Sae-Hwan Park, PhD
Associate Director of Research;
Research Staff, Statistical Center for Translational Research in Medicine (SC-TRM)
Dr. Sae-Hwan Park is Associate Director of Research at Penn Medicine and a member of the Statistical Center for Translational Research in Medicine (SC-TRM). His work sits at the intersection of healthcare AI, statistical methodology, real-world data, and research software, with an emphasis on developing rigorous and reproducible methods for clinical and translational research.
His research focuses on clinical risk prediction, algorithmic fairness, health equity, and the evaluation of predictive models across heterogeneous patient populations. His current work includes methods for risk-distribution-adjusted model evaluation, race-neutral approaches to incorporating social determinants of health, analyses of bias and performance drift in healthcare algorithms, and methodological extensions of net benefit and decision curve analysis. He collaborates on research using VA, Medicare, and Penn Medicine data, as well as studies involving pragmatic clinical trials, wearable devices, and Hospital at Home care.
Dr. Park also develops open-source research software that translates statistical methods into reproducible tools for researchers. His projects include Metrics-Adjuster, for conventional and risk-distribution-adjusted evaluation of binary prediction models, and Risk-Bridge, for developing calibrated risk-prediction models using constrained maximum likelihood when available data sources differ in representativeness and information content. His broader software work includes risk adjustment, synthetic healthcare data, and reproducible infrastructure for working with CMS data and documentation. Dr. Park holds a PhD in Public Health Science with a concentration in Health Outcomes Research, an MSc in Policy Economics, and a bachelor’s degree in Electrical Engineering. He also earned a Master of Applied Science in Computer Science from the University of Pennsylvania. His interdisciplinary background informs an approach that combines statistical reasoning, clinical research, and software engineering to make methodological advances more usable, transparent, and reproducible.