多维尺度综合信息监测模式下的医院感染早期预警系统设计与应用  

Design and application of early-warning system of nosocomial infection based on the comprehensive information surveillance with multi-dimensional and multi-scale mode

作　　者：范允舟 曹雄晶 熊煌果 张玉鹏 朱璇 黎明 熊莉娟[1] Fan Yunzhou;Cao Xiongjing;Xiong Huangguo;Zhang Yupeng;Zhu Xuan;Li Ming;Xiong Lijuan(Department of Nosocomial Infection Management,Union Hospital,Tongji Medical College,Huazhong University of Science and Technology,Wuhan 430022,China;Wuhan Run He De Kang Medical Data Co.,Ltd.,Wuhan 430070,China)

机构地区：[1]华中科技大学同济医学院附属协和医院医院感染管理科,武汉430022 [2]武汉润和德康医疗数据有限公司,武汉430070

出　　处：《中华医院管理杂志》2024年第5期348-355,共8页Chinese Journal of Hospital Administration

基　　金：国家自然科学基金(72004068)。

摘　　要：医院感染可危害患者安全并增加疾病负担,医院感染暴发是医院感染危害性的集中体现,具有社会敏感性。医院感染监测预警是医院感染管理的核心内容之一,准确的医院感染早期预警技术能够体现感染防控的管理水平。本研究基于多维尺度综合信息监测模式,针对医院感染暴发设计开发了早期预警系统,并于2023年3月开始上线运行。该系统可从医院的各类业务系统中抓取医院感染相关的过程类指标数据,并存储至医院感染数据中心库;在多维尺度模式下,监测对象的各类过程指标将会被按照预设的指标维度和统计尺度进行分层统计,从而生成若干亚组的时间序列数据;系统后台每周按照预警规则对所有维度和尺度的亚组监测数据集进行分析,并向管理者推送预警信号。管理者可通过数据核查、病例核查和流行病学核查方式核实预警信号,一旦确认感染暴发可尽快实施干预措施。在实证应用中,该系统在2023年8月7日对12个科室发出医院感染聚集性事件预警,而医院感染病例监测系统中最终报告有54例医院感染病例,分布于13个科室,其中6个科室存在医院感染聚集情况(病例数均超过3例),并且这6个科室均接收到了多维尺度综合信息监测系统的预警。因此,该系统预警的阳性预测值为50.0%,预警灵敏度为100.0%,且预警时间比病例监测预警平均提早5.3 d。多维尺度综合信息监测模式下的早期预警可改变医院感染暴发预警管理的被动状态,且在隐性暴发的早期预警方面具有优势,可为提升医院感染管理水平提供支持。Nosocomial infection poses a significant threat to patient safety and increase their disease burden.Outbreaks of nosocomial infection are the main harmfulness associated with nosocomial infection,which making them socially sensitive issues.Nosocomial infection surveillance and warning are core contents of nosocomial infection management.Accurate early warning technology for nosocomial infection outbreaks can reflect the management capability of infection prevention and control.This study designed an early warning system based on a multi-dimensional and multi-scale comprehensive information surveillance mode for nosocomial infection outbreaks which was launched in March 2023.This system extracted the process-related indicators of nosocomial infection from various hospital information systems into the nosocomial infection database center.Under the multi-dimensional and multi-scale surveillance mode,the process-related data were stratified according to the predefined dimensions and scales,then generating time-series datasets of numerous subgroups.The system conducted weekly for all time-series datasets of subgroups based on warning rules,and subsequently sent out warning signals to managers.These warning signals could be verified by managers through data check,case verification and epidemiological investigation.Once a nosocomial infection outbreak was confirmed,intervention measures could be implemented promptly.In practical application,the system generated warning signals for nosocomial infection clusters in 12 departments on August 7th,2023.The traditional nosocomial infection case report system ultimately reported 54 nosocomial infection cases,which distributed across 13 departments,with clusters(more than 3 cases)observed in 6 departments.All these 6 departments received warning signals generated from our system.It has been demonstrated that our system could predicted the nosocomial infection clusters 5.3 days earlier than the traditional nosocomial infection case report system on average,with warning sensitivit

关 键 词：信息系统 医学信息学应用 医院感染暴发 早期预警 过程类指标 

分 类 号：TP277[自动化与计算机技术—检测技术与自动化装置] R197.323[自动化与计算机技术—控制科学与工程]