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 the different graphics that you can produce with ClinReport.

First step

Start by loading all usual libraries.

library(ClinReport)
library(officer)
library(flextable)
library(dplyr)
library(reshape2)
library(nlme)
library(emmeans)
library(car)

Load your data.

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

Quantitative graphics

First you need to produce a quantitative table using report.quanti() function then you use the plot() function on the result:

If there is no explicative variable, the corresponding graphic of the call to report.quanti() is a barplot of the mean of the response value.

# Only one numerical response

tab1=report.quanti(data=datafake,y="y_numeric")
plot(tab1,title="Mean (+/-SD) of y",add.sd=T,ylab="Mean Response value",xlab="")

If there is one explicative variable, the corresponding graphic will be a barplot per treatment group.

You can personnalize the outputs using, the usual, title xlab ylab ylim etc.. options.

#  one numerical response ; one explicative variable

tab2=report.quanti(data=datafake,y="y_numeric",x1="GROUP")
plot(tab2,title="Mean (+/-SD) of y per treatment group",add.sd=T,ylab="Mean of y",xlab="",
ylim=c(-1,8))

If there is also a second explicative variable. The corresponding plot will be a lineplot:

The means will be connected according to this second variable to capture the evolution across its different levels.

#  one numerical response ; two explicative variable

tab3=report.quanti(data=datafake,y="y_numeric",x1="GROUP",x2="TIMEPOINT")
plot(tab3,title="Mean (+/-SD) of y per treatment group",add.sd=T,ylab="Mean of y",xlab="",
ylim=c(-1,8))

Qualitative graphics

For qualitative statistics, the corresponding graphics will be barplots.

If only the argument y is supplied, the corresponding graphic will looks like:

# Only one categorical response

tab1=report.quali(data=datafake,y="y_logistic")
plot(tab1,title="Distribution of y (%)",ylab="Percentages",xlab="",ylim=c(0,100))

If there is one explicative variable, it will be split by color and split on the x-axis:

# one categorical response ; one categorical explicative variable

tab1=report.quali(data=datafake,y="y_logistic",x1="GROUP")
plot(tab1,title="Distribution of y per treatment group (%)",
ylab="Percentages",xlab="",ylim=c(0,100),legend.label="Levels of y")

If there is two explicative variables, it will be split by color, split on the x-axis and split by panels:

# one categorical response ; two categorical explicative variables

tab1=report.quali(data=datafake,y="y_logistic",x1="GROUP",x2="VAR")
plot(tab1,title="Distribution of y per treatment group\nand by level of VAR (%)",
ylab="Percentages",xlab="",ylim=c(0,100),legend.label="Levels of y")

LS-means graphics

For LS-Means graphics, it will be pretty much the same graphics as the graphics of quantitative statistics.

Instead of displaying the standard deviations, it's the 95% confidence intervals that are displayed.

For one explicative variable:

# Only one categorical response


mod=lme(y_numeric~GROUP,
random=~1|SUBJID,data=datafake,na.action=na.omit)

test=emmeans(mod,~GROUP)

tab.mod=report.lsmeans(lsm=test)

plot(tab.mod,title="LS-means of y per treatment group",
ylab="LS-Means",xlab="",add.ci=TRUE)

With two explicative variables:

# Only one categorical response

mod=lme(y_numeric~baseline+GROUP+TIMEPOINT+GROUP*TIMEPOINT,
random=~1|SUBJID,data=datafake,na.action=na.omit)

test=emmeans(mod,~GROUP|TIMEPOINT)

tab.mod=report.lsmeans(lsm=test)

plot(tab.mod,title="LS-means evolution of y per treatment group\nas a function of time",
ylab="LS-Means",xlab="",add.ci=TRUE)

jfrancoiscollin/ClinReport documentation built on Sept. 20, 2020, 7:33 a.m.