Tutorial: How to Read a Forest Plot

Introduction

• Forest plots are graphical representations of data from multiple studies used in systematic reviews and meta-analyses.
• They help summarize and compare results from different papers asking the same question.

Learning Objectives

• Understand the purpose of forest plots.
• Learn to read forest plots and interpret individual study results and averaged results.
• Comprehend the appearance variations in forest plots based on analyzed statistics.
• Recognize the importance of heterogeneity in forest plots and its impact on interpretation.

Why Use Forest Plots?

• Forest plots consolidate data from many studies into one graph, allowing for straightforward comparison.
• They identify a common statistic across various studies, facilitating an overall view of the data.

Analyzing Forest Plots

Part 1: The Axis

• Horizontal axis represents statistics (e.g., odds ratio, relative risk).
• Vertical line is the "line of null effect"; no difference or association exists at this line.
  • Relative statistics (OR, RR) null value: 1
  • Absolute statistics (ARR, SMD) null value: 0

Part 2: The Study Lines

• Each horizontal line represents a study; black box = point estimate and study size.
• Horizontal line length = 95% confidence interval (CI).
• Null effect line crossing indicates non-significant results.
• Larger studies: smaller confidence interval and larger black box.

Part 3: Combining Studies

• Diamond represents averaged results across studies.
• Vertical points of diamond = point estimate; horizontal points = 95% CI.
• Diamond crossing null effect line implies non-significance.

Part 4: Contextual Information

• Left of plot: author names and publication years.
• Immediate left: outcome numbers (n = event occurrences; N = total sample size).
• Far right: numerical forest plot summary (point estimate and CI).
• Bold statistics associated with combined results.

Part 5: Heterogeneity

• Consistency among studies checked using I² statistic.
• I² < 50% indicates acceptable consistency; >50% suggests potential non-chance inconsistency.

Summary of Key Points

1. Horizontal line = individual study; black box = result and size.
2. Side of vertical line a study falls on indicates implication of result.
3. Crossing vertical line implies non-significance.
4. Diamond = combined study results; interpret like other study lines.
5. I² statistic measures study consistency; high value = questionable conclusions.

Additional Examples

• Cochrane Collaboration logo is a representation of a forest plot.
• Original systematic review: steroids reduce preterm birth complications.

References

• Roberts D, Dalziel SR. Antenatal corticosteroids for fetal lung maturation. Cochrane Database of Systematic Reviews, 2006.
• University of Oxford: Interpreting forest plots and heterogeneity.
• Cochrane Collaboration logo origin.

---

For further practice, examine additional forest plots and apply the outlined interpretation steps.