Di Shu, PhD
Di Shu is an Assistant Professor of Biostatistics in the Department of Biostatistics, Epidemiology and Informatics in the Perelman School of Medicine at the University of Pennsylvania. Her research focuses on developing and applying suitable statistical methods to assess comparative safety and effectiveness of drugs and other medical products with real-world data — including electronic health records and administrative claims — which may have complex structures that complicate statistical analysis.
To address concerns about misspecification of the treatment decision process using a single propensity score model, Dr. Shu and her colleagues have developed a robust method that allows for using a set of propensity score models simultaneously. The resulting estimators of causal effect measures achieve statistical consistency when the set of propensity score models contains a correct one. They also have developed a one-step method to allow data partners to share only summary-level risk set tables to estimate overall and site-specific hazard ratios in distributed data network studies. This method has been implemented as part of the routine querying tools in Sentinel, the U.S. Food and Drug Administration's national medical product safety surveillance system. To correct for outcome misclassification, a common data quality issue, they derived a closed-form, bias-corrected estimator of causal relative risk as well as an efficient method using validation data.
Dr. Shu is also passionate about programming with R. She and her colleagues have written three packages that are publicly available on the Comprehensive R Archive Network.
Assistant Professor of Biostatistics in Pediatrics
Associate Director for Biostatistics, Center for Pediatric Clinical Effectiveness (CPCE)
The Children’s Hospital of Philadelphia
Causal inference, measurement error, privacy-protecting methods