报告题目：Cross-Cluster Referral and Opt-out Incentive in Healthcare Systems
报 告 人：Weixin SHANG（Lingnan University，Associate Professor）
主 持 人：傅科（中山大学岭南学院，教授）
时 间： 2018年12月27日（周四）下午16:30-17:30
语 言： 中文+英文
We study the problem of long patient waiting times in a public hospital system having several hospital clusters. These hospitals provide identical services together with outside private hospitals that provide the same services but for substantial fees. We examine two approaches that can be used to enhance system performance. Cross-cluster referral and opt-out incentive, which can be observed in some healthcare practice but are not well studied in the literature, relate to economics of queues and, in our practical hospital system setting, exhibit properties that are new to the literature and can be used to help policy development and operations quality improvement in public healthcare systems. We construct a multi-cluster service network to study patients' individual decisions, and a Markov decision process model to investigate the optimal patient routing policy under centralized control. We also examine analytically and numerically how the two approaches can be applied effectively. We find that, though the option of cross-cluster referral in general can improve system performance, it may, under certain conditions, worsen system performance, which resembles Braess's paradox examined extensively in the economics and transportation literature. Opt-out incentive is well justified for the societal well being because compared with a socially optimized system, an individually optimized system admit more patients due to negative externalities but which can be mitigated by opt-out incentive. However, unlike the single-node case where the system can be perfectly coordinated by charging a toll to an entering customer, a static incentive scheme may not fully coordinate the multi-cluster network due to the cross-cluster effect of externalities generated by the switching option. Cross-cluster referral is effective when the cost of receiving service from outside the system is very high, whereas the system workload is intermediate and relatively imbalanced; good operational control is needed to avoid the trap of over-facilitated referral which would worsen the overall system performance. While not completely eliminating negative externalities, a carefully chosen fixed opt-out subsidy is publicly acceptable and can induce self-optimizing patients to behave very closely to the way desired in a socially optimal system, and this mechanism is especially effective when system utilization is sufficiently high.
Weixin Shang is an Associate Professor at Faculty of Business, Lingnan University (Hong Kong). He holds a Ph.D. degree from the Department of Industrial Engineering and Logistics Management, Hong Kong University of Science and Technology. His research interests include service operations management, information management in supply chains, and economic models in supply chain management. His work has been published in Management Science, Manufacturing and Service Operations Management, and Production and Operations Management.