Exercises

Data

Hypoxia MAP Treatment Dataset

The Hypoxia MAP dataset was contributed by Dr. Amy Nowacki, Associate Professor, Cleveland Clinic. Please refer to this resource as: Amy S. Nowacki, “Hypoxia MAP Treatment Dataset”, TSHS Resources Portal (2022). Available at causeweb.org/tshs/hypoxia. The data is licensed under a Creative Commons Attribution-Non Commerical-Share Alike 4.0 International (CC BY-NC-SA 4.0) license, and is intended for educational purposes only.

Download CSV: hypoxia.csv

Download Excel: hypoxia.xlsx

Note:

  • Female
    • 0 = male
    • 1 = female
  • Race
    • 1 = African American
    • 2 = Caucasian
    • 3 = Other
  • Type Surg
    • 1 = gastroenterostomy
    • 2 = gastric restrictive procedure
    • 3 = gastroplasty
    • 4 = removal of gastric restrictive device

Exercises

Exercise 1: Performing operations in R

  1. Create a vector called age (like the one below). Multiply each element by 5.
age <- c(15, 25, 32, 87, 12)
Solution 
age * 5
[1]  75 125 160 435  60
  1. What does the sqrt() function do? Apply the sqrt() function to each element of age.
Solution 
sqrt(age)
[1] 3.872983 5.000000 5.656854 9.327379 3.464102
  1. Install the {ggplot2} package.
Solution 
install.packages("ggplot2")
  1. Load the {ggplot2} package into R.
Solution

Exercise 2: Loading data into R

  1. Make an R project file (File –> New Project). Download either the CSV or Excel file of the Hypoxia data from the section below and save it into your R project folder. Load the data into R.
Solution

You can use read.csv():

hypoxia <- read.csv("data/hypoxia.csv")

But the read_csv() function from {readr} is a little bit nicer:

library(readr)
hypoxia <- read_csv("data/hypoxia.csv")

For Excel files

library(readxl)
hypoxia <- read_xlsx("data/hypoxia.xlsx")
  1. Inspect the data using View().
Solution 
View(hypoxia)
  1. How many rows and columns are in the data?
Solution 
nrow(hypoxia)
[1] 281
ncol(hypoxia)
[1] 36
  1. Create a summary of the data.
Solution 
summary(hypoxia)
      Age            Female            Race            BMI       
 Min.   :16.40   Min.   :0.0000   Min.   :1.000   Min.   :23.40  
 1st Qu.:39.60   1st Qu.:0.0000   1st Qu.:2.000   1st Qu.:41.50  
 Median :48.20   Median :1.0000   Median :2.000   Median :46.00  
 Mean   :47.29   Mean   :0.7153   Mean   :1.826   Mean   :46.75  
 3rd Qu.:54.70   3rd Qu.:1.0000   3rd Qu.:2.000   3rd Qu.:51.70  
 Max.   :73.80   Max.   :1.0000   Max.   :3.000   Max.   :71.70  
                                                                 
   Sleeptime        Min Sao2          AHI           Smoking      
 Min.   : 0.00   Min.   :28.00   Min.   :1.000   Min.   :0.0000  
 1st Qu.: 1.40   1st Qu.:74.00   1st Qu.:2.000   1st Qu.:0.0000  
 Median : 8.10   Median :82.00   Median :3.000   Median :1.0000  
 Mean   :18.71   Mean   :78.71   Mean   :2.871   Mean   :0.5018  
 3rd Qu.:26.90   3rd Qu.:86.00   3rd Qu.:4.000   3rd Qu.:1.0000  
 Max.   :99.60   Max.   :95.00   Max.   :4.000   Max.   :1.0000  
                                 NA's   :3                       
    Diabetes          Hyper             CAD          Preop AntiHyper Med
 Min.   :0.0000   Min.   :0.0000   Min.   :0.00000   Min.   :0.0000     
 1st Qu.:0.0000   1st Qu.:0.0000   1st Qu.:0.00000   1st Qu.:0.0000     
 Median :0.0000   Median :1.0000   Median :0.00000   Median :0.0000     
 Mean   :0.3452   Mean   :0.7082   Mean   :0.09253   Mean   :0.2384     
 3rd Qu.:1.0000   3rd Qu.:1.0000   3rd Qu.:0.00000   3rd Qu.:0.0000     
 Max.   :1.0000   Max.   :1.0000   Max.   :1.00000   Max.   :1.0000     
                                                                        
