Configuring surgical instrument trays to reduce costs

摘要

Most hospitals in the United States provide and manage significant inventories of durable surgical instruments used in operating rooms. The sheer volume and variety of instruments introduces considerable complexity in ensuring that the right instruments are available at the right time. Surgical instruments are usually stored and delivered to an operating room (OR) as procedure-specific sets of trays with multiple instruments included in a single tray. Because procedure trays are used by multiple surgeons trained at different institutions, procedure trays often include surgeon-specific instruments. Hospital materials managers and surgeons appear to weigh differently the various attributes of different tray configurations. Materials managers want to limit the cost of inventory and the variety of trays. Surgeons, on the other hand, prefer trays with the minimum number of unneeded instruments. Clearly, the kitting of surgical instruments into trays has many benefits, yet the actual tray design is a complex combinatorial problem. We propose a mathematical programming formulation to decide on the composition of trays to minimize the costs of owning, maintaining, and using both the trays and the instruments. We show that the optimal configuration depends not only on physician instrument preferences but also on the actual operating rooms' schedules. This dependency implies that changing surgery schedules can have a significant impact on how trays should be configured. Our numerical experiments suggest that currently, hospital materials managers overestimate the cost of tray variety and underestimate the cost of re-processing the extra instruments in a tray. Using real-world hospital data, we demonstrate that optimizing trays can result in substantial cost savings for the hospital while increasing surgeons' satisfaction. We introduce a fast heuristic algorithm for finding a near-optimal low-cost tray configuration given surgeons' preferences and surgical schedules.