In felix-hof/hMean: Confidence Curves and P-Value Functions for Meta-Analysis

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>",
  out.width = "100%",
  fig.width = 9
)

The confMeta package implements methods related to meta-analysis. The functions that confMeta provides, can be categorized roughly into three categories:

• An S3 class confMeta and a plot method for it
• p-value functions
• Heterogeneity adjustments for standard errors

Installation

Currently, the package is developed on Github. Thus, the easiest way to install it is via the remotes package. If the remotes package is not installed, you can run the following line of code to do so.

install.packages("remotes")

Once remotes has been installed, you can install confMeta by running

remotes::install_github("felix-hof/confMeta")

Once installed, the package can be used by loading it into the R session. This can be achieved by running

library(confMeta)

Usage examples

Simulating individual studies

Consider the hypothetical scenario where we have n = 3 individual studies that should be combined into a single confidence interval using a confidence level of 1 - alpha = 0.95. We can simulate these by running the following code

n <- 3
conf_level <- 0.95
estimates <- rnorm(n)
SEs <- rgamma(n, 5, 5)

Here, the object estimates contains the individual study estimates, whereas the object SEs contains the corresponding standard errors.

Creating the confMeta object

With these individual studies, a confMeta object can be created. However, this requires the specification of a p-value function, i.e. a method, that takes the individual studies (argument estimates), their standard errors (argument SEs), and the mean under the null-hypothesis (argument mu) as input and returns the corresponding p-value at the specified mean value. The confMeta package provides implementations for the following p-value functions

• Harmonic mean (Function: p_hmean)
• Wilkinson (Function: p_wilkinson)
• Pearson (Function: p_pearson)
• Edgington (Function: p_edgington)
• Fisher (Function: p_fisher)
• Tippett (Function: p_tippett)

In this example, we choose Edgington's method. Thus, we can create the confMeta object as follows

cm <- confMeta(
    estimates = estimates,
    SEs = SEs,
    conf_level = conf_level,
    fun = p_edgington,
    fun_name = "Edgington"
)

As the variable cm now contains the confMeta object, we can inspect it by running the following code

# See what elements it has
names(cm)

# Check out the combined confidence interval(s)
cm$joint_cis

Visualizations

The package also contains an autoplot method that can be used to visualize the p-value function. The documentation for this function can be inspected by running

?autoplot.confMeta

The method provides essentially two plots, one showing the p-value function and one constructing a forest plot. Which one is returned can be specified using the type argument.

# show the p-value function
autoplot(cm, type = "p")

# show the forest plot
autoplot(cm, type = "forest")

# show both
autoplot(cm, type = c("p", "forest"))

You can also compare different p-value functions with each other. In order to illustrate how this works, we need a second confMeta object, that uses a different p-value function.

cm2 <- confMeta(
    estimates = estimates,
    SEs = SEs,
    conf_level = conf_level,
    fun = p_fisher,
    fun_name = "Fisher"
)

Now, we can compare the two p-value functions to each other with in the following way

# show the p-value function
autoplot(cm, cm2, type = "p")

# show the forest plot
autoplot(cm, cm2, type = "forest")

# show both
autoplot(cm, cm2, type = c("p", "forest"))

felix-hof/hMean documentation built on Jan. 26, 2025, 4:59 p.m.