Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

updated with code for Adverse Events table #3

Open
wants to merge 1 commit into
base: master
Choose a base branch
from
Open

updated with code for Adverse Events table #3

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2,974 changes: 2,974 additions & 0 deletions README.html
View file
Open in desktop

Large diffs are not rendered by default.

175 changes: 174 additions & 1 deletion book/02-tables.Rmd
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,9 +1,182 @@
# (PART\*) Tables {-}
A common way to record data summaries in the medical literature is with tables. Two of the most common tables for clinical trial manuscripts are "Table 1", which is traditionally a description of the characteristics of the study population, and a table of adverse events, which is usually Table 2. <br>
We will walk through how to create Table 1 using the *tableby* function from the **arsenal** package, and then a table of Adverse Events using the *tabyl* function from the **janitor** package and then the **flextable** package to format this table. <br>
Let's get started by taking a quick look at our study data using the *head* function and the *glimpse* function. Run the code chunk below to read in our mockstudy data and then add two lines of code:
1. head(study) to show the first 6 observations
2. glimpse(study) to look at the structure of the dataset
You can also use View(study) to view the whole dataset in the top left pane.

```{r}
library(readr)
library(here)
library(tidyverse)
study <- read_csv(here("data/mockstudy2.csv")) %>% select(-X1)
study %<>%
mutate(low_wbc= factor(low_wbc)) %>%
mutate(neuropathy= factor(neuropathy)) %>%
mutate(diarrhea= factor(diarrhea)) %>%
mutate(vomiting= factor(vomiting)) %>%
mutate(blood_clot= factor(blood_clot))

```



# Table one

# Markdown tables

# kable tables

# flextable
# Table 2: Adverse Events with janitor::tabyl and flextable
In our mock study data, we have five adverse events reported, which are listed as the last 5 variables in the dataset, from low_wbc to blood_clot.

First we will load the janitor and arsenal libraries, then use some piping to select the adverse events columns.
Run the chunk below. All of the adverse events data are coded as 0 (did not happen) or 1 (did occur in this patient).

```{r}
library(janitor)
library(arsenal)
study %>%
select(case, arm, low_wbc:blood_clot)
```

In the next chunk we will use janitor::tabyl to count the cases and percentages for leukopenia.
We will use the functions to "adorn" basic tables with percentages, and percent formatting, and Ns, then save this as an object named low_wbc_tab.

Fiddle with the arguments to adorn_percentages
and adorn_pct_formatting.
use help() to learn the arguments. Experiment with affix_sign and rounding
Try commenting out different lines with a hashtag #.
Figure out what each step of the pipe does.


```{r}
study %>%
select(case, arm, low_wbc:blood_clot)%>%
tabyl(low_wbc, arm) %>%
adorn_percentages("col") %>%
adorn_pct_formatting(digits = 1) %>%
adorn_ns() ->
low_wbc_tab
```

Once you are happy with your small table, repeat the formatting for the other 4 adverse events (edit the code below), and run this code chunk.
```{r}
study %>%
select(case, arm, low_wbc:blood_clot) %>%
tabyl(neuropathy, arm) %>%
adorn_percentages("col") %>%
adorn_pct_formatting(digits = 1) %>%
adorn_ns() ->
neuropathy_tab

study %>%
select(case, arm, low_wbc:blood_clot) %>%
tabyl(diarrhea, arm) %>%
adorn_percentages("col") %>%
adorn_pct_formatting(digits = 1) %>%
adorn_ns() ->
diarrhea_tab

study %>%
select(case, arm, low_wbc:blood_clot) %>%
tabyl(vomiting, arm) %>%
adorn_percentages("col") %>%
adorn_pct_formatting(digits = 1) %>%
adorn_ns() ->
vomiting_tab

study %>%
select(case, arm, low_wbc:blood_clot) %>%
tabyl(blood_clot, arm) %>%
adorn_percentages("col") %>%
adorn_pct_formatting(digits = 1) %>%
adorn_ns() ->
blood_clot_tab
```

