Optimal Planning and Scheduling in Outpatient Procedure Centers

BERG, BJORN P. Optimal Planning and Scheduling in Outpatient Procedure Centers. (Under the direction of Brian T. Denton.) Outpatient procedure centers (OPCs) are increasingly becoming an important component of health care delivery systems. They offer benefits such as high patient safety outcomes, low cost of care, and convenience to patients; they also introduce new challenges for planning and scheduling. These challenges arise due to the complex structure of activities in outpatient procedure centers and the high level of demand uncertainty. In this thesis, we propose new models and methods to optimize planning and scheduling decisions in the context of OPCs. The focus of the thesis is on the use of stochastic models to understand how the uncertainty in demand affects optimal decisions related to planning and scheduling in OPCs. The first model is a discrete event simulation model based on an endoscopy suite. The purpose of this model is to analyze the costs attributed to no-shows in the endoscopy suite. Further, we compare the benefits of overbooking methods with no-show mitigation strategies that are common in practice. The effects of overbooking methods on the daily performance of the endoscopy suite are presented along with implications for increasing patient access to care. Extending the overbooking framework analyzed in the discrete event simulation model, a two-stage stochastic integer program is formulated for the purpose of identifying optimal booking, sequencing, and scheduling decisions for outpatient procedure centers in the presence of no-shows and patient classes. A new decomposition-based solution method is presented along with results comparing the computational performance of several solution methods. Theoretical and numerical results are presented that provide insights into the structure of the optimal solution for larger problems. The benefits of optimal booking, sequencing, and scheduling decisions are demonstrated based on a case study of a real OPC. In contrast to demand uncertainty manifesting as no-shows, the uncertainty of urgent add-on procedures is considered in the final model. The model represents a new formulation of the dynamic extensible bin packing problem. We develop a multi-stage stochastic integer programming formulation to capture the decision making process of allocating routine and urgent procedures to procedure rooms. A comparison of the solution quality and computational performance for several decomposition-based solution methods is presented. Heuristics are developed based on algorithms for traditional bin packing. Our numerical results show that the easy-to-implement heuristics perform very well, obtaining optimal or near-optimal solutions for a high percentage of problem instances. c © Copyright 2012 by Bjorn P. Berg