      CPAP          Type Surg     Duration of Surg Duration of Surg1
 Min.   :0.0000   Min.   :1.000   Min.   :2.000    Min.   :0.2000   
 1st Qu.:0.0000   1st Qu.:1.000   1st Qu.:3.700    1st Qu.:0.4000   
 Median :1.0000   Median :1.000   Median :4.200    Median :0.5000   
 Mean   :0.6335   Mean   :1.224   Mean   :4.315    Mean   :0.5131   
 3rd Qu.:1.0000   3rd Qu.:1.000   3rd Qu.:4.900    3rd Qu.:0.6000   
 Max.   :1.0000   Max.   :4.000   Max.   :7.900    Max.   :1.1000   
                                                   NA's   :6        
 Duration of Surg2    TWA MAP          TWA MAP1         TWA MAP2     
 Min.   :1.400     Min.   : 66.00   Min.   : 51.96   Min.   : 66.17  
 1st Qu.:3.000     1st Qu.: 81.24   1st Qu.: 70.70   1st Qu.: 82.76  
 Median :3.500     Median : 88.55   Median : 78.02   Median : 89.73  
 Mean   :3.552     Mean   : 89.01   Mean   : 79.24   Mean   : 90.30  
 3rd Qu.:4.100     3rd Qu.: 95.94   3rd Qu.: 86.40   3rd Qu.: 98.16  
 Max.   :7.000     Max.   :127.20   Max.   :111.10   Max.   :132.71  
 NA's   :3                          NA's   :4                        
     TWA HR          TWA HR1          TWA HR2       Intraop AntiHyper Med
 Min.   : 54.22   Min.   : 50.32   Min.   : 53.61   Min.   :0.0000       
 1st Qu.: 68.89   1st Qu.: 66.28   1st Qu.: 68.52   1st Qu.:0.0000       
 Median : 75.38   Median : 73.98   Median : 75.89   Median :0.0000       
 Mean   : 76.10   Mean   : 75.58   Mean   : 76.14   Mean   :0.3096       
 3rd Qu.: 82.91   3rd Qu.: 83.48   3rd Qu.: 83.61   3rd Qu.:1.0000       
 Max.   :109.09   Max.   :124.59   Max.   :105.25   Max.   :1.0000       
 NA's   :78       NA's   :82       NA's   :78                            
  Vasopressor       Ephedrine      Ephedrine Amt     Epinephrine      
 Min.   :0.0000   Min.   :0.0000   Min.   : 0.000   Min.   :0.000000  
 1st Qu.:0.0000   1st Qu.:0.0000   1st Qu.: 0.000   1st Qu.:0.000000  
 Median :1.0000   Median :0.0000   Median : 0.000   Median :0.000000  
 Mean   :0.5445   Mean   :0.3452   Mean   : 5.196   Mean   :0.007117  
 3rd Qu.:1.0000   3rd Qu.:1.0000   3rd Qu.:10.000   3rd Qu.:0.000000  
 Max.   :1.0000   Max.   :1.0000   Max.   :50.000   Max.   :1.000000  
                                                                      
