Research Interests
My research is mainly methodologic, with particular interest in Bayesian Nonparametrics and Causal Inference. However, I find it extremely interesting to apply new methods to understand the real data world, such as environmental epidemiology and public health.
Nonparametric Bayesian Model for Heterogeneity in Causal Effects
an application on air pollution epidemiology
We introduce a novel Confounder-Dependent Bayesian Mixture Model, leveraging the flexibility of the dependent Dirichlet process, to characterize causal effect heterogeneity. Our method enables us to identify heterogeneous and mutually exclusive population groups defined by similar Group Average Treatment Effects in a data-driven way, and estimate and characterize the causal effects within each of the identified groups. We apply our method to claims data from Medicare enrollees in Texas, identifying mutually exclusive groups where the causal effects of PM2.5 on mortality are heterogeneous.
A Bayesian Nonparametric Method to Adjust for Unmeasured Confounding with Negative Controls
Negative controls have been introduced in the causal inference literature as a promising approach to account for unmeasured confounding bias. We develop a Bayesian nonparametric method to estimate a causal exposure-response function (CERF). We model the CERF as a mixture of linear models. This strategy offers the dual advantage of capturing the potential nonlinear shape of CERFs while maintaining computational efficiency. Additionally, it leverages closed-form results that hold under the linear model assumption.
Bayesian Nonparametrics for Principal Stratification with Continuous Post-Treatment Variables
Principal stratification provides a causal inference framework that allows adjustment for confounded post-treatment variables when comparing treatments. Although the literature has focused mainly on binary post-treatment variables, there is a growing interest in principal stratification involving continuous post-treatment variables. However, characterizing the latent principal strata with a continuous post-treatment presents a significant challenge, which is further complicated in observational studies where the treatment is not randomized. In this paper, we introduce the Confounders-Aware SHared atoms BAyesian mixture (CASBAH), a novel approach for principal stratification with continuous post-treatment variables that can be directly applied to observational studies. CASBAH leverages a dependent Dirichlet process, utilizing shared atoms across treatment levels, to effectively control for measured confounders and facilitate information sharing between treatment groups in the identification of principal strata membership.