How can we better understand causal pathways when dealing with count data that has many zeros? This paper introduces a new mediation methodology for zero‐inflated count outcomes using the marginalized zero‐inflated Poisson (MZIP) model and the counterfactual approach to mediation. This novel work gives population‐average mediation effects whose variance can be estimated rapidly via delta method. This methodology addresses limitations of current methods that are computationally intensive, biased, or challenging to interpret. This methodology is extended to cases with exposure‐mediator interactions. Applied to explore if diabetes diagnosis can explain BMI differences in healthcare utilization, the proposed MZIP method minimizes bias and computation time compared to alternative approaches while allowing for straight-forward interpretations.
Published in Statistics in Medicine, this paper fits well within the journal's focus on statistical methods applied to medical research. The development of a new mediation methodology for zero-inflated count outcomes directly aligns with the journal’s objective to advance statistical techniques relevant to health sciences.