Using SAS Macros for Multiple Mediation Analysis in R

Area of Science:

• Biostatistics
• Statistical Modeling
• Computational Statistics

Background:

• Mediation analysis assesses third-variable effects on exposure-response relationships.
• Generalized linear models (GLM) are standard but limited for nonlinearities and interactions.

Purpose of the Study:

• To develop SAS macros for implementing advanced mediation analysis.
• To extend the capabilities of the R package 'mma' to the SAS environment.

Main Methods:

• Developed SAS macros to interface with the 'mma' R package.
• Utilized generalized linear models, multiple additive regression trees, and smoothing splines.
• Applied the method to perform mediation analysis within SAS.

Main Results:

• Successfully created SAS macros for mediation analysis.
• Enabled the use of advanced statistical techniques (regression trees, splines) in SAS.
• Facilitated the analysis of complex exposure-response relationships.

Conclusions:

• The developed SAS macros provide a powerful tool for mediation analysis.
• This integration enhances statistical capabilities for researchers using SAS.
• The method accommodates nonlinear relationships and interactions effectively.