Stochastic programming analysis and solutions to schedule overcrowded operating rooms in China

As a result of the growing demand for health services, China's large city hospitals have become markedly overstretched, resulting in delicate and complex operating room scheduling problems. While the operating rooms are struggling to meet demand, they face idle times because of (human) resources being pulled away for other urgent demands, and cancellations for economic and health reasons. In this research we analyze the resulting stochastic operating room scheduling problems, and the improvements attainable by scheduled cancellations to accommodate the large demand while avoiding the negative consequences of excessive overtime work. We present a three-stage recourse model which formalizes the scheduled cancellations and is anticipative to further uncertainty. We develop a solution method for this three-stage model which relies on the sample average approximation and the L-shaped method. The method exploits the structure of optimal solutions to speed up the optimization. Scheduled cancellations can significantly and substantially improve the operating room schedule when the costs of cancellations are close to the costs of overtime work. Moreover, the proposed methods illustrate how the adverse impact of cancellations (by patients) for economic and health reasons can be largely controlled. The (human) resource unavailability however is shown to cause a more than proportional loss of solution value for the surgery scheduling problems occurring in China's large city hospitals, even when applying the proposed solution techniques, and requires different management measures. HighlightsPresents multi-stage stochastic programming models the complex stochastic nature of operating room scheduling in overcrowded Chinese big city hospitals.Develops mathematical results and techniques to solve the formulated models to almost optimality.Attains substantial and significant schedule improvements for single operating room instances derived from real life data.