Cart

Frequently Asked Questions




Graph tip - How can I plot an odds ratio plot, also known as a Forest plot, or a meta-analysis plot?

FAQ# 809    Last Modified 1-January-2009

screen shot

This example shows how to make an odds ratio plot, also known as a Forest plot or a meta-analysis plot, graphs odds ratios (with 95% confidence intervals) from several studies. It also shows how to place a custom grid line on a graph.


How to do it:

GraphPad Prism can make this kind of graph easily.

  1. When you start the program, or use New table/graph to create a Column data table,.
     
  2. Enter the data. Enter in each column the Odds ratio itself, as well as the high and low confidence limits (computed elsewhere). Each treatment group should be entered in a separate column, and you'll enter three values in each column (the odds ratio, and the lower and upper confidence limits). The order of those three values doesn't matter. Label the groups by entering column titles. You may use Greek symbols, super and subscript, etc. This example plots results from three studies, but use columns D, E, F, etc. to plot results from more studies.

    screen shot
  3. Choose the thumbnail for how you want your graph to look. Also choose to plot the median and range since the median of the three values you entered is the middle one, the odds ratio, and the range of the three values you entered will cover the confidence interval computed elsewhere. 

  1. View the graph.
  2. Some metaanalysis plots use larger symbols to denote studies with more subjects. Prism can't do this automatically. But you can adjust the size of each symbol in the Format Graph dialog, so make some larger and some smaller. 
     
  3. Edit the graph and axis titles, add an additional custom line (actually a dotted line) at X=1.0 (from Format Axes dialog as shown below), and you are done.

screen shot

 
Epidemiologists often like to make the x axis logarithmic. This makes sense as it makes odd ratios greater than 1.0 and less than 1.0 symmetrical  (for example, an odds ratio of 2.0 becomes symmetrical with an odds ratio of  0.5).

This is easy to do with Prism. Format the X axis with a Log 10 scale and set the range from 0.1 to 10 with "antilog" numbering and a grid line at X=1.

 

Download this file to see how this graph was made.



Keywords: horizontal error bars odd's ratio metanalysis meta metaanalysis Forrest