ORAHS2024
Abstract Submission

38. Overcoming the bottleneck: data-driven strategies and analytics of operating room management from an integrative view

Contributed abstract in session TA-5: Operating Room Scheduling /1, stream Regular talks.

Tuesday, 9:00-10:30
Room: Room S6

Authors (first author is the speaker)

1. Haoliang Pan
Industrial Engineering, Tsinghua University
2. Peng Gao
China-Japan Friendship Hospital
3. Xiaolei Xie
Industrial Engineering, Tsinghua University

Abstract

Operating rooms represent bottleneck units in most hospitals. Operating room managers are devoted to balancing the two contrasting goals of maximizing prime time utilization and minimizing overtime work. This task can be very difficult because it involves the uncertainty of surgery durations.
In this study, we develop a data-driven framework to solve the surgery scheduling problem faced by many large hospitals. We utilize machine learning methods to capture the nature of uncertainty of surgery durations. Then, we construct a stochastic optimization model that addresses the managerial goals of resource utilization and the constraints of resource overuse of operating rooms. The model is hard to solve because of its complex structure under uncertainty, especially the chance constraints. In the light of this, we develop a novel data-driven approach to obtain a polyhedral approximation of the feasible region of the original model.
By using real data from a large hospital in China, we find that our proposed approach outperforms classical sample average approximation approach. We emphasize that our approach demonstrates great potential in rapid decision-making scenarios, which are frequent and critical cases in healthcare systems. Our approach also yields managerial insights of quantitative relations among various operative goals of interrelated units in the hospital system, which could help hospital managers determine proper strategies and goals to pursue integrative benefits.

Keywords

Status: accepted

Back to the list of papers