TITLE:  Revenue Management for Outpatient Care Service

ABSTRACT:

The Affordable Care Act expands health insurance coverage for Americans, and will increase patient volumes and visits significantly.  However, a tremendous amount of provider time is wasted due to the inefficiency of appointment scheduling systems currently in use.  Our research proposes and demonstrates that, by factoring patient choice behavior into the design of appointment scheduling systems, providers can significantly improve their service efficiency with only minimal changes to their current practice.  Specifically, we develop models that account for heterogeneous patient preferences in appointment choice offering.  The provider has a certain number of appointment slots to fill in a day.  We consider two possible ways that the provider may interact with patients.  First, the provider offers a set of appointment choices to patients, who then choose according to their preferences.  This resembles an online appointment scheduling system (e.g., zocdoc.com), which has become increasingly popular among patients and providers.  Second, the scheduler may sequentially offer several sets of appointment choices.  Imagine in the telephone scheduling context, the scheduler may offer additional choices to patients if the initial offers are not accepted.  This may also work for online settings designed in a way that patients do not see all choices in their first attempt of booking appointments.  We will discuss the forms and structures of the optimal non-sequential offering and sequential offering policies.  We also propose effective heuristics to solve large-scale instances and for practical use.  We estimate that, based on medical reimbursement data, the efficiency gains by adopting these policies as opposed to ignoring patient choice behavior can translate into additional revenues of $26-150k per year for a primary care provider that sees 30 patients a day.  Theoretically, our work is related to assortment planning in revenue management, bipartite matching in graph theory and admission control in queueing theory.  This is joint work with Peter van de Ven (CWI, Netherlands) and Bo Zhang (IBM Watson Research Center). 

Bio: Nan Liu is an Assistant Professor of Health Policy and Management at Columbia University’s Mailman School of Public Health.  His research focuses on stochastic models, dynamic decision making and applied statistical analysis, with applications to the (health) service industry.  Much of his work bridges data analytics and systems modeling, and has focused on the design and control of service systems facing complex customer behaviors.  His recent work investigates customer preference and choice in the healthcare market and their impact on healthcare management.  In addition, he has done extensive work on health policy research using data-driven operations modeling.  His research has been published in leading journals in both fields of operations management and health administration/policy, including Manufacturing & Service Operations Management, Operations Research, Production and Operations Management, Health Services Research, Medical Care Research and Review, and Public Administration Review.  His work has received a wide range of media attention including the Washington Post and Crain's New York Business.  He received a third prize in the INFORMS Junior Faculty Interest Group (JFIG) Paper Competition and the Calderone Junior Faculty Research Prize awarded by Mailman School of Public Health.  Dr. Liu has a B.E. in Civil Engineering from Tsinghua University (Beijing, China) and receives his Ph.D. in Operations Research and M.S. in Statistics from the University of North Carolina at Chapel Hill.