In jfrancoiscollin/ClinReport: Statistical Reporting for Scientific Publications, Epidemiological Studies and Clinical Trials

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>"
)

Objective

In this vignette we will see how to modify your outputs. Basically you can:

Mix qualitative and quantitative descriptive statistics: regroup() function
Choose the quantitative statistics you want to display: report.quanti() function
Report row percentages instead of column percentages: report.quali() function
Add a specific statistics to an existing table: add.stat() function
Split an existing table in 2 parts: split() function
Drop one or several levels of a variable in a table
Transpose the statistics in columns

Load the usual R packages

Start by loading all usual libraries.

library(ClinReport)
library(officer)
library(flextable)
library(emmeans)

Load your data.

# We will use fake data
data(datafake)
print(head(datafake))

Mixed Quantitative and Qualitative statistics

You can mix qualitative and quantitative outputs. It's possible only if there is only 1 explicative variable, and it should be the same variable for both response:

tab1=report.quanti(data=datafake,y="y_numeric",
        x1="GROUP",subjid="SUBJID",y.label="Y numeric")

tab2=report.quali(data=datafake,y="y_logistic",
        x1="GROUP",subjid="SUBJID",y.label="Y logistic")

tab3=regroup(tab1,tab2,rbind.label="The label of your choice")


report.doc(tab3,title="Mixed Qualitative and Quantitative output",
colspan.value="Treatment group")

Report a specific quantitative statistic

You will have to specify the function corresponding to the statistics and use the func.stat argument

es=function(x) mean(x,na.rm=T)/sd(x,na.rm=T) 

tab=report.quanti(data=datafake,y="y_numeric",x1="GROUP",
total=TRUE,subjid="SUBJID",
func.stat=es,
func.stat.name="Effect size")

report.doc(tab,title="Example of a specific statistic reporting",
colspan.value="Treatment group")

Add a specific statistics to an existing table

You can also add a specific (or several) statistics to an existing table.

Let's say you want the default statistics (mean, median, sd etc..) plus some specific statistics like the mode or the coefficient of variation.

In this case, you can use the add.stat() function with the pos argument to choose where you want to add this statistic:

# The default statistics are given here:

tab1=report.quanti(data=datafake,y="y_numeric",x1="GROUP",total=TRUE,subjid="SUBJID")

 # Define the function corresponding to the coefficient of variation for example

cv=function(y) sd(y,na.rm=TRUE)/mean(y,na.rm=TRUE)

 # We use the add.stat function to add CV at the second row:

tab1.cv=add.stat(tab1,datafake,func.stat=cv,func.stat.name="Coef. Var",
 pos=2)

report.doc(tab1.cv,title="Example of adding a coefficient of variation")

 # Same with 2 explicative variables

tab=report.quanti(data=datafake,y="y_numeric",x1="GROUP",
 x2="TIMEPOINT",total=TRUE,subjid="SUBJID",
        at.row="TIMEPOINT")

 tab=add.stat(tab,datafake,func.stat=cv,func.stat.name="Coef. Var",
 pos=2)


 # And on position 5, we can add for example the mode

 mode=function(x)
 {
   x=na.omit(x)
   ux <- unique(x)
   ux[which.max(tabulate(match(x, ux)))]
 }


 tab=add.stat(tab,datafake,func.stat=mode,func.stat.name="Mode",pos=5)

 report.doc(tab,title="Example of adding 2 more statistics in an existing table",
 colspan.value="Treatment Group")

Row or column percentages

If you want to display the row percentages instead of column percentages in report.quali() function, you just have to set the argument percent.col to FALSE:

tab=report.quali(data=datafake,y="y_logistic",x1="GROUP",
total=TRUE,subjid="SUBJID",percent.col=FALSE)

report.doc(tab,title="Example of row percentage reporting",
colspan.value="Treatment group")

Drop levels

If you want to drop some levels in the table, you can use drop.x1 drop.x2 or drop.y if 'y' is a qualitative variable.

For example. If we just want the statistics of group A:

tab=report.quali(data=datafake,y="y_logistic",x1="GROUP",
subjid="SUBJID",drop.x1=c("B","C"))

report.doc(tab,title="Example of row percentage reporting",
colspan.value="Treatment group")

Drop missing values

Sometimes it can be useful to remove the missing values so that percentages are computed only on non missing observations. For that, just use the remove.missing argument in report.quali:

tab=report.quali(data=datafake,y="y_logistic",x1="GROUP",
remove.missing=TRUE)

report.doc(tab,title="Example of dropping missing values",
colspan.value="Treatment group")

Transpose Descriptive statistics (experimental)

tab=report.quali(data=datafake,
y="y_logistic",x1="GROUP",
subjid="SUBJID",remove.missing=T)

# The default output
report.doc(tab)

# The transposed output
report.doc(transpose(tab))

Transpose LS-Means

mod=lm(y_numeric~GROUP,data=datafake)
pairs=pairs(emmeans(mod,~GROUP))

tab=report.lsmeans(pairs,transpose=TRUE)

report.doc(tab,title="Example of transposing LS-Means in column",
colspan.value="Treatment group")