A Review of Methods for Research Synthesis

Area of Science:

• Statistics
• Biostatistics
• Epidemiology

Background:

• Meta-analysis aggregates research using statistical methods.
• Fixed and random effects models are dominant but assume study homogeneity.
• Research heterogeneity (e.g., design, population) necessitates advanced methods.

Purpose of the Study:

• To review advanced meta-analysis techniques beyond standard models.
• To describe the theoretical basis and applications of these methods.
• To discuss the advantages, limitations, and potential combinations of advanced methods.

Main Methods:

• Review of hierarchical models.
• Discussion of bias adjustment and quality weighting methods.
• Exploration of Bayesian meta-analysis and decision-centered approaches.

Main Results:

• Advanced methods address heterogeneity in meta-analysis.
• Techniques include hierarchical, bias adjustment, quality weighting, Bayesian, and decision-centered approaches.
• Each method offers unique benefits and drawbacks for summarizing diverse research.

Conclusions:

• Advanced meta-analysis methods are crucial for handling research heterogeneity.
• The choice of method depends on study characteristics and research questions.
• Combining methods may offer more robust and comprehensive results.