临床路径变异检测与分析方法研究

Clinical pathway (CP) is a beneficial and effective management tool that aims at improving healthcare service delivery and controlling medical cost. The implementation and incremental refinement of current CPs rely heavily on knowledge and experience of clinical experts. As a result, there generally exist various deviations of practical medical processes from CPs. Variations, as exceptional or inappropriate medical behaiovrs in CPs, may be related to caring performance or even the patient survival. It is imperative that the variations be timely detected and analyzed so that useful and actionable knowledge of interest could be extracted to clinical experts, and CPs could be more efficient. .Existing techniques for CP varation detection and analysed are based on the experiences and knowledge of clinical experts, or are oriented to clinical data statisitcal analysis such as the statistics of pathway coincidence rate and abort rate, etc. In such techniques, the analysts interpret large amounts of collected medical behaviors, and elaborate CP variations, piece after piece, which can be a very tedious process..The target of this research is to propose a data mining approach that can be applied to clinical data to detect and analyze varaitions in CPs. Clinical Pathway Ontology (CPO) is to be constructed and a representative language will be designed to describe practical clinical process. Based on CPO and the representative language, novel data mining algorithms are proposed to detect critial variations of medical behaviors in CPs, to identify associations between the detected variations, and to analyze the correlations between variations and patient's symptoms so as to provide diagnostic information on the detected varitions in an informative manner, and assist the designing and improvement of CPs. Process mining technologies are the frontier research areas of process intelligent computing. Our research proposes novel approaches to design and optimize CPs and contributes to the extensions of theory of process intelligence to complex medical settings.

临床路径是提高医疗质量，降低医疗费用的重要手段，但在其实施过程中，会不可避免地出现偏离路径的情况，即发生变异。变异检测与分析是实施、评估和修正临床路径的关键，可以为临床路径提供持续的、及时的正反馈，有利于整体医疗流程的优化。现有变异检测和分析大多依赖人工分析，费时费力、繁琐易错。为此，本项目目标是研究出临床路径变异检测与分析方法。首先通过建立临床路径知识本体，研制出一种能有效表达临床路径的建模语言。在此基础上，检测临床路径中的关键变异，分析变异之间的依赖关系，以及变异与患者疾病和症状的关联信息，发现变异发生的根因，为临床路径的设计和完善提供优化建议。项目研究为临床路径设计与优化提供了新的方法与手段，而且对于发展复杂多变诊疗环境下的过程智能理论具有重要科学意义。

临床路径变异检测与分析是实施、评估和修正临床路径的关键，可以为临床路径提供持续的、及时的正反馈，有利于整体医疗流程的优化。本项目系统研究了基于电子病历数据挖掘的临床路径变异检测与分析方法。通过建立临床路径知识本体，对电子病历数据进行结构化提取和语义预处理。在此基础上，设计和实现了若干基于电子病历数据挖掘的临床路径变异检测与分析方法，检测临床路径中的关键变异，分析变异发生的根因，支持从全局视角描述临床路径的整体结构和特征，量化分析医疗行为内容差异和时间差异，准确的识别路径变异行为等。前期的研究成果已经在解放军总医院开展临床转化应用。项目研究不但为临床路径设计与优化提供了新的方法与手段，而且对于电子病历的研究与应用具有重要科学意义。