Emergency Response Systems

 

Mass-casualty events such as natural disasters and terrorist attacks are not common occurrences but when they happen they can easily overwhelm emergency response systems in the affected area.   Together with their students, Professors Argon and Ziya have built an extensive research program on resource allocation problems in the aftermath of mass-casualty events supported by the National Science Foundation   (CMMI-0620736-0715020; CMMI-0927607, CMMI-1234212, CMMI-1635574). Our medical collaborators on these projects include Dr. James Winslow, Dr. Brian Hiestand, and Dr. Lane Smith of School of Medicine at Wake Forest University, Winston-Salem, NC.

Below are two examples of resource allocation problems that our research group has studied:

  1. Mass-casualty triage: Following a mass-casualty event, one of the most challenging decisions faced by emergency responders is to prioritize patients competing for scarce resources such as ambulances and operating rooms. In medical terminology, the practice of categorizing and prioritizing patients according to their health conditions is called triage. The most commonly used triage policy in the U.S. classifies patients into four criticality groups and largely ignores the level of resource limitations with respect to the size of the event. With the objective of developing effective resource-based triage rules, our research group formulated multiple mathematical models that capture the main tradeoff between prioritizing patients who are in more critical conditions and those who are either faster to save or more likely to survive. The analysis of our first model with patient abandonments (i.e., deaths) showed that taking into account the availability of resources and number of patients improves the expected number of survivors substantially [1]. Later, we developed an alternative model based on a fluid relaxation to incorporate more realistic features such as time-dependent survival probability functions [2, 3]. The resulting policy, which we call ReSTART, shows exceptional performance in realistic simulations. For an interactive demonstration of this policy, visit here.

 

  1. Patient distribution to hospitals: In the aftermath of a mass-casualty event, another major decision that heavily affects the outcome of the response effort is the distribution of casualties among multiple hospitals in the area. The most common practice is to transfer all patients to the nearest hospital, which is easy-to-implement and intuitive. However, such a simplistic approach to patient distribution proved to result in poor outcomes in past disasters. For a more effective policy, one needs to incorporate several factors into the decision such as travel times to hospitals, number of ambulances, hospital service capabilities, and hospital census levels. For this purpose, we formulated a mathematical decision making problem that led to dynamic policies that are not difficult to implement and are shown to be very effective in realistic simulations based on data from a U.S. trauma database and software developed by the U.S. government [4].

 

Selected Publications

  1. Uzun Jacobson, E., N. T. Argon, and S. Ziya, “Priority Assignment in Emergency Response,” Operations Research 60 (2012), No. 4, 813–832.
  2. Mills, A. F., N. T. Argon, and S. Ziya, “Resource-Based Patient Prioritization in Mass-Casualty Incidents,” Manufacturing and Service Operations Management 15 (2013), No. 3, 361–377.
  3. Mills, A. F., N. T. Argon, S. Ziya, B. Hiestand, and J. Winslow, “ReSTART: A Novel Framework for Resource-Based Triage in Mass-Casualty Events,” Journal of Special Operations Medicine 14 (2014), No. 1, 30–39.
  4. Mills, A. F., N. T. Argon, and S. Ziya, “Dynamic Distribution of Patients to Medical Facilities in the Aftermath of a Disaster.” Forthcoming in Operations Research.