 Epinephrine Amt     Phenylephrine    Phenylephrine Amt      MAC       
 Min.   :0.000e+00   Min.   :0.0000   Min.   :0.0000    Min.   :0.510  
 1st Qu.:0.000e+00   1st Qu.:0.0000   1st Qu.:0.0000    1st Qu.:2.080  
 Median :0.000e+00   Median :0.0000   Median :0.0000    Median :2.830  
 Mean   :7.117e-05   Mean   :0.3879   Mean   :0.1232    Mean   :2.841  
 3rd Qu.:0.000e+00   3rd Qu.:1.0000   3rd Qu.:0.1000    3rd Qu.:3.410  
 Max.   :1.000e-02   Max.   :1.0000   Max.   :2.5700    Max.   :6.560  
                                                        NA's   :1      
 Propofol Induction IV Morphine Eq     Crystalloids     Colloids     
 Min.   :  0.0      Min.   :   6.70   Min.   : 400   Min.   :   0.0  
 1st Qu.:200.0      1st Qu.:  25.00   1st Qu.:2400   1st Qu.:   0.0  
 Median :200.0      Median :  30.30   Median :3000   Median : 500.0  
 Mean   :206.9      Mean   :  43.43   Mean   :2994   Mean   : 482.4  
 3rd Qu.:250.0      3rd Qu.:  40.00   3rd Qu.:3500   3rd Qu.:1000.0  
 Max.   :400.0      Max.   :1101.10   Max.   :6950   Max.   :2000.0

Exercise 3: Plotting single variables

  1. Create a histogram of the age of all patients in the study. What does the bins argument in geom_histogram() do?
Solution

Load ggplot2:

ggplot(
  data = hypoxia,
  mapping = aes(x = Age)
) +
  geom_histogram()
`stat_bin()` using `bins = 30`. Pick better value with `binwidth`.

Choose a different number of bins with bins:

ggplot(
  data = hypoxia,
  mapping = aes(x = Age)
) +
  geom_histogram(bins = 20)

  1. Create a histogram of the age of all female patients in the study. Hint: remember that 0 = male and 1 = female in the Female column.
Solution 
library(dplyr)
# Filter data to only females
f_hypoxia <- hypoxia %>%
  filter(Female == 1)
# Plot histogram
ggplot(
  data = f_hypoxia,
  mapping = aes(x = Age)
) +
  geom_histogram()

  1. Create a bar chart of the number of people who had each surgery type. Hint: geom_bar().
Solution 
ggplot(
  data = hypoxia,
  mapping = aes(x = `Type Surg`)
) +
  geom_bar()

  1. Bonus: Edit the axis labels using the labs() function. Can you also add a title?
Solution 
ggplot(
  data = hypoxia,
  mapping = aes(x = `Type Surg`)
) +
  geom_bar() +
  labs(
    x = "Surgery Type", y = "Number of patients",
    title = "Number of patients by surgery type"
  )

Exercise 4: Reading help files in R

  1. The {ggplot2} package is often used for plotting in R. What does the geom_count() function do?

Hint: use the ? and try the examples!

Solution 
?geom_count

The geom_count() function creates a bubble plot where the size of the points is proportional to the number of observations at each unique x and y combination.

  1. What is the difference between geom_bar() and geom_col() ?
Solution 
?geom_bar

  • geom_bar() counts the number of observations of a categorical variable and puts the count on the y-axis (i.e. R does the counting for you)

  • geom_col() puts the values in a column on the y-axis (i.e. you do the counting)

  1. Does this code do what you expect? Can you fix it?
ggplot(
  data = hypoxia,
  mapping = aes(x = `Type Surg`, fill = "blue")
) +
  geom_bar()
Solution 
ggplot(
  data = hypoxia,
  mapping = aes(x = `Type Surg`)
) +
  geom_bar(fill = "blue")

  • the fill argument changes the colour inside the bars
  • the colour isn’t mapped to a column in the data so it shouldn’t be inside the aes() function.

Exercise 5: Plotting multiple variables

  1. Create a scatter plot of age (on the x-axis) and BMI (on the y-axis).
Solution 
ggplot(
  data = hypoxia,
  mapping = aes(x = Age, y = BMI)
) +
  geom_point()

  1. Create a boxplot of age for each surgery type. Hint: make `Type Surg` a factor().
Solution 
ggplot(
  data = hypoxia,
  mapping = aes(x = Age, y = factor(`Type Surg`))
) +
  geom_boxplot()

  1. Create a bar chart of the number of people who had each surgery type. Colour the bars based on whether people had diabetes. Hint: should Diabetes be a numeric or a factor?
Solution 
ggplot(
  data = hypoxia,
  mapping = aes(x = `Type Surg`, fill = factor(Diabetes))
) +
  geom_bar()

