This field-based case was developed through interviews with individuals at the University of Virginia (UVA) Health System. It follows John Ainsworth, administrator of analytics, as he and his team tackle inefficiencies in the hospital’s discharge process. The narrative describes their efforts to use machine learning to predict patient discharge types, thereby reducing patient stay durations and improving bed utilization. The case offers a detailed exploration of the hospital’s operational challenges and the potential of data science to drive significant improvements in health care delivery.

Students will explore techniques to handle datasets with over 800 features, many of which are not easily interpretable. The primary learning objectives are building and evaluating both binary and multi-class classification models to improve hospital operational efficiency. This case was taught at the UVA Darden School of Business toward the end of a second-year MBA elective called “Machine Learning and AI in Business” and is also well-suited to be taught after an introduction to neural networks. Students should be familiar with concepts such as basic classification techniques and out-of-sample testing.

This case was taught using Jupyter notebooks, which are provided as instructor supplements and will require Jupyter software to use them.