Online Program Home
My Program

Abstract Details

Activity Number: 168 - Causal Inference
Type: Contributed
Date/Time: Monday, July 30, 2018 : 10:30 AM to 12:20 PM
Sponsor: Biometrics Section
Abstract #330231 Presentation
Title: Generalized Causal Mediation and Path Analysis Using the R Gmediation Package
Author(s): Jeffrey Albert*
Companies: Case Western Reserve University
Keywords: Causal Inference; Dental Caries; Mediation formula; Potential Outcome

Causal mediation analysis involves the decomposition of the effect of a treatment or exposure into causally interpretable indirect (through a mediator) and direct effects. The basic single-mediator situation has been extended to allow multiple ordered or unordered mediators. Recently, a generalized causal mediation and path analysis methodology has been developed, allowing for a sequence (two stages) of mediators with multiple mediators at each stage. Outcomes (including mediators) may be of different types following generalized linear models, and unsaturated models and clustered data are accommodated. This methodology can be implemented using a new R package called gmediation. Following an overview of the statistical method, the use of the package will be explained and illustrated using data from a dental caries cohort study. Possible future directions for causal mediation analysis and the gmediation package will be discussed.

Authors who are presenting talks have a * after their name.

Back to the full JSM 2018 program