Forest Plot

Overview

The forest plot is useful for presenting the results of a meta-analysis — that is, a study of studies.  Each row in a forest plot represents one study.  A linear graph for each study indicates the result estimate and the lower and upper 95% confidence interval bounds.  A vertical line represents the “no effect” point.  Additional columns can provide the plotted information in text format as well as additional information, such as number of events, population size, etc.

[image here]

Typically a forest plot uses a linear scale for continuous outcomes or a logarithmic scale for binary outcomes (e.g., risk ratios, odds ratios, etc.). The WCMS allows you to specify the scale type.

[note on Total row if available]

With the Subtext option, you can provide notes on details such as p value, heterogeneity, etc.

Data Requirements

For each study represented in a forest plot, your source data must include a study label/ID, a numeric result estimate (“Relative Risk” in example data below), and the lower and upper bounds for the 95% confidence interval (in separate columns).  Each study graph is automatically generated from the result and the CI values.  A weight column can be specified for determining the size of each result indicator.  The chart designer also supports additional text columns from the source data, such as the number of study participants.

[image here]

Guidelines

• Many general audience members may not be familiar with forest plot conventions, so, unless your chart is intended solely for scientific audiences, include a summary of chart results in plain English.
• Interpretation of a forest plot depends on the type of outcome being studied (particularly the desirability of the outcome). The WCMS chart designer allows you to provide X-axis labels on the left and right sides of the “line of no effect” to help readers interpret the results.