【作者】 李远敏；

【导师】 马光志；

【作者基本信息】 华中科技大学 ， 计算机应用技术， 2004， 硕士

【摘要】 电子病历是实现以病人为中心管理模式的基本要求,是数字化医疗发展的必然趋势,它包括了患者在医院诊断、治疗过程的全部信息。由于HL7标准提出不久,且还处于进一步完善的过程中,电子病历数据抽取和模式挖掘在国内外都少有研究。非标准的电子病历必须转换为标准的电子病历,才能实现不同单位和部门之间的信息交换。标准电子病历遵守HL7V3标准,其CDA消息结构用XML表示。由于XML形式的病历文档存在对应的XML 模式(Schema),从而为标准电子病历消息的转换和数据抽取提供了可能。据此,以HL7V3参考信息模型RIM的类图为桥梁,基于对象模型驱动映射实现了二者之间的转换。在此基础上,基于.NET FrameWork中XML的读写类、文档类等,利用文档对象模型和拉模型技术,对标准化病历消息进行了消息有效性验证,并实现了电子病历消息的数据抽取。最后,通过电子病历数据抽取接口获取数据,利用Apriori算法的连接运算产生候选频繁集,并根据候选频繁集构造条件模式树,通过对FP-growth算法的改进实现了频繁模式挖掘。通过以上工作,完成了基于HL7电子病历数据抽取及模式挖掘的任务,满足了电子病历信息交换及数据分析的要求,从而为基于HL7的电子病历知识发现奠定一定的基础。更多还原

【Abstract】 The electronic medical record is the basic request for realization of the management pattern which takes the patient as the central, and it is an inevitable trend of the digitized medical service development .It includes all the information of during the patient in the hospital diagnosis and treatment. Because the HL7 standard has just proposed and is also in the process of further consummation, electronic medical record data extract and pattern mining have been studied little. In order to realize the information exchange between the different departments, Thenon- standard electronic medical record must be transformed into the standard one. The standard electronic medical record observes the HL7V3 standard, its CDA (Clinical Document Architecture) messages are expressed with XML. Because medical record documents expressed by XML exists the correspondence XML Schema, the standard electronic medical record messages transformation and data extract become possible. According to the above, mapping based on the object-oriented model actuation has realized transformation by class graph of HL7V3 reference information model. In this foundation, by NET FrameWork’s XML read-write and the documents class, the use the technology of documents object model and pull model, has carried on the message validity confirmation to the standardized medical record messages, and has realized electronic medical records data extract. Finally, system gain data through the electronic medical record data extract interface, produces the candidate frequent collection using the Apriori algorithm connection operation, and structure condition pattern tree according to the candidate frequent collection, realized the frequent pattern excavation to improve the FP-growth algorithm to realize the frequent pattern minning. Worked through above, system has completed based on HL7 electron medical record data extract and the pattern minning duty, has satisfied the electronic medical record exchange of information and the data analysis request, thus it laid the certain foundation for knowledge discovered based on the HL7 electronic medical record.更多还原