Abstract:
|
Causal mediation analysis sheds light on how a treatment or exposure can affect an outcome of interest through one or more mediators on causal pathway. When multiple mediators on the causal pathway are causally ordered, it requires a sensitivity parameter to identify mediation effects on certain causal pathways. In this work a mixed model-based approach was proposed in the Bayesian framework to connect potential outcomes at different treatment levels and to identify mediation effects independent of such sensitivity parameter for natural direct and indirect effects on all causal pathways. The method was first proposed for continuous mediators and outcomes with possible mediator by treatment interactions, and later extended to handle more complex situations with nonlinear or mixed type of mediators and outcome. Sensitivity analysis was performed for prior choices in the Bayesian models in simulation studies. The proposed method was applied to an adolescence dental health study, to see how social economic status can affect dental caries through a sequence of causally ordered mediators in dental visit and oral hygiene index.
|