Optimizing Hospital Surgery Schedules with Robust Mathematical Modeling
Hospitals face a constant balancing act when scheduling surgeries. Unexpected delays or faster-than-anticipated recoveries can create a ripple effect, disrupting subsequent procedures and straining limited resources like operating rooms and ICU beds. This complexity, explained Dr. [Shehadeh], a professor in the daniel J.Epstein Department of Industrial and Systems Engineering, can lead to difficult choices for hospital managers – prematurely discharging patients or even cancelling surgeries, both with potentially negative consequences for patient care.
Recognizing this challenge, Dr. Shehadeh and a team of researchers from Carnegie Mellon University, Texas Tech University, and the Medical University of South Carolina developed a novel approach to surgery scheduling. They collaborated with a hospital, analyzing real-world surgery data and case studies to build complex mathematical models. These models utilize a technique called distributionally robust optimization (DRO),a powerful method that acknowledges the inherent uncertainty in predicting surgery duration and post-operative recovery times.
Rather of relying on perfect forecasts, the DRO approach prepares for a range of possibilities. “It’s a mathematical model that can make all of these decisions – how many surgeries to schedule, when, and where – considering operating room capacity, ICU and ward availability, and the variability in surgery length and recovery,” Dr. Shehadeh explained.
The potential impact of this integrated scheduling system is important. Early results suggest hospitals could reduce operational costs by 24% to 60%. Beyond cost savings, the models promise tangible benefits for patients. By creating more reliable schedules, they drastically reduce the chances of last-minute cancellations due to bed shortages and minimize delays in the operating room - a major source of anxiety for patients and their families. Furthermore,the system could improve overall access to surgical procedures.
The research also highlighted a crucial trade-off: maximizing surgery volume doesn’t necessarily equate to optimal performance. Scheduling too many procedures can overwhelm recovery units, leading to further disruptions. the team’s models empower hospital administrators to find the ideal balance based on their specific resource constraints.
“Our findings offer valuable insights…and demonstrate the practical impact of our integrated approaches,” noted Rema Padman, Trustees Professor of management Science and Healthcare Informatics at carnegie Mellon university’s Heinz College, a co-author on the study.
The next phase of the project focuses on translating these powerful models into user-pleasant software tools for hospitals. ”Implementation is a big challenge,especially in healthcare,” Dr. Shehadeh acknowledged. “Our goal is to continue collaborating with health systems to make this accessible and create a decision support tool where they can input their data and generate optimized schedules.”
