performance

# Check cache
pnadc_cache_status()

# Clear cache
pnadc_clear_cache()

# Lazy evaluation
lazy_df <- pnadc_data(
  module = "deficiencia",
  year = 2022,
  lazy = TRUE,
  backend = "arrow"
)

Further reading