This is great, but now you have 5 distinct tables, and you would like to put them into one table.
So bind them together with bind_rows in the chunk below
```{r}
bind_rows(low_wbc_tab, neuropathy_tab, diarrhea_tab, vomiting_tab, blood_clot_tab) %>%
select(`A: IFL`, `F: FOLFOX`, `G: IROX`, low_wbc, neuropathy, diarrhea,
vomiting, blood_clot) ->
ae_table1
ae_table1
```
This is OK, but your labels for each row are in distinct columns.
And you only want the rows where the event value = 1 (occurred).
This is a good time to use gather to make your table taller,
filter to select event = 1,
and select to re-order your variables.
See if you can figure out what each line in the pipe below does.

```{r}
ae_table1%>%
gather(key = "adv_event", value = "present", low_wbc:blood_clot) %>%
filter(present == "1") %>%
select(adv_event, `A: IFL`:`G: IROX`) ->
ae_table2
ae_table2
```

That looks much better.
But some of the names could be improved,
particular the headers and the names of adverse events in column 1.

Change the adv_event values for the 5 rows to
c("Leukopenia", "Neuropathy", "Diarrhea", "Vomiting",
"Venous Thromboembolism")
and the names of the 4 columns to
c("Adverse Event", "IFL", "FOLFOX", "IROX")

by editing the chunk below and running it.

```{r}
ae_table2$adv_event <- c("Leukopenia", "Neuropathy", "Diarrhea", "Vomiting",
"Venous Thromboembolism")
names(ae_table2) <- c("Adverse Event", "IFL", "FOLFOX", "IROX")
ae_table2
```

This is fine, but you could use some nice custom formatting.

Let's use the flextable package and neaten this up.
Run the code chunk below to make a basic flextable.

```{r}
library(flextable)
library(officer)
ae_table2 %>%
flextable()

```

Now check out the formatting options at the [flextable website]
(https://cran.r-project.org/web/packages/flextable/vignettes/overview.html)

Now, add the following formatting:

1. Add a header row that spans columns 2-4, and says "Study Arm"
2. Autofit the column widths
3. Fix the alignments of chemo regimens and rates
4. Set arms to bold
5. Control font and font sizes
6. Use conditional formatting to make any AE > 20% in red font and italic

Be creative!



Below is another approach to another adverse event table, albeit with less formatting control, using tableby.
Really quick and easy.
```{r}
# low effort tableby version
tableby(arm ~ low_wbc + neuropathy + diarrhea + vomiting + blood_clot, data = study) ->
table2
summary(table2)

```


206 changes: 107 additions & 99 deletions data/expanded_mockstudy.R
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,99 +1,107 @@
#### expanding mockstudy
library(arsenal)
library(randomNames)
library(tidyverse)
library(here)
data(mockstudy)
head(mockstudy)

# note 7 missing race, assign to "Other"
mockstudy$race[is.na(mockstudy$race)] <- "Other"

mockstudy$race_num <- factor(case_when(
mockstudy$race == "African-Am" ~ 3,
mockstudy$race =="Asian" ~ 2,
mockstudy$race =="Caucasian" ~ 5,
mockstudy$race =="Hawaii/Pacific" ~ 2,
mockstudy$race =="Hispanic" ~ 4,
mockstudy$race =="Native-Am/Alaska" ~ 1,
mockstudy$race =="Other" ~ 6
))


mockstudy$gender_num <- factor(case_when(
mockstudy$sex == "Male" ~ 0,
mockstudy$sex =="Female" ~ 1
))

mockstudy$firstname <-
randomNames(ethnicity=mockstudy$race_num,
gender = mockstudy$gender_num,
which.names = "first",
sample.with.replacement = TRUE)

mockstudy$lastname <-
randomNames(ethnicity=mockstudy$race_num,
gender = mockstudy$gender_num,
which.names = "last",
sample.with.replacement = TRUE)

mockstudy %>%
mutate(ae_random = runif(n=nrow(.), min=0, max=1)) %>%
mutate(ae_random = ifelse(arm=="IROX",
ae_random+0.14, ae_random)) %>%
mutate(ae_random = ifelse(arm=="IFL",
ae_random+ 0.07, ae_random)) %>%
mutate(low_wbc = ifelse(ae_random<0.23, 1, 0)) %>%
mutate(neuropathy = ifelse(ae_random >0.05 & ae_random<0.23, 1, 0)) %>%
mutate(diarrhea = ifelse(ae_random >0.02 & ae_random<0.26, 1, 0)) %>%
mutate(vomiting = ifelse(ae_random >0.06 & ae_random<0.26, 1, 0)) %>%
mutate(blood_clot = ifelse(ae_random >0.13 & ae_random<0.18, 1, 0)) %>%
rename(mdquality = mdquality.s) %>%
select(case, firstname, lastname, age, age.ord, sex, race, bmi,
fu.time,fu.stat, low_wbc:blood_clot) ->
mockstudy

write.csv(mockstudy, here("mockstudy2.csv"))



