医院紧急救助系统建模分析研究

The top priority of healthcare operations management is always to ensure patient safety. The astonishingly high number of annual preventable death prompted an initiative to create and deploy rapid response team (RRT) to quickly react to patients with clinical signs of deterioration to reduce the frequency and severity of negative outcomes. However, there is a lack of strong evidence that implementation of RRT will achieve that goal. Although RRT plays a critical role in hospital rapid response, collaborative and integrated operations involving care providers from different department should be examined, thus the study of rapid response system (RRS) is necessary. An analytical modeling framework is necessary to uncover the underlying principle of such process. Since such models do not exist in the current literature, the goal of this project is intended to contribute to this end. Under the analytical framework, multi-patient network with care provider resource sharing is studied. Mathematical methods are developed to obtain three important performance measures, which are mean decision time, CV of decision time and the probability that care delivery is carried out, response time performance (RTP).Using the three measures, a bottleneck analysis method is introduced to identify and improve the process whose improvement can lead to the largest improvement in those measures, respectively. The derived bottleneck indicator can quickly identify those bottlenecks using the data from floor data. Lastly, optimal design problem is studied given the systems parameters to reduce the mean and cv of decision time. Also optimal configuration of those parameters is examined given that RTP is greater than a desired value. In summary, this project studies the hospital rapid response in systems perspectives. Not only the obtained result can help build a continuous improvement system, but the derived methods to the scientific problems in the complex network can provide insights in the systems with similar underlying structures.

医院运作管理的首要目标是保障医疗安全。每年极高的可避免死亡数量促使医院建立和实施快速救助小组(RRT)用来救助体征突然恶化的患者，尽管RRT在紧急救助中起到的关键的作用，RRT所在的涉及多科室的医院紧急救助系统(RRS)更为重要，而文献中缺乏该类研究。本课题拟建立量化分析模型框架研究RRS，开发理论方法得到三个系统性能指标：平均救助时间，救助时间变异系数和按时救助率。应用这三个指标，进行持续改善，通过开发瓶颈指标，利用医院数据直接确定系统瓶颈。最后本课题研究系统优化问题，实现基于现有系统参数，配置RRS中所需的医疗资源，使得平均救助时间和救助时间变异系数的最小化，并研究在保障按时救助率指标的前提下，最优的系统参数。本课题从系统角度研究RRS，得到的科研成果可以帮助医院运作管理；通过研究复杂网络模型，开发一系列的数学方法。这些方法也适用于有类似网络结构的其它系统。

医疗安全的保障是医疗运营管理永恒的主题，而医疗运营效率的提升也至关重要，但临床文献和多数医院管理现状表明潜在提升空间巨大。本项目聚焦患者安全和医疗服务效率，基于建模、分析和改善优化的工业和系统工程思想，围绕和患者安全高度相关的医院紧急救助系统和重症监护系统，搭建基于随机模型与优化模型的分析框架，选取临床专家认可的关键指标求解，在对系统性质通过数学证明或数值实验进行探讨后，进行瓶颈分析和最优化决策评估。通过数值实验和基于真实医疗数据的案例分析，展示模型和方法的相对优势，并给出管理实践建议。此外，基于迭代等思想开发的研究方法，在生产系统领域做拓展，开展同时考虑生产线耗能和效率的联合建模、分析和改善研究，并探讨针对本质结构相似的问题解决方法的通用性。本项目重视结合医院实际需求，和国内三家医院开展了手术停台分析和基于随机混合整数规划模型的多因素下的最优排班方案确定。最后，项目团队围绕从医院获取的多个医疗数据库为以上研究提供案例分析的数据，也开展了基于数据挖掘的医疗资源消耗联合预测，为医疗资源的优化配置打下基础。.成果方面，已发表和录用的国际期刊论文13篇，其中项目负责人作为第一或通讯作者的论文12篇，并完成英文著作1部。发表在IISE Transactions on Healthcare Systems Engineering的工作被作为期刊总主编选中，作为国际工业和系统工程学会的亮点科研成果被报道。发表在IEEE期刊的5篇论文中包括3篇IEEE Transactions on Automation Science and Engineering的汇刊长文。项目成果有1篇发表在公共卫生领域的国际权威期刊Preventive Medicine(2017 IF:3.48)，并被国家卫健委主管的《健康报》报道。论文发表后被来自美国密西根大学、威斯康星大学、梅奥医院等境外知名学术机构的学者引用和评述。.项目负责人在国内外多次开展学术交流活动，团队共有22人次参加国际会议。2名博士生和2名硕士生毕业，5名博士生进展顺利，多名学生获得国奖、优秀论文奖和国际会议旅行奖。项目负责人为国家卫生健康委撰写咨询报告一份并得到批示，并创建网站进行成果展示。