Case study: length of hospital stay

Part 1: EDA

The data liggetid was collected at the Geriatric Department at Ullevål Sykehus. This dataset has been used for teaching at University of Oslo, MF9130E. The course material can be accessed here.

We will focus on the following variables:

• admission_year, year of hospital admission
• age
• sex
• admission_from, where 1 = home, 2 = Div. of Medicine, 3 = Div. of Surgery, 4 = Other division, 5 = Other hospital, 6 = Nursing home
• stroke, where 1 = yes, 0 = no
• los, hospital stay, in days

los <- readRDS('data/los.rds')
head(los, 3)

  admission_year age    sex admission_from stroke los
1           1987  81 female           home      0  13
2           1987  96 female           home      0  17
3           1987  79 female           home      0   6

Some variables need to be recoded and put into factor to display nicely.

# library(dplyr)
library(ggplot2)

# remove NA
los <- dplyr::filter(los, !is.na(sex) & !is.na(stroke) & !is.na(admission_from))

# code admission from with text
# unique(los$admission_from)
los$admission_from <- factor(los$admission_from, 
                            levels = c('home', 'div_surgery', 
                                       'div_medicine', 'div_other', 
                                       'other_hospital', 'nursing_home'), 
                            labels = c('home', 'div_surgery', 
                                       'div_medicine', 'div_other', 
                                       'other_hospital', 'nursing_home'))

# code admission year with text
los$admission_year <- factor(los$admission_year,
                            levels = c(1981:1987),
                            labels = as.character(1981:1987))

los$stroke <- factor(los$stroke, 
                          levels = c(0, 1), 
                          labels = c('no','yes'))

Visualization

LOS vs age and sex

plt_scat2 <- ggplot(data = los, 
                    mapping = aes(x = age, y = los, shape = sex, color = sex))
plt_scat2 <- plt_scat2 + geom_point(size = 2, alpha = 0.7)
# customize
plt_scat2 <- plt_scat2 + labs(
  x = 'Age', 
  y = 'Length of hosptial stay (days)', 
  title = 'Length of stay versus age'
)
plt_scat2 <- plt_scat2 + theme_bw() # make white background
# change text size
plt_scat2 <- plt_scat2 + theme(
  axis.text = element_text(size = 12),
  axis.title = element_text(size = 12), 
  plot.title = element_text(size = 15)
)
# change color
plt_scat2 <- plt_scat2 + scale_color_brewer(palette = 'Set1')
plt_scat2

library(ggridges)
plt_ridge <- ggplot(data = los, 
                    mapping = aes(x = los, y = admission_year, fill = sex))
plt_ridge <- plt_ridge + geom_density_ridges(alpha = 0.6) 
plt_ridge <- plt_ridge + theme_ridges()
plt_ridge <- plt_ridge + labs(
  x = 'Length of hosptial stay (days)', 
  y = 'Admission year', 
  title = 'Length of stay in each year, for each gender'
)
# change color
plt_ridge <- plt_ridge + scale_fill_brewer(palette = 'Set1')
plt_ridge

Picking joint bandwidth of 32.7

LOS vs year of admission

plt_box <- ggplot(data = los, 
                  mapping = aes(x = admission_year, y = los, fill = sex))
plt_box <- plt_box + geom_boxplot(outlier.size = 1)
# plt_box <- plt_box + facet_wrap( ~ sex)
plt_box <- plt_box + coord_flip()

# customize
plt_box <- plt_box + theme_bw() # make white background
plt_box <- plt_box + labs(
  x = 'Admission year', 
  y = 'Length of hosptial stay (days)', 
  title = 'Length of stay in each year, both men and women'
)
plt_box <- plt_box + theme(
  axis.text = element_text(size = 12),
  axis.title = element_text(size = 12), 
  plot.title = element_text(size = 15), 
  strip.text = element_text(size = 12)
)

plt_box <- plt_box + scale_fill_brewer(palette = 'Set1')
plt_box 

LOS vs types of admission

plt_box2 <- ggplot(data = los, 
                   mapping = aes(x = admission_year, y = los, fill = sex))
plt_box2 <- plt_box2 + geom_boxplot(outlier.size = 0.8)
plt_box2 <- plt_box2 + facet_wrap( ~ admission_from)


# customize
plt_box2 <- plt_box2 + theme_bw() # make white background
plt_box2 <- plt_box2 + labs(
  x = 'Admission year', 
  y = 'Length of hosptial stay (days)', 
  title = 'Length of stay in each year, each type of admission'
)
plt_box2 <- plt_box2 + theme(
  axis.text = element_text(size = 11),
  axis.title = element_text(size = 12), 
  plot.title = element_text(size = 15), 
  strip.text = element_text(size = 12), 
  axis.text.x = element_text(angle = 45) # more readable
)

plt_box2 <- plt_box2 + scale_fill_brewer(palette = 'Set1')
plt_box2