Meta-analysis as the core of evidence-based behavioral medicine: tools and pitfalls of a statistical approach

Abstract

Purpose of review: Meta-analyses have a significant impact on clinical decision-making. In behavioral medicine, they are regularly used to derive clinical practice guidelines. Owing to an increasing complexity of data integration methods used, their interpretation has become a challenge to clinicians and researchers. This review aims to provide an accessible introduction to recent advances in the methodology and reporting style of quantitative reviews within the field of behavioral medicine.

Recent findings: Meta-analytical findings are sensitive to the precise methods used. State-of-the-art criteria for reporting of clinical trials and meta-analyses have been put forward. Further adaptations of these criteria result from recent meta-analyses of the placebo effect in clinical trials. There is a need to carefully develop and evaluate methods to deal with patient dropout and missing data. Useful methods to evaluate diagnostic test accuracy and meta-analytically evaluate direct and indirect treatment comparisons have recently been developed.

Summary: In the interdisciplinary field of behavioral medicine, meta-analyses are applied to evaluate interventions, diagnostic instruments, and procedures, and to derive evidence-based treatment recommendations. Criteria to improve reporting quality and methods to control for potential biases have been adapted to meet the requirements of research in behavioral medicine today.