Causal Inference Approaches in Childhood Trauma-Inflammation Research: A Monte Carlo Simulation Study exploring Adiposity

In psychoneuroimmunology research, controlling for adiposity is often done without consideration for the underlying data-generating process, potentially leading to underestimation of associations between childhood trauma and inflammation suppressing indirect pathways. We conducted a Monte Carlo simulation study using MIDUS-informed effect sizes to model a theorized causal structure (Childhood trauma → Adiposity → CRP) where trauma influences CRP indirectly via adiposity with no direct effect. Results showed models estimating the total association between childhood trauma and CRP produced effect estimates consistent with the simulated data-generating process, whereas models adjusting for adiposity yielded a null direct effect, reflecting statistical absorption of the mediating pathway. These findings illustrate how analytic adjustment decisions alter the estimand and, consequently, the biological interpretation of trauma–inflammation associations.