Increasing Operating Room Profits and Decreasing Wait Lists by Use of a Data-Driven Overbooking Model

Background and Objectives: Operating rooms (ORs) are precious resources in hospitals, as they constitute more than 40% of the hospital revenues. As such, surgical cancellations are very costly to hospitals. Same-day surgery cancellations or no-shows were found to be the main contributing factor to underutilization of ORs in a public-sector hospital despite the existence of long surgical wait lists. Methods: To demonstrate the feasibility of overbooking surgical procedures, a Monte-Carlo simulation model to predict unused o‎r time was built and validated using 6 months of historical data. We first fitted statistical distributions to the random parameters, using Easyfit© software. Then, a surgical case-mix optimization problem was formulated to prescribe the surgeries for overbooking that maximize the profits over the predicted unused o‎r time. The optimization model was solved both deterministically based on average historical surgery durations using Lingo software package and heuristically by taking into account the random nature of surgical durations using Risk Solver Platform© add-on to Microsoft Excel. Findings: We conducted simulation-optimization of the stochastic model for the three selected dates. Results show significant improvements over the base case of no overbooking. The increases in the average surgical profits due to overbooking on three randomly selected dates were 89%, 36%, 93%. Conclusions: To the best of our knowledge, this study is the first applied demonstration of descriptive, predictive, and prescriptive analytics to improve surgical operations through overbooking. The analysis shows significant opportunities in generating surgical profits (36%-93% on randomly selected validation dates) and reducing average wait times (by 4.16 weeks) without risking o‎r overtime.