Introduction to user interface functions

In regmedint: Regression-Based Causal Mediation Analysis with Interaction and Effect Modification Terms

## knitr configuration: http://yihui.name/knitr/options#chunk_options
library(knitr)
showMessage <- FALSE
showWarning <- TRUE
set_alias(w = "fig.width", h = "fig.height", res = "results")
opts_chunk$set(comment = "##", error= TRUE, warning = showWarning, message = showMessage,
               tidy = FALSE, cache = FALSE, echo = TRUE,
               fig.width = 7, fig.height = 7,
               fig.path = "man/figures")

Data preparation

The package require all variables to be numerical. So a multi-categorical factor needs to be converted to dummy variables or multiple dichotomous indicators. For survival outcome models, the indicator variable is for the event (1 = event, 0 = censored).

library(regmedint)
library(tidyverse)
## Prepare dataset
data(vv2015)

regmedint object

Following typical modeling workflow in R (e.g., lm and glm), a constructor function is used to create a model fit object. The summary method is the main user function for examining the results in the object. Lower-level methods such as coef, vcov, and confint are also provided for flexibility. The print method is mainly for meaningful implicit printing when only the object name is evaluated. All methods for the regmedint object has arguments a0, a1, m_cde, and c_cond. These are used to re-evaluate the results without re-fitting the underlying models.

regemedint() object constructor

regmedint_obj <- regmedint(data = vv2015,
                           ## Variables
                           yvar = "y",
                           avar = "x",
                           mvar = "m",
                           cvar = c("c"),
                           eventvar = "event",
                           ## Values at which effects are evaluated
                           a0 = 0,
                           a1 = 1,
                           m_cde = 1,
                           c_cond = 0.5,
                           ## Model types
                           mreg = "logistic",
                           yreg = "survAFT_weibull",
                           ## Additional specification
                           interaction = TRUE,
                           casecontrol = FALSE)

summary() for regmedint

summary(regmedint_obj)
summary(regmedint_obj, exponentiate = TRUE)
summary(regmedint_obj, m_cde = 0, c_cond = 1)
summary(regmedint_obj, m_cde = 0, c_cond = 1, level = 0.99)

coef() for regmedint

coef(regmedint_obj)
coef(regmedint_obj, m_cde = 0, c_cond = 1)

vcov() for regmedint

vcov(regmedint_obj)
vcov(regmedint_obj, m_cde = 0, c_cond = 1)

confint() for regmedint

confint(regmedint_obj)
confint(regmedint_obj, m_cde = 0, c_cond = 1)
confint(regmedint_obj, m_cde = 0, c_cond = 1, level = 0.99)

print() for regmedint

regmedint_obj # Implicit printing
print(regmedint_obj)
print(regmedint_obj, m_cde = 0, c_cond = 1)

Methods for summary_regmedint

The summary method for the regmedint object returns an object of class summary_regmedint. To extract the mediation analysis result table as a matrix, use the coef method.

coef() for summary_regmedint

coef(summary(regmedint_obj))

print() for summary_regmedint

regmedint_summary_obj <- summary(regmedint_obj)
regmedint_summary_obj # Implicit printing
print(regmedint_summary_obj)

regmedint documentation built on April 7, 2022, 1:17 a.m.