Improving Patient flow in Acute Care

* 在处理卫生保健管理问题时，我们经常被迫将卫生系统划分开来，以便将挑战简化为易于处理的模型。这种划分未能考虑到病人通过急性护理流程的复杂性。近似动态规划（ADP）在解决大规模序贯决策问题中的成功意味着某些划分不再是必要的。本研究的第一个目标是合并一些传统的隔间在健康的OR模型，以包括这些决定的高度依赖性。我们开始的模型，结合先进的调度和预约调度问题。这些问题传统上被孤立地处理，但实际上高度依赖于约会调度问题的输入是来自高级调度模型的输出，并且高级调度模型的性能取决于约会调度的形式。然后，我们将通过两种方式增加这个问题的复杂性：首先，通过考虑顺序消耗的多个资源（具体参考手术调度，患者消耗手术室时间，然后在病房恢复时间），其次，通过纳入容量决策。然而，即使这样也涉及显著的划分，因为手术时间表不仅影响手术患者，而且影响通过急诊科入院的其他患者，因为两种类型的患者竞争相同的床位。因此，我们将寻求建立预测模型，可以预测的影响，医院拥挤的实施从ADP模型手术调度的政策。最后，由于社区能力不足，病人出院的任何延误都严重阻碍了这些模式的影响。因此，我们会继续探讨社区服务网络的容量规划模式，以确保病人能适时出院。因此，我们的建议有三个方面的目标：*1）建立一个模型，将高级调度和预约调度问题结合到一个模型中，可以处理顺序消耗的多个资源以及容量规划决策。2)建立一个模型来预测医院内的拥挤情况，并帮助评估目标1的模型输出。* 3)建立社区护理服务能力规划模型（排队理论和优化），以保持急诊护理的适当流出。因此，拟议的研究将开发一系列相互关联的模型，这些模型共同允许卫生当局确定从进入急性护理到出院到社区服务的整个护理路径沿着的必要能力，同时还提供智能调度政策，以确保患者服务的及时性。

***In dealing with health care managerial problems, we are often forced to compartmentalize the health system in order to simplify the challenges into a tractable model. This compartmentalization fails to take into account the complexity involved in patient flow through acute care. The success of approximate dynamic programming (ADP) in solving large scale sequential decision problems means that some of that compartmentalization is no longer necessary. The first objective of this research is to merge some of the traditional compartments in OR models in health in order to include the highly dependent nature of these decisions. We begin with a model that incorporates the advanced scheduling and appointment scheduling problems. These problems have traditionally been dealt with in isolation but are in truth highly dependent with the input of the appointment scheduling problem being the output from an advanced scheduling model and the performance of the advanced scheduling model depending on the form of the appointment schedule. We will then increase the complexity of this problem in two ways: first by considering multiple resources consumed in sequence (with specific reference to surgical scheduling where patients consume operating room time followed by recovery time in a ward) and second by incorporating capacity decisions. However even this involves significant compartmentalization as the surgical schedule impacts not only surgical patients but also other patients admitted through the emergency department as both types of patients compete for the same beds. We will thus look to build predictive models that can predict the impact on hospital congestion of implementing the policies derived from the ADP models for surgical scheduling. Finally, the impact of these models is significantly hampered by any delays in discharging patients from the hospital due to a lack of capacity in the community. We will thus continue to explore the capacity planning models for the network of community services that ensure timely discharge from the hospital. Thus our proposal has a three pronged objective:***1) Build a model that combines the advanced scheduling and appointment scheduling problems into one model and that can handle multiple resources consumed in sequence as well as capacity planning decisions.***2) Build a model to predict congestion within a hospital and help assess the output of the model from objective 1.***3) Build a capacity planning model for community care services (queuing theory and optimization) in order to maintain proper flow out of acute care.***Thus, the proposed research will develop a series of interconnected models that together allow a health authority to determine the necessary capacity along the entire care pathway from entry into acute care through to discharge to community services while also providing intelligent scheduling policies that ensure the timeliness of patient service.**