A Two-Stage Method for Extending Inferences From a Collection of Trials

Area of Science:

• Biostatistics
• Epidemiology
• Clinical Trials

Background:

• Synthesizing evidence from multiple randomized controlled trials (RCTs) is crucial for understanding treatment effects.
• Traditional meta-analysis often yields non-causally interpretable estimates due to between-study heterogeneity.
• Existing methods struggle to provide causally interpretable average treatment effects in specific target populations.

Purpose of the Study:

• To propose a novel two-stage meta-analytic approach for obtaining causally interpretable average treatment effects.
• To address between-study heterogeneity in conditional average treatment effects.
• To extend inferences from multiple RCTs to a defined target population.

Main Methods:

• Developed a two-stage meta-analytic framework inspired by existing methods.
• Established assumptions for identifying target population average treatment effects with heterogeneous conditional effects.
• Introduced a two-stage weighted averaging estimator for study-specific treatment effect estimates.

Main Results:

• The proposed method provides causally interpretable estimates of average treatment effects in target populations.
• Simulation studies demonstrated the performance of the new approach.
• Applications included a Hepatitis-C trial and pediatric traumatic brain injury therapy studies.

Conclusions:

• The novel two-stage meta-analysis effectively synthesizes RCT data despite between-study heterogeneity.
• This method enhances the ability to derive causally interpretable treatment effect estimates for specific populations.
• The approach has practical implications for evidence-based medicine and clinical decision-making.