Abstract

Sepsis is a life-threatening condition, caused by the body’s extreme response to an infection. Worldwide, Sepsis remains to be one of the leading causes of death, contributing to 19.7% total global deaths. In the United States, 1.7 million cases of sepsis occur annually, resulting in 265,000 deaths. Delayed diagnosis and treatment are associated with higher mortality rates. The exponential rise in the availability of medical data has allowed for the development of machine learning algorithms to predict sepsis earlier than the onset. However, these models often underperform, generating excessive false alarms when implemented in real time. In this project, we aim to improve early sepsis predictions by augmenting machine learning algorithms with sequential decision making models. Hierarchical structures are adopted for the developed frameworks. Prediction outcomes from two distinct patient cohorts are measured through a set of temporal metrics featuring real-time implementations. The hierarchical frameworks produce interpretable results with improved precision and reduced false alarms that alleviates workloads of the bedside staff.

Biography

Zeyu Liu is an Assistant Professor in the Department of Industrial and Management Systems Engineering at West Virginia University. He received his Ph.D. in Industrial Engineering from The University of Tennessee, Knoxville in 2022. His research interests include optimization under uncertainty, data-driven analytics, reinforcement learning, and agent-based simulation, with applications in healthcare, infrastructure resilience, energy systems, transportation systems, and space logistics. Dr. Liu received the “Harvey J. Greenberg Research Award” from the INFORMS Computing Society in 2022. He is a member of INFORMS and IISE.