  1. Bonus: Edit your bar chart to put the bars next to each other instead of stacked on top. Hint: look at the position argument in geom_bar().
Solution 
ggplot(
  data = hypoxia,
  mapping = aes(x = `Type Surg`, fill = factor(Diabetes))
) +
  geom_bar(position = position_dodge())

Note that the bar for Type 3 surgery is twice as wide because there are no people with diabetes who had this surgery. To fix this:

ggplot(
  data = hypoxia,
  mapping = aes(x = `Type Surg`, fill = factor(Diabetes))
) +
  geom_bar(position = position_dodge(preserve = "single"))

Exercise 6: Computing summary statistics

  1. For each of the columns `Duration of Surg` and AHI, calculate the following summary statistics: mean and standard deviation.

If there are any missing values, calculate the mean of the non-missing values. Hint: Look at the na.rm argument for ?mean.

Solution

Using the summarise() function:

hypoxia %>%
  summarise(
    mean_dur = mean(`Duration of Surg`),
    mean_ahi = mean(AHI),
    sd_dur = sd(`Duration of Surg`),
    sd_ahi = sd(AHI),
  )
# A tibble: 1 × 4
  mean_dur mean_ahi sd_dur sd_ahi
     <dbl>    <dbl>  <dbl>  <dbl>
1     4.31       NA   1.12     NA

Exclude missing values from the calculation:

hypoxia %>%
  summarise(
    mean_dur = mean(`Duration of Surg`),
    mean_ahi = mean(AHI, na.rm = TRUE),
    sd_dur = sd(`Duration of Surg`),
    sd_ahi = sd(AHI, na.rm = TRUE),
  )
# A tibble: 1 × 4
  mean_dur mean_ahi sd_dur sd_ahi
     <dbl>    <dbl>  <dbl>  <dbl>
1     4.31     2.87   1.12   1.08
  1. Repeat the calculations, but group the summary statistics by Surgery Type (`Type Surg`).
Solution 
hypoxia %>%
  group_by(`Type Surg`) %>% 
  summarise(
    mean_dur = mean(`Duration of Surg`),
    mean_ahi = mean(AHI, na.rm = TRUE),
    sd_dur = sd(`Duration of Surg`),
    sd_ahi = sd(AHI, na.rm = TRUE),
  )
# A tibble: 4 × 5
  `Type Surg` mean_dur mean_ahi sd_dur sd_ahi
        <dbl>    <dbl>    <dbl>  <dbl>  <dbl>
1           1     4.61     2.92  0.928  1.04 
2           2     2.91     2.75  0.640  1.21 
3           3     4.57     1.67  0.808  0.577
4           4     6.7      2.5   1.41   2.12
  1. Bonus: also calculate the median, minimum, and maximum.
Solution 
hypoxia %>%
  group_by(`Type Surg`) %>% 
  summarise(
    mean_dur = mean(`Duration of Surg`),
    mean_ahi = mean(AHI, na.rm = TRUE),
    sd_dur = sd(`Duration of Surg`),
    sd_ahi = sd(AHI, na.rm = TRUE),
    median_dur = median(`Duration of Surg`),
    median_ahi = median(AHI, na.rm = TRUE),
    min_dur = min(`Duration of Surg`),
    min_ahi = min(AHI, na.rm = TRUE),
    max_dur = max(`Duration of Surg`),
    max_ahi = max(AHI, na.rm = TRUE),
  )
# A tibble: 4 × 11
  `Type Surg` mean_dur mean_ahi sd_dur sd_ahi median_dur median_ahi min_dur
        <dbl>    <dbl>    <dbl>  <dbl>  <dbl>      <dbl>      <dbl>   <dbl>
1           1     4.61     2.92  0.928  1.04         4.4        3       2.9
2           2     2.91     2.75  0.640  1.21         2.8        3       2  
3           3     4.57     1.67  0.808  0.577        4.1        2       4.1
4           4     6.7      2.5   1.41   2.12         6.7        2.5     5.7
# ℹ 3 more variables: min_ahi <dbl>, max_dur <dbl>, max_ahi <dbl>

