Collaborative Causal and Bayesian Modeling
Average cost reductions across multiple healthcare cost types - fit using a Bayesian hierarchical zero-inflated model that partially pools estimates across types.
This is a collection of some applied work with particularly interesting Bayesian and/or causal modeling. For instance, Hubbard et al. use a Bayesian mixture model to infer patients' unknown diabetes status in noisy EHR data. Harrison et al. use a Bayesian hierarchical zero-inflated model to assess difference in costs associated with a cost-lowering healthcare intervention. Takvorian et al. use a difference-in-differences strategy to assess the impact of medicare expansion under ACA on cancer treatment delivery.
