Speakers: Xing Song, PhD

Xing Song, PhD

Assistant Professor, Biomedical Informatics, Biostatistics and Medical Epidemiology
University of Missouri

 

Dr. Xing Song’s work sits at the intersection of large-scale data infrastructure and real-world evidence analytics. She led the design and operation of secure, cloud-based platforms for multi-site EHR and claims data integration (the GPC Reusable Observable Unified Environment), supporting standardized ETL, governance, and reproducible analytics across networks. Leveraging this infrastructure, she has developed and deployed scalable machine learning and causal inference frameworks to generate robust, decision-relevant evidence from observational data. Her expertise spans building high-quality, cross-site data systems and translating those data into rigorous, longitudinal analyses of dynamic treatment strategies, addressing challenges of heterogeneity, bias, and model generalizability.

 

From Data Access to Causal Insight: Building a Secure Real-World Evidence Engine with GROUSE and GPC

Generating rigorous real-world evidence (RWE) requires more than data access, it demands tightly integrated infrastructure, governance, and analytic workflows. In this talk, I will present GROUSE (GPC Reusable Observable Unified Study Environment), a secure, cloud-based research platform developed within the Greater Plains Collaborative (GPC) to enable scalable, multi-site studies using standardized EHR data. I will highlight how GROUSE operationalizes a “data enclave” model with fine-grained while supporting reproducible ETL pipelines and federated analytics across PCORnet CDM sites. Beyond infrastructure, I will discuss how these capabilities support advanced analytic paradigms, including machine learning and causal inference. I will also share practical lessons from security reviews, compliance alignment (e.g., NIST 800-53), and building production-grade ETL workflows that integrate heterogeneous clinical and non-clinical data sources.