The Research Of Electronic Medical Record Retrieval In The Context Of Hadoop

Electronic medical record as an important part of hospital information system, is the complete and detailed clinical information resources which record the process of patients’ all medical institutions. Electronic medical record is of great significance to the sharing of information between doctors,the comparison and analysis of medical case, academic research in the medical field. However, the development of electronic medical record appears some problems in some big hospitals in china. For example, the system of electronic medical record is developed in accordance with difference standards. The medical record information is inquired only through some keywords such as, the patient admission number or the patient name. The increase of the electronic medical record data leads to the slow retrieval. So how to store the electronic medical record data, how to improve the retrieval efficiency of electronic medical record and how to analyze the massive electronic record data have became one of the research hospital in medical field.For this kind of situation, after a in-depth study on the international standards of electronic medical record, Hadoop platform, Lucene search framework, this paper proposes a solution of electronic medical record storage, retrieval and analysis in Hadoop environment. The corresponding prototype system is designed and implemented.First of all, the article combines the centralized storage of electronic medical record with the distributed storage in Hadoop HDFS file system. The hybrid storage is used to solve the problem of electronic medical record data explosion. Second, the electronic medical record is preprocessed, the analyzer of electronic medical record is implemented and the Lucene index is established through the distributed, high performance Map Redcue programming framework. The method can realize the full-text