戴秋玉1,张伟1,黄钊1,陈杰2,陶震寰1
目的 针对区域胸痛中心建设中存在的问题,提出一种利用信息手段优化区域胸痛综合管理平台的设计方案及具体实
现路径。方法 在现有胸痛管理系统的基础上,加入人工智能心电图(Artificial Intelligence Electrocardiograph,AI-ECG)系
统、整合多个医疗救治系统、打造移动应用服务,建立“患者-基层-急救中心-胸痛中心”四方联动的区域协同救治体系。
采集区域心电中心海量标准化数据,训练AI-ECG模型,并将AI算法应用于疾病筛查、快速甄别和术后监测等方面。结果 通过
AI-ECG及胸痛管理平台的优化,提高了基层医院的诊断准确性和效率,心电图十八分类诊断模型诊断特异性、敏感度、准
确率均值分别是95.79%、87.88%和96.74%;AI辅助诊断平均每份心电图的分析时间约为0.12 s,心内科临床医生诊断时间
约为1 min;发病到首次医疗接触和首次医疗接触到球囊再灌注的平均时间均明显缩短。结论 平台的建设及应用,实现了
胸痛患者院前、院中、院后的全流程闭环管理,增强了胸痛患者自我健康管理意识,对智慧胸痛中心的建设具有一定的参
考价值。
Methods On the basis of the existing chest pain management system, artificial intelligence electrocardiograph (AI-ECG) system
was added, multiple medical treatment systems were integrated, and mobile application services were created, and then a four-way
regional cooperative treatment system of “patient-basic-first-aid center-chest pain center” was established. By collecting massive
standardized data from regional ECG centers, AI-ECG auxiliary diagnosis model was trained, and AI algorithm was applied to
various aspects of disease screening, rapid screening, and postoperative monitoring. Results Through the optimization of AI-ECG
and chest pain management platform, the diagnostic accuracy and efficiency of primary hospitals were improved. The diagnostic
specificity, sensitivity and accuracy of ECG 18 classification diagnosis model were 95.79%, 87.88% and 96.74%, respectively. The
average AI-assisted diagnosis time of each ECG was about 0.12 s, and the diagnosis time of cardiology clinicians was about 1 min.
The average time of symptom onset to first medical contact time and first medical contact to balloon was significantly shortened.
Conclusion The construction and application of the platform can realize the closed-loop management of the whole process of chest
pain patients before, during and after the hospital, enhance the self-health management awareness of chest pain patients, and have
certain reference significance for the construction of the smart chest pain center.