Summarize variables and regression models for final publication
Switch branches/tags
Nothing to show
Clone or download
Latest commit 73edbc1 May 7, 2018

README.md

summarizer

The summarizer package provides functions that help you create elegant final results tables and charts when modelling. Its design follows Hadley Wickham's tidy tool manifesto.

Installation and Documentation

You can install summarizer from github with:

# install.packages("devtools")
devtools::install_github("ewenharrison/summarizer")

It is not a dependent, but it is recommended that this package is used together with dplyr which can be installed via:

install.packages("dplyr")

To install off-line (or in a Safe Haven), download the zip file and use devtools::install_local().

Main Features

1. Summarise variables/factors by a categorical variable

summary.factorlist() is a simple wrapper used to summarise any number of variables by a single categorical variable. This is usually "Table 1" of a study report. The categorical variable can have a maximum of five levels.

library(summarizer)
library(dplyr)

# Load example dataset, modified version of survival::colon
data(colon_s)

# Table 1 - Patient demographics ----
explanatory = c("age", "age.factor", "sex.factor", "obstruct.factor")
dependent = "perfor.factor"
colon_s %>%
  summary.factorlist(dependent, explanatory, p=T)

summary.factorlist() is also commonly used to summarise any number of variables by an outcome variable (say dead yes/no).

# Table 2 - 5 yr mortality ----
explanatory = c("age.factor", "sex.factor", "obstruct.factor", "perfor.factor")
dependent = 'mort_5yr'
colon_s %>%
  summary.factorlist(dependent, explanatory)

Tables can be knitted to PDF, Word or html documents. We do this in RStudio from a .Rmd document. Example chunk:

knitr::kable(example_table, row.names=FALSE, align=c("l", "l", "r", "r", "r"))

2. Summarise regression model results in final table format

The second main feature is the ability to create final tables for logistic glm(), hierarchical logistic lme4::glmer() and Cox proprotional hazard survival::coxph() regression models.

The summarizer() "all-in-one" function takes a single dependent variable with a vector of explanatory variable names (continuous or categorical variables) to produce a final table for publication including summary statistics, univariable and multivariable regression analyses. The first columns are those produced by summary.factorist().

e.g. glm(depdendent ~ explanatory, family="binomial")

explanatory = c("age.factor", "sex.factor", "obstruct.factor", "perfor.factor")
dependent = 'mort_5yr'
colon_s %>%
  summarizer(dependent, explanatory)

Where a multivariable model contains a subset of the variables specified in the full univariable set, this can be specified.

explanatory = c("age.factor", "sex.factor", "obstruct.factor", "perfor.factor")
explanatory.multi = c("age.factor", "obstruct.factor")
dependent = 'mort_5yr'
colon_s %>%
  summarizer(dependent, explanatory, explanatory.multi)

Random effects.

e.g. lme4::glmer(dependent ~ explanatory + (1 | random_effect), family="binomial")

explanatory = c("age.factor", "sex.factor", "obstruct.factor", "perfor.factor")
explanatory.multi = c("age.factor", "obstruct.factor")
random.effect = "hospital"
dependent = 'mort_5yr'
colon_s %>%
  summarizer(dependent, explanatory, explanatory.multi, random.effect)

metrics=TRUE provides common model metrics.

colon_s %>%
  summarizer(dependent, explanatory, explanatory.multi,  metrics=TRUE)

Cox proportional hazards

e.g. survival::coxph(dependent ~ explanatory)

explanatory = c("age.factor", "sex.factor", "obstruct.factor", "perfor.factor")
dependent = "Surv(time, status)"

colon_s %>% 
	summarizer(dependent, explanatory)

Rather than going all-in-one, any number of subset models can be manually added on to a summary.factorlist() table using summarizer.merge(). This is particularly useful when models take a long-time to run or are complicated.

Note requirement for glm.id=TRUE. fit2df is a subfunction extracting most common models to a dataframe.

explanatory = c("age.factor", "sex.factor", "obstruct.factor", "perfor.factor")
explanatory.multi = c("age.factor", "obstruct.factor")
random.effect = "hospital"
dependent = 'mort_5yr'

# Separate tables
colon_s %>%
  summary.factorlist(dependent, explanatory, glm.id=TRUE) -> example.summary

colon_s %>%
  glmuni(dependent, explanatory) %>%
  fit2df(estimate.suffix=" (univariable)") -> example.univariable

colon_s %>%
  glmmulti(dependent, explanatory) %>%
  fit2df(estimate.suffix=" (multivariable)") -> example.multivariable


colon_s %>%
  glmmixed(dependent, explanatory, random.effect) %>%
  fit2df(estimate.suffix=" (multilevel") -> example.multilevel

# Pipe together
example.summary %>% 
  summarizer.merge(example.univariable) %>% 
  summarizer.merge(example.multivariable) %>% 
  summarizer.merge(example.multilevel) %>% 
  select(-c(glm.id, index)) -> example.final
example.final

Cox Proportional Hazards example with separate tables merged together.

explanatory = c("age.factor", "sex.factor", "obstruct.factor", "perfor.factor")
explanatory.multi = c("age.factor", "obstruct.factor")
dependent = "Surv(time, status)"

# Separate tables
colon_s %>%
	summary.factorlist(dependent, explanatory, glm.id=TRUE) -> example2.summary

colon_s %>%
	coxphuni(dependent, explanatory) %>%
	fit2df(estimate.suffix=" (univariable)") -> example2.univariable

colon_s %>%
  coxphmulti(dependent, explanatory.multi) %>%
  fit2df(estimate.suffix=" (multivariable)") -> example2.multivariable

# Pipe together
example2.summary %>% 
	summarizer.merge(example2.univariable) %>% 
	summarizer.merge(example2.multivariable) %>% 
	select(-c(glm.id, index)) -> example2.final
example2.final

3. Summarise regression model results in plot

Models can be summarized with odds ratio/hazard ratio plots using or.plot or hr.plot (hr.plot not fully tested).

# OR plot
explanatory = c("age.factor", "sex.factor", "obstruct.factor", "perfor.factor")
dependent = 'mort_5yr'
colon_s %>%
  or.plot(dependent, explanatory)
# Previously fitted models (`glmmulti()` or `glmmixed()`) can be provided directly to `glmfit`  
  
# HR plot (not fully tested)
explanatory = c("age.factor", "sex.factor", "obstruct.factor", "perfor.factor")
dependent = "Surv(time, status)"
colon_s %>%
  hr.plot(dependent, explanatory, dependent_label = "Survival")
# Previously fitted models (`coxphmulti`) can be provided directly using `coxfit`

Our own particular Rstan models are supported and will be documented in the future. Broadly, if you are running (hierarchical) logistic regression models in Stan with coefficients specified as a vector labelled beta, then fit2df() will work directly on the stanfit object in a similar manner to if it was a glm or glmerMod object.

4. Kaplan-Meier survival plots

KM plots can be produced using the library(survminer)

# KM plot
explanatory = c("perfor.factor")
dependent = "Surv(time, status)"
colon_s %>% 
  surv.plot(dependent, explanatory, xlab="Time (days)", pval=TRUE, legend="none")

Notes

Use Hmisc::label() to assign labels to variables for tables and plots.

label(colon_s$age.factor) = "Age (years)"

Export dataframe tables directly or to R Markdown using knitr::kable().

Note wrapper summary.missing() can be useful. Wraps mice::md.pattern.

colon_s %>%
  summary.missing(dependent, explanatory)