OVERVIEW
In recent years, biomedical and health research has entered an era characterized by unprecedented data richness and complexity. Large-scale electronic health records, multi-site clinical databases, digital health technologies, and real-world surveillance systems have created new opportunities to study disease mechanisms, treatment effects, and health system performance at population scale. At the same time, these data present fundamental challenges, including heterogeneity across sources, irregular observation processes, missing and error-prone outcomes, high-dimensional covariates, and limited ability to pool individual-level data. Addressing these challenges requires principled statistical foundations together with modern computational and algorithmic tools.
Our research develops statistical, causal, and machine learning methods to support reliable inference and decision making in complex biomedical and health settings. We contribute to foundational statistical inference theory under nonstandard conditions, communication-efficient federated learning and distributed inference, causal machine learning and AI, surrogate-powered inference for incomplete outcomes, and advanced meta-analysis and evidence synthesis. These methodological advances are tightly integrated with substantive applications in pharmacovigilance and vaccine safety, pediatric COVID-19 research, metabolic therapies, and telemedicine within learning health systems. Across these domains, our work emphasizes rigor, robustness, and scalability, with the goal of enabling trustworthy real-world evidence generation that informs clinical practice, regulatory decisions, and health system operation.