#next - add random adverse events for FOLFOX, IROX, IFL
# FOLFOX more than 10%
# infection
# low WBC 23%
# fever 5%
# SOB
# Anemia
# bruising, bleeding
# fatigue
# neuropathy 18%
# nausea
# diarrhea
# mouth sores
# 1-10%
# peeling skin on hands
# runny nose
# sunburn
# hair loss
# chipped nails
# blood clot
#
# less than 1%
# difficulty swallowing
# tinnitus

#IROX
# low WBC 30%
# fever 29%
# diarrhea 24%
# n/V 12%%
# neuropathy 18%

#IFL
# vomiting 11%
# diarrhea 16%
#mouth sores 3%
# low WBC 26%
# fever 5%


#### expanding mockstudy
library(arsenal)
library(randomNames)
library(tidyverse)
library(here)
data(mockstudy)
head(mockstudy)

# note 7 missing race, assign to "Other"
mockstudy$race[is.na(mockstudy$race)] <- "Other"

mockstudy$race_num <- factor(case_when(
mockstudy$race == "African-Am" ~ 3,
mockstudy$race =="Asian" ~ 2,
mockstudy$race =="Caucasian" ~ 5,
mockstudy$race =="Hawaii/Pacific" ~ 2,
mockstudy$race =="Hispanic" ~ 4,
mockstudy$race =="Native-Am/Alaska" ~ 1,
mockstudy$race =="Other" ~ 6
))


mockstudy$gender_num <- factor(case_when(
mockstudy$sex == "Male" ~ 0,
mockstudy$sex =="Female" ~ 1
))

mockstudy$alive <- factor(case_when(
mockstudy$fu.stat == 2 ~ 0,
mockstudy$fu.stat == 1 ~ 1
))

mockstudy$firstname <-
randomNames(ethnicity=mockstudy$race_num,
gender = mockstudy$gender_num,
which.names = "first",
sample.with.replacement = TRUE)

mockstudy$lastname <-
randomNames(ethnicity=mockstudy$race_num,
gender = mockstudy$gender_num,
which.names = "last",
sample.with.replacement = TRUE)

mockstudy %>%
mutate(ae_random = runif(n=nrow(.), min=0, max=1)) %>%
mutate(ae_random = ifelse(arm=="G: IROX",
ae_random+0.12, ae_random)) %>%
mutate(ae_random = ifelse(arm=="A: IFL",
ae_random+ 0.07, ae_random)) %>%
mutate(low_wbc = ifelse(ae_random>0.9, 1, 0)) %>%
mutate(neuropathy = ifelse(ae_random >0.97, 1, 0)) %>%
mutate(diarrhea = ifelse(ae_random >0.95, 1, 0)) %>%
mutate(vomiting = ifelse(ae_random >0.88, 1, 0)) %>%
mutate(blood_clot = ifelse(ae_random >0.99, 1, 0)) %>%
rename(mdquality = mdquality.s) %>%
select(case, firstname, lastname, age, age.ord, sex, race, bmi, hgb, alk.phos,
ast, arm, mdquality, fu.time, alive, low_wbc:blood_clot) ->
mockstudy

write.csv(mockstudy, here("data/mockstudy2.csv"))

# next - eliminate G:, F:, I: at beginning of names
# next - diff random for each of 5 AEs
# next - differential different between arms for each of 5 AEs
#

#next - add random adverse events for FOLFOX, IROX, IFL
# FOLFOX more than 10%
# infection
# low WBC 23%
# fever 5%
# SOB
# Anemia
# bruising, bleeding
# fatigue
# neuropathy 18%
# nausea
# diarrhea
# mouth sores
# 1-10%
# peeling skin on hands
# runny nose
# sunburn
# hair loss
# chipped nails
# blood clot
#
# less than 1%
# difficulty swallowing
# tinnitus

#IROX
# low WBC 30%
# fever 29%
# diarrhea 24%
# n/V 12%%
# neuropathy 18%

#IFL
# vomiting 11%
# diarrhea 16%
#mouth sores 3%
# low WBC 26%
# fever 5%


Loading