Hospitals are highly complex systems and as a result they provide researchers an endless stream of important, interesting, and challenging research problems. In many cases, a patient’s visit to a hospital typically starts in the emergency department (ED) or in the surgery unit as an elective patient. The patient goes through different stages of “service” during her stay in the hospital. Viewed from a distance, the ED-Hospital system can be seen as a complex network of queues and it is tempting to draw an analogy with a production system, which also can be seen as a network of queues. Nevertheless, while the analogy works up to a certain point and both types of systems are difficult to analyze, healthcare operations pose significant additional challenges. At the top of these challenges is the fact that a hospital typically has multiple objectives some of which might in fact be conflicting with each other. A hospital aims to give the best possible care to its patients – which is a vague goal by itself as it has to entail giving priority to patients with life threatening injuries along with delivering an overall short waiting times to the rest of the patients – but also reduce costs, make profits, and better compete with the other hospitals. There are complex interactions among the “servers.” A server might correspond to different entities (a bed, a diagnostic equipment, a nurse, a physician etc.) depending on the service stage and more than one type of server might be needed at the same time for the execution of any given part of service.
Argon and Ziya and their graduate and undergraduate students, in collaboration with faculty members Abhi Mehrotra from the UNC School of Medicine, Debbie Travers from the UNC School of Nursing, Jeffrey Strickler from UNC Healthcare, and Thomas Bohrmann and Kenneth Lopiano from Roundtable Analytics, have been involved in research projects that aim to better understand these complex systems, come up with the “right” models and tools for their analysis, and ultimately generate insights and develop methods that can help in making decisions and improving upon the current practice. More specifically, the project team is interested in questions like how one can measure emergency department and hospital crowdedness, how physician and nurse decisions are affected by census levels, how one can predict imminent bottlenecks for making better staffing and patient flow decisions, and how one can predict hospital admissions at triage and use this information when making better operational decisions. The team puts together methodologies from both operations research and statistics with expertise in emergency medicine in the service of providing answers to these questions.
Students who have been involved in this research project include Emily Riederer (undergraduate), Virgina Ahalt, names of two other Master’s students (Masters students), and Wanyi Chen (Ph.D. student). If you are interested in being involved in this research project, please send an email to Argon or Ziya for any potential opportunities.
Selected Publications: