We will here use the create_table_forestplot and dessin_forest_plot functions in the Data & Functions page.
Download Forest_autom.R

# Libraries
library(rms) # or library(survival)
library(cmprsk)
library(forestplot) # version 3.1.1

# Data used
set.seed(143)
lung$bras <- sample(c('A','B'), nrow(lung), replace=TRUE)
lung$status2 <- sample(0:2, nrow(lung), replace=TRUE, prob=c(0.6, 0.25, 0.15))

lung$status <- lung$status-1
lung$ecog[lung$ph.ecog %in% c(0,1)] <- '01'
lung$ecog[lung$ph.ecog %in% c(2,3)] <- '23'

lung <- lung[,c("bras","time","status","status2","age","sex","ecog","meal.cal")]

`create_table_forestplot` function

cov <- c("age","sex","ecog","meal.cal")
quali <- rep(1,length(cov))
quali[cov %in% c("age","meal.cal")] <-0

tabforest1 <- create_table_forestplot(data_frame=lung,
                                      covariables=cov,
                                      var_quali_ou_quanti=quali,
                                      modele_souhaite="cox",
                                      bras="bras",
                                      var_dinteret="status",
                                      delai_dinteret="time")

Subset	A (Events/Pts)	B (Events/Pts)	HR (95%CI)	Pvalue_interaction
age				> 0.2
≤63	46/67	34/50	1.17 (0.75 - 1.82)
>63	51/63	34/48	0.97 (0.63 - 1.5)
sex				> 0.2
1	62/77	50/61	1.09 (0.75 - 1.58)
2	35/53	18/37	0.91 (0.51 - 1.61)
ecog				> 0.1
01	67/98	52/78	1.22 (0.85 - 1.75)
23	30/32	15/19	0.71 (0.38 - 1.32)
meal.cal				0.99
≤975	40/55	25/36	1.1 (0.67 - 1.82)
>975	37/48	32/42	1.1 (0.68 - 1.76)
All patients	97/130	68/98	1.06 (0.78 - 1.45)

Here is another example where optional parameters are used:

nom.var <- c("Age (years)","Gender","ECOG","Calories consumed")
levels.var <- c("Male","Female","0-1","2-3")

tabforest2 <- create_table_forestplot(data_frame=lung,
                                      covariables=cov, nom_var=nom.var,
                                      levels_var=levels.var,
                                      var_quali_ou_quanti=quali,
                                      modele_souhaite="cox",
                                      bras="bras", nom_bras=c("Arm A","Arm B"),
                                      var_dinteret="status",
                                      delai_dinteret="time",
                                      percent=TRUE)

Subset	Arm A (Events/Pts)	Arm B (Events/Pts)	HR (95%CI)	Pvalue_interaction
Age (years)				> 0.2
≤63	46/67 (69%)	34/50 (68%)	1.17 (0.75 - 1.82)
>63	51/63 (81%)	34/48 (71%)	0.97 (0.63 - 1.5)
Gender				> 0.2
Male	62/77 (81%)	50/61 (82%)	1.09 (0.75 - 1.58)
Female	35/53 (66%)	18/37 (49%)	0.91 (0.51 - 1.61)
ECOG				> 0.1
0-1	67/98 (68%)	52/78 (67%)	1.22 (0.85 - 1.75)
2-3	30/32 (94%)	15/19 (79%)	0.71 (0.38 - 1.32)
Calories consumed				0.99
≤975	40/55 (73%)	25/36 (69%)	1.1 (0.67 - 1.82)
>975	37/48 (77%)	32/42 (76%)	1.1 (0.68 - 1.76)
All patients	97/130 (52%)	68/98 (69%)	1.06 (0.78 - 1.45)

In the example above, the covariates, the levels of the categorical variables and the treatment arms are named. The percentages of the numbers are also displayed.

`dessin_forest_plot` function

Here is an example of a forest plot without the optional arguments:

dessin_forest_plot(data_frame=lung,
                   covariables=cov,
                   var_quali_ou_quanti=quali,
                   modele_souhaite="cox",
                   bras="bras",
                   var_dinteret="status",
                   delai_dinteret="time")

We can:

Give a name to the covariates, categorical variable levels and treatment arms

dessin_forest_plot(data_frame=lung,
                   covariables=cov, nom_var=nom.var,
                   levels_var=levels.var,
                   var_quali_ou_quanti=quali,
                   modele_souhaite="cox",
                   bras="bras", nom_bras=c("Arm A","Arm B"),
                   var_dinteret="status",
                   delai_dinteret="time")

Change the position and color of the graph, as well as the shape of the estimate

dessin_forest_plot(data_frame=lung,
                   covariables=cov, nom_var=nom.var,
                   levels_var=levels.var,
                   var_quali_ou_quanti=quali,
                   modele_souhaite="cox",
                   bras="bras", nom_bras=c("Arm A","Arm B"),
                   var_dinteret="status",
                   delai_dinteret="time",
                   pos_graphique=2,
                   col=fpColors(box="palegreen4", lines="palegreen3", summary="palegreen4"),
                   forme=fpDrawNormalCI)

Change the size of the estimates

dessin_forest_plot(data_frame=lung,
                   covariables=cov, nom_var=nom.var,
                   levels_var=levels.var,
                   var_quali_ou_quanti=quali,
                   modele_souhaite="cox",
                   bras="bras", nom_bras=c("Arm A","Arm B"),
                   var_dinteret="status",
                   delai_dinteret="time",
                   pos_graphique=2,
                   col=fpColors(box="palegreen4", lines="palegreen3", summary="palegreen4"),
                   forme=fpDrawNormalCI,
                   boxsize=0.8)

Change the threshold values of quantitative variables

dessin_forest_plot(data_frame=lung,
                   covariables=cov, nom_var=nom.var,
                   levels_var=levels.var,
                   var_quali_ou_quanti=quali,
                   modele_souhaite="cox",
                   bras="bras", nom_bras=c("Arm A","Arm B"),
                   var_dinteret="status",
                   delai_dinteret="time",
                   cut_off=c(70,950),
                   pos_graphique=2,
                   col=fpColors(box="palegreen4", lines="palegreen3", summary="palegreen4"),
                   forme=fpDrawNormalCI,
                   boxsize=0.8)

Choose a log scale

dessin_forest_plot(data_frame=lung,
                   covariables=cov, nom_var=nom.var,
                   levels_var=levels.var,
                   var_quali_ou_quanti=quali,
                   modele_souhaite="cox",
                   bras="bras", nom_bras=c("Arm A","Arm B"),
                   var_dinteret="status",
                   delai_dinteret="time",
                   cut_off=c(70,950),
                   xlog=TRUE,
                   pos_graphique=2,
                   col=fpColors(box="palegreen4", lines="palegreen3", summary="palegreen4"),
                   forme=fpDrawNormalCI)

Specified the difference in direction below the axis

dessin_forest_plot(data_frame=lung,
                   covariables=cov, nom_var=nom.var,
                   levels_var=levels.var,
                   var_quali_ou_quanti=quali,
                   modele_souhaite="cox",
                   bras="bras", nom_bras=c("Arm A","Arm B"),
                   var_dinteret="status",
                   delai_dinteret="time",
                   cut_off=c(70,950),
                   xlog=TRUE,
                   pos_graphique=2,
                   col=fpColors(box="palegreen4", lines="palegreen3", summary="palegreen4"),
                   forme=fpDrawNormalCI,
                   favors=c("Favors A","Favors B"))

Use a logistic model

dessin_forest_plot(data_frame=lung,
                   covariables=cov, nom_var=nom.var,
                   levels_var=levels.var,
                   var_quali_ou_quanti=quali,
                   modele_souhaite="log",
                   bras="bras", nom_bras=c("Arm A","Arm B"),
                   var_dinteret="status",
                   pos_graphique=2,
                   col=fpColors(box="palegreen4", lines="palegreen3", summary="palegreen4"),
                   forme=fpDrawNormalCI)

Use a Fine & Gray model

dessin_forest_plot(data_frame=lung,
                   covariables=cov, nom_var=nom.var,
                   levels_var=levels.var,
                   var_quali_ou_quanti=quali,
                   modele_souhaite="finegray",
                   bras="bras", nom_bras=c("Arm A","Arm B"),
                   var_dinteret="status2", failcode=1,
                   delai_dinteret="time",
                   pos_graphique=2,
                   col=fpColors(box="palegreen4", lines="palegreen3", summary="palegreen4"),
                   forme=fpDrawNormalCI)

Use more options : remove interaction pvalues, add pvalues for each subgroup, change the place of the plot and fix the number of digits

dessin_forest_plot(data_frame=lung,
                   covariables=cov, nom_var=nom.var,
                   levels_var=levels.var,
                   var_quali_ou_quanti=quali,
                   modele_souhaite="finegray",
                   bras="bras", nom_bras=c("Arm A","Arm B"), label_bras="# success/ # patients",
                   var_dinteret="status2", failcode=1,
                   delai_dinteret="time",
                   digit=2, fixed_digit=TRUE,
                   test=NULL,
                   pos_graphique=0,
                   col=fpColors(box="palegreen4", lines="palegreen3", summary="palegreen4"),
                   forme=fpDrawNormalCI,
                   print_pval = TRUE)

Contact

This document is a work of the statistics team in the Biostatistics and Medical Information Department at Saint-Louis Hospital in Paris (SBIM).
Based on The R Graph Gallery by Yan Holtz.

SBIM

create_table_forestplot function

dessin_forest_plot function

Contact

`create_table_forestplot` function

`dessin_forest_plot` function