brfss_project

<!DOCTYPE html>

brfss_project

<!DOCTYPE html>

BRFSS Project

knitr::opts_chunk$set(echo = FALSE)

Setup

library(dplyr)
library(ggplot2)

setwd("~/Desktop/BRFSS_project")
load("4tiY2fqCQa-YmNn6gnGvzQ_1e7320c30a6f4b27894a54e2de50a805_brfss2013.RData")

set.seed(123)
brfss <- sample_n(brfss2013, 50000)

Data and Scope of Inference

The Behavioral Risk Factor Surveillance System (BRFSS) is a large cross-sectional health survey conducted annually by the Centers for Disease Control and Prevention (CDC) in the United States. Data are collected through telephone interviews from non-institutionalized adults aged 18 years or older. Because the data are observational and self-reported, results show association rather than causation, but due to the large random sample they can be generalized to the adult U.S. population.

Research Question 1

Is sleep duration related to poor mental health days across employment groups?

rq1 <- brfss %>%
  transmute(
    sleep = sleptim1,
    menthlth = menthlth,
    employ = employ1
  ) %>%
  filter(!is.na(sleep), !is.na(menthlth), !is.na(employ),
         sleep >= 0, sleep <= 24,
         menthlth >= 0, menthlth <= 30)

rq1 %>% summarise(
  n = n(),
  mean_sleep = mean(sleep),
  mean_menthlth = mean(menthlth)
)

##       n mean_sleep mean_menthlth
## 1 48160   7.050208      3.351184

ggplot(rq1, aes(x = sleep, y = menthlth)) +
  geom_point(alpha = 0.2) +
  geom_smooth(method = "lm", se = FALSE) +
  facet_wrap(~ employ) +
  labs(x = "Hours of sleep", y = "Poor mental health days")

## `geom_smooth()` using formula = 'y ~ x'

Findings: Less sleep is associated with more poor mental health days across employment categories.

Research Question 2

Is BMI different between smokers and non-smokers, and does this difference vary by sex?

rq2 <- brfss %>%
  transmute(
    bmi = as.numeric(X_bmi5)/100,
    smoker = smoke100,
    sex = sex
  ) %>%
  filter(!is.na(bmi), bmi >= 10, bmi <= 80,
         !is.na(smoker), !is.na(sex))

rq2 %>%
  group_by(sex, smoker) %>%
  summarise(mean_bmi = mean(bmi), .groups="drop")

## # A tibble: 4 × 3
##   sex    smoker mean_bmi
##   <fct>  <fct>     <dbl>
## 1 Male   Yes        28.1
## 2 Male   No         28.1
## 3 Female Yes        27.7
## 4 Female No         27.5

ggplot(rq2, aes(x = smoker, y = bmi)) +
  geom_boxplot() +
  facet_wrap(~ sex) +
  theme_minimal()

Findings: BMI distributions differ slightly between smokers and non-smokers and vary by sex.

Research Question 3

Is physical activity associated with general health across age groups?

rq3 <- brfss %>%
  transmute(
    exercise = exerany2,
    genhlth = genhlth,
    agegrp = X_ageg5yr
  ) %>%
  filter(!is.na(exercise), !is.na(genhlth), !is.na(agegrp))

tab3 <- rq3 %>%
  count(agegrp, exercise, genhlth) %>%
  group_by(agegrp, exercise) %>%
  mutate(p = n/sum(n)) %>%
  ungroup()

ggplot(tab3, aes(x = genhlth, y = p)) +
  geom_col() +
  facet_grid(agegrp ~ exercise) +
  theme(axis.text.x = element_text(angle = 45, hjust = 1))

Findings: Individuals who exercise generally report better health across all age groups.