Year : 2015  |  Volume : 57  |  Issue : 3  |  Page : 313--314

A simple rearrangement can improve visual understanding of a Forest plot


Nilesh Shah1, Chittaranjan Andrade2,  
1 Department of Psychiatry, Lokmanya Tilak Municipal Medical College and General Hospital, Sion, Mumbai, Maharashtra, India
2 Department of Psychopharmacology, National Institute of Mental Health and Neurosciences, Bengaluru, Karnataka, India

Correspondence Address:
Chittaranjan Andrade
Department of Psychopharmacology, National Institute of Mental Health and Neurosciences, Bengaluru, Karnataka
India




How to cite this article:
Shah N, Andrade C. A simple rearrangement can improve visual understanding of a Forest plot.Indian J Psychiatry 2015;57:313-314


How to cite this URL:
Shah N, Andrade C. A simple rearrangement can improve visual understanding of a Forest plot. Indian J Psychiatry [serial online] 2015 [cited 2021 Sep 25 ];57:313-314
Available from: https://www.indianjpsychiatry.org/text.asp?2015/57/3/313/166620


Full Text

By tradition, Forest plots in meta-analyses usually present studies either in alphabetical order, byfirst author name, or in chronological order, by year of publication. Alphabetical ordering allows readers to quickly locate a study in the plot. Chronological ordering allows readers to identify the evolution of results across time. Most readers only cursorily glance at the individual study data, focusing, instead, upon the diamond at the bottom of the plot, which represents the summary statistic. This, in a way, is how it should be; after all, the entire purpose of the meta-analysis is to generate that summary statistic.

A discerning reader also looks at the effect size and confidence intervals of the individual studies in the plot. This information educates the reader about the extent to which each study influenced the summary statistic. The information is useful because different studies would have differed in methodological rigor, or merely in some important aspect of methodology, such as sample size or drug dose. An idea of how studies might have influenced the summary statistic could guide the reader in his interpretation of that summary statistic, and in the possible need for additional sensitivity analyses.

The conventional arrangement of studies in the Forest plot does not assist this discerning reader in his efforts. We suggest that, instead, studies can be arranged in order of increasing effect size; that is, keeping the study with the smallest (or most negative) effect size at the top, and the study with the largest (or most positive) effect size at the bottom. This small rearrangement of data immediately allows the reader to see which study influenced the summary statistic in which direction and to what extent; it also allows the reader to see how the numerically smaller effect sizes counterbalanced the numerically larger effect sizes to generate the summary statistic.

We illustrate our point with data extracted from a published meta-analysis,[1] presented in two Forest plots. In thefirst plot [Figure 1], the studies are arranged alphabetically. In the second plot [Figure 2], the studies are arranged from top to bottom in the increasing order of the value of the standardized mean difference (SMD). Note that the second plot allows the reader to more easily and visually understand the significance and size of the individual SMDs, and their influence on the value of the summary SMD.{Figure 1}{Figure 2}

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.

References

1Andrade C, Kisely S, Monteiro I, Rao S. Antipsychotic augmentation with modafinil or armodafinil for negative symptoms of schizophrenia: Systematic review and meta-analysis of randomized controlled trials. J Psychiatr Res 2015;60:14-21.