Drug safety meta-analysis: promises and pitfalls

Area of Science:

• Pharmacovigilance
• Biostatistics
• Clinical Research

Background:

• Meta-analysis is crucial for detecting drug and vaccine adverse effects, especially rare ones not evident in small efficacy studies.
• However, meta-analysis results in this field are often controversial due to various limitations.

Purpose of the Study:

• To examine the challenges and controversies in using meta-analysis for identifying adverse drug and vaccine reactions.
• To propose methods for improving the reliability and precision of meta-analysis in pharmacovigilance.

Main Methods:

• Review of meta-analysis applications in drug and vaccine safety studies.
• Discussion of statistical challenges including rare events, heterogeneity, and inconsistent reporting.
• Exploration of advanced statistical approaches like random-effects models and multiple model testing.

Main Results:

• Meta-analyses can yield different conclusions due to inherent data limitations.
• Rare adverse events, clinical/methodological heterogeneity, and incomplete reporting complicate standard meta-analysis.
• Results' precision may be overestimated.

Conclusions:

• Standard meta-analysis methods may be inadequate for assessing drug and vaccine adverse effects.
• Analysts should explore heterogeneity, employ advanced statistical models (e.g., random-effects), and use multiple models to assess result robustness.