Exercise 7: Summary tables

  1. Create a descriptive table of patient characteristics, which includes the following variables: age, gender, race, smoking.
Solution 
Warning: package 'gtsummary' was built under R version 4.4.1
tbl1_data <- hypoxia %>%
  select(Age, Female, Race, Smoking)
tbl1_data %>%
  tbl_summary()
Characteristic N = 2811
Age 48 (40, 55)
Female 201 (72%)
Race
    1 57 (20%)
    2 216 (77%)
    3 8 (2.8%)
Smoking 141 (50%)
1 Median (Q1, Q3); n (%)
  1. Are Female and Smoking represented in the table in a way that makes sense? Change the Female and Smoking columns to a factor. Hint: you might want to look at the mutate() function to edit the columns in the data.
Solution 
tbl1_data %>%
  mutate(
    Smoking = factor(Smoking),
    Female = factor(Female)
  ) %>%
  tbl_summary()
Characteristic N = 2811
Age 48 (40, 55)
Female
    0 80 (28%)
    1 201 (72%)
Race
    1 57 (20%)
    2 216 (77%)
    3 8 (2.8%)
Smoking
    0 140 (50%)
    1 141 (50%)
1 Median (Q1, Q3); n (%)
  1. Group the table by Smoking.

Hint: look at the by argument in tbl_summary().

Solution 
tbl1_data %>%
  mutate(
    Smoking = factor(Smoking),
    Female = factor(Female)
  ) %>%
  tbl_summary(by = Smoking)
Characteristic 0
N = 1401 		1
N = 1411
Age 45 (35, 53) 51 (44, 57)
Female

    0 35 (25%) 45 (32%)
    1 105 (75%) 96 (68%)
Race

    1 29 (21%) 28 (20%)
    2 106 (76%) 110 (78%)
    3 5 (3.6%) 3 (2.1%)
1 Median (Q1, Q3); n (%)
  1. Bonus: Change the labels for Smoking to Smoking and No smoking instead of 1 and 0.
Solution 
tbl1_data %>%
  mutate(
    Smoking = factor(Smoking,
      levels = c(0, 1),
      labels = c("No smoking", "Smoking")
    ),
    Female = factor(Female)
  ) %>%
  tbl_summary(Smoking)
Characteristic No smoking
N = 1401 		Smoking
N = 1411
Age 45 (35, 53) 51 (44, 57)
Female

    0 35 (25%) 45 (32%)
    1 105 (75%) 96 (68%)
Race

    1 29 (21%) 28 (20%)
    2 106 (76%) 110 (78%)
    3 5 (3.6%) 3 (2.1%)
1 Median (Q1, Q3); n (%)

Exercise 8: Statistical tests

  1. Perform a t-test to test whether the age of patients is significantly different for males and females. Assume the variances of the two groups are equal.
Solution 
# Subset data
f_hypoxia <- filter(hypoxia, Female == 1)
m_hypoxia <- filter(hypoxia, Female == 0)

# Unpaired t-test
t.test(f_hypoxia$Age, m_hypoxia$Age, var.equal = TRUE)

    Two Sample t-test

data:  f_hypoxia$Age and m_hypoxia$Age
t = -1.9334, df = 279, p-value = 0.0542
alternative hypothesis: true difference in means is not equal to 0
95 percent confidence interval:
 -5.70622115  0.05134553
sample estimates:
mean of x mean of y 
 46.48756  49.31500
  1. Test if the variances of the two groups are actually equal.
Solution 
var.test(f_hypoxia$Age, m_hypoxia$Age)

    F test to compare two variances

data:  f_hypoxia$Age and m_hypoxia$Age
F = 0.78536, num df = 200, denom df = 79, p-value = 0.1829
alternative hypothesis: true ratio of variances is not equal to 1
95 percent confidence interval:
 0.5341363 1.1205117
sample estimates:
ratio of variances 
         0.7853552