Introduction to methodology of dose-response meta-analysis for binary outcome: With application on software

Area of Science:

• Biostatistics
• Epidemiology
• Health Research Methodology

Background:

• Dose-response meta-analysis (DRMA) is a statistical technique used for over 30 years.
• DRMA is increasingly vital in healthcare for clinical decision-making and understanding exposure-outcome relationships.

Purpose of the Study:

• To provide an overview of the methodologies employed in dose-response meta-analysis.
• To illustrate the application of common DRMA methods using a practical example.

Main Methods:

• Summarizes commonly used regression models for fitting dose-response curves: linear, piecewise, natural polynomial, fractional polynomial, and restricted cubic spline regression.
• Explains two pooling approaches: one-stage and two-stage methods for combining results across studies.
• Demonstrates DRMA application with a case example.

Main Results:

• The study illustrated five distinct regression models to analyze dose-response relationships.
• Both one-stage and two-stage pooling approaches were presented for meta-analyzing dose-response data.
• The example analysis indicated comparable results across the different regression models.

Conclusions:

• Multiple DRMA methods are available for examining the link between exposure levels and outcome risks.
• Further methodological advancements are needed to enhance the utility and precision of DRMA.