Abstract
Massive casualty incidents (MCIs) such as natural disasters and accidents have drawn increasing attention recently. According to the official Japanese account, the tsunami on March 11, 2011 resulted in 15,839 dead, 5,950 injured, and 3,642 missing. In November 2015, terrorist attacks in Paris killed 132 people and injured more than 300. MCIs usually result in a large number of casualties and have significant effects on the community. The most important task for emergency departments is to provide timely treatment to these casualties by dispatching them to medical facilities.In practice, emergency responders follow policy and transport all patients to the closest hospitals. However, a large number of patients will overwhelm the capacity of the nearby medical facilities, creating a major challenge for emergency services. Considering the limited transportation and medical resources, a properly designed dispatch plan is vital to achieving a better chance of survival for casualties and to minimize their recovery cost.
The medical and operations research (OR) literature emphasizes the importance of taking limited resources into account and providing managerial insights into this problem. In this paper, consistent with the insights of previous work, we generate a framework approach for allocating limited resources by dispatching patients for timely treatment. The objective is to maximize “utility” based on each patient’s probability of survival and to reduce their recovery cost. Two essential decisions are suggested in our plan: the priority of patients for transport/treatment and the destination of the patients to be dispatched. This is a complex and challenging task requiring careful coordination of two critical resources: available ambulances and medical facility capacity.
We first remove the complexity due to uncertainty and develop two deterministic models that allow us to analyze how dispatching ambulances and prioritizing casualties affects the effectiveness of the overall plan. More specifically, in both models, we assume that the medical condition of patients can be predicted (a significant body of medical literature and technological advancements such as wearable devices have made such predictions increasingly accurate). In the first model, to analyze the effect of transportation resources, we formulate the exact transportation throughput: the number of ambulances and the travel time from the incident site to each involved hospital. We assume that the capacity of each hospital is constant and cannot be changed at the time we make the dispatch decision. We then develop a more sophisticated model based on the first by specifying hospital capacity as the number of operating rooms and taking into account the treatment time for each patient.
The deterministic model enables the analysis of the optimal structure of solutions and the development of effective heuristics based on such structure. We then analyze the effects of the three main sources of uncertainty on the robustness of the plan created by our heuristics: 1) the inaccuracy of the diagnosis of the severity of an injury; 2) uncertainty about the treatment time of a casualty; and 3) the uncertainty of travel time due to traffic conditions. A comprehensive set of problems is created using realistic settings to cover various common scenarios. An extensive numeric study is carried out to evaluate the effectiveness and robustness of our approach.
| Date of Award | 16 Jun 2016 |
|---|---|
| Original language | English |
| Awarding Institution |
|
| Supervisor | Biying SHOU (Supervisor) & Yanzhi David LI (Supervisor) |