Data-driven robust admission and discharge control for hospitals with readmissions

• Applicant: Professor Dr. Jingui Xie, Ph.D.
• Subject Area: Operations Management and Computer Science for Business Administration ; Public Health, Healthcare Research, Social and Occupational Medicine
• Term: since 2024
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 543063591

This project seeks to advance a dynamic, data-driven support system for medical decision-making, enhancing the management of hospital admissions and discharge processes when considering readmissions. It aims to refine patient readmission risk prediction by addressing intrinsic challenges with new instrumental variables (IVs) in predictive analytics, moving beyond traditional IVs such as the admission day-of-week, which may not be as effective for elective patients. The cornerstone of the project is a dynamic optimization model informed by a 'riskiness index,' which will strategically schedule admissions and evaluate potential early discharges against the backdrop of ward capacity constraints. This model is designed to harmonize the immediate demands of hospital workload with the longer-term risks associated with patient readmission, a calculation that prior research has not fully integrated. Acknowledging that no predictive model is immune to estimation errors, our approach incorporates a robustness optimization framework to fortify the decision-making model against such inaccuracies, thereby enhancing the reliability of outcomes. Different from the traditional predict-then-optimize method, we aim to develop a framework which can seamless combine predictive and prescriptive analytics. The project will culminate in the development of a data-driven simulation to rigorously assess the proposed models' effectiveness compared to established benchmarks in the field. This comprehensive approach aims to deliver a robust decision-making framework that can significantly improve operational efficiency and patient care in hospitals.

• DFG Programme: Research Grants