Dealing with effect size multiplicity in systematic reviews and meta-analyses

Area of Science:

• Biostatistics
• Epidemiology
• Medical Research Methodology

Background:

• Systematic reviews frequently include studies reporting multiple effect sizes from the same participants, potentially causing statistical dependency in meta-analyses.
• This dependency can compromise the validity and reliability of meta-analytic results, necessitating careful methodological consideration.

Purpose of the Study:

• To provide a tutorial on managing effect size multiplicity within studies for meta-analyses.
• To recommend a structured, three-step approach for addressing statistical dependency arising from multiple effect sizes.
• To offer guidance on selecting appropriate meta-analytic strategies for intervention and association studies.

Main Methods:

• The tutorial proposes a three-step approach: 1. Define the research question (convergent vs. divergent approach).
• 2. Identify and categorize types of effect size multiplicities within the study database.
• 3. Select a suitable meta-analytic strategy (reductionist or integrative) based on the identified multiplicities.

Main Results:

• The paper categorizes different types of effect size multiplicities encountered in systematic reviews.
• It outlines available meta-analytic strategies, including reductionist (single effect size per study) and integrative (multiple effect sizes per study) approaches.
• Recommendations are provided to aid researchers in choosing the most appropriate strategy for their specific meta-analysis.

Conclusions:

• Effect size multiplicity is a common challenge in systematic reviews and meta-analyses that requires specific handling.
• The proposed three-step approach and strategy selection guidance can help researchers manage statistical dependency effectively.
• Adopting appropriate methods for dealing with multiplicity enhances the accuracy and interpretability of meta-analytic findings.