Page 26

Health System Management • December 2016

26 WWW.HEALTHSYSTEMMGMT.COM information, the proactive policies achieved shorter wait times at any patient traffic intensity. Imagine a scenario where the ED is about to encounter a “bursty” episode, during which there are more arrivals than service availability. Not having access to information about the future, a threshold policy will have to wait until the queue builds up to significant length before the ED starts to make diversions, after which the long queue is bound to cause large delays for subsequent arrivals. In contrast, by predicting the onset of a “bursty” episode beforehand, a proactive policy can make diversions earlier, and potentially prevent the queue from building up in the first place. Our model shows that such a proactive policy can achieve a substantially smaller queue length than that of a threshold policy at the same diversion rate. Instead of waiting for congestion to build up, it may be more effective for ED managers to be proactive and divert more aggressively towards the beginning of a “bursty” episode of ED arrivals, when there will be significantly more arrivals than service availability. DO THESE POLICIES WORK? The effectiveness of these proactive diversion policies were examined in the ED through statistical simulation, and showed that proactive policies consistently deliver up to a 15% reduction in average waiting time. Integrating predictive analytics into ED policy making can be an important is a step towards reducing the overcrowding and long wait times that plague many of today’s EDs, ensuring patients receive the care they need, when they need it. OPERATIONS HEALTH SYSTEM MANAGEMENT | DECEMBER | 2016 “Patients on their way to the emergency room want their health crisis to be handled expeditiously.” severity index from one through five, with one indicating the highest severity. Following triage, patients wait until they are taken into an examination room and assigned a bed. While in the examination room, the patient may interact with a physician and nurses, have blood drawn, be taken for various tests, and/or wait between any of these events. The number of patients going through this process can quickly escalate creating congestion. Finally, the patient will either be discharged from the ED or admitted to an inpatient unit. As the ED experiences more congestion, the need to manage patient crowding can increase the amount of time a hospital spends on diverting patients to other institutions. Many solutions have been suggested to address this overcrowding problem. Some hospitals have resorted to increasing bed capacity, or using queueing theory to improve staffing decisions. Others have tried to educate patients about when it’s appropriate to visit their primary care physician instead of going to the ED. Still another approach is diverting ambulances elsewhere when the ED is overcrowded. But even with the effective implementation of traditional diversion strategies, random variations can still result in periods of overcrowding. USE PREDICTIVE ANALYTICS FOR PROACTIVE POLICY DEVELOPMENT Predictive modeling can be used to reduce congestion in the ED and wait times for patients by leveraging future information when making admission decisions. A primary motivation in developing these predictive models has been to guide operational decision making or policy development such as staff roster and resource planning or decisions related to on-call staffing. However, while there has been substantial attention paid to developing such predictive models, there has been limited work demonstrating how they can best be utilized to improve system performance. Our research is an important first step towards a methodology to consider how predictive models of patient arrival counts could be used to create proactive diversion policies that may improve quality of care. Currently, hospital diversion decisions are typically made solely on information about current congestion, so if a maximum threshold is reached, new patients will be diverted. UTILIZING ALGORITHMS FOR DIVERSION DECISIONS We focused on creating an algorithm that utilizes predictions of future arrivals to the ED to make better diversion decisions. A major part of this work was dedicated to creating a framework for understanding when and how predictive information can be used to reduce waiting times. The research tested the design of a family of proactive diversion policies to leverage predictions of future patient arrival and service patterns. The results demonstrated that, given sufficient future WEBEXTRA For further suggestions on ways to improve the efficiency of your emergency room, read “How to Shorten Emergency Room Wait Time” at www.HealthSystemMgmt.com


Health System Management • December 2016
To see the actual publication please follow the link above