基于自然语言处理的临床医学检查危急值自动识别平台的设计与应用

Design and application of automatic identification platform for critical values in clinical medical examinations based on natural language processing

目的:面对人工式检查检验危急值管理存在的效率低、易出错等痛点,借助自然语言处理技术进一步提升危急值识别率,减少人工差错。方法:采用自然语言分词的方法,利用Jieba中文分词工具,创新利用正反双向规则库,与相关系统进行集成,构建了一个具备检查检验危急值处理等功能的危急值自动识别平台。结果:该平台使得危急值报告上报率达到了100%,显著提高了危急值的识别效率,减少了危急值上报遗漏,进一步提升了医疗安全质量。结论:自然语言处理技术可有效应用于检查检验危急值自动识别和提示,对于辅助临床危急值管理识别具有积极作用。

Objective In the face of the problems of low efficiency and error-prone in critical value management of manual inspection,natural language technology can be used to further improve the identificationrate of critical values and reduce human errors.Methods Using the natural language word segmentation method and the Jieba Chinese word segmentation tool,and with the innovativeuse of positive and negative bidirectional rule libraries,whichwereintegrated with relevant systems,an automatic critical value recognition platform with the functions of checkingand processing critical values was constructed.Results The platform has achieved a 100%reporting rate for critical value alerts,significantly improved the efficiency of critical value identification and reduced the risk of critical value reporting omissions,further enhancing medical safety.Conclusion Natural language processing technology can be effectively applied to the automatic identification and prompting of critical values in laboratoryexaminations,which has a positive impact on assisting clinical critical value management.