Data-Driven Surgical Tray Optimization to Improve Operating Room Efficiency

Surgical procedures account for over 60% of the operating cost of a hospital. About 15% of these costs are related to surgical instruments and supplies. Hospitals spend several million dollars annually on instrument sterilization, instrument tray assembly, and instrument repurchase costs. However, in a large majority of hospitals, less than 20%–30% of reusable instruments supplied to a surgery are used on average. Prior implementations of surgical tray rationalizations have typically been expert driven. This is a labor-intensive effort typically focused on a small set of trays. On the other hand, past mathematical programming model-based studies have typically tested models with simulated data because of the difficulty in obtaining actual instrument-level usage data. We obtained actual surgical instrument usage at a large multispecialty hospital in partnership with OpFlow, a healthcare software company. We formulate a data-driven mathematical optimization model for surgical tray configuration and assignment with the goal of reducing costs of unused instruments, such as sterilization, instrument purchase, and tray assembly costs. We develop a solution methodology that scales to thousands of surgeries, thousands of instruments, and hundreds of surgical trays. This methodology decomposes the problem into a tray rationalization and an add-on tray creation step. At each step, we solve a restricted version of the full problem. We perform extensive out-of-sample testing of our solution. Our model-based approach identifies improvements in tray configuration and assignment, which would lead to a 54% reduction in unused instruments per surgery compared with the current tray configuration used at this hospital. We also validated our model with an expert-recommended solution for a subset of trays. We find that our model-based solution leads to 20% lower overage and 21% lower underage than the expert-recommended solution. We estimate projected annual cost savings of 35% in instrument sterilization, tray assembly costs, and instrument repurchase costs from using the recommendations of our model. Our solution was implemented at the UNC Rex Hospital, and we report on the results of our implementation. This analysis has quantified the value of collecting point-of-usage data to be at least $1.39 million per year from using the model-recommended solution at the hospital. Supplemental Material: The electronic companion is available at https://doi.org/10.1287/opre.2022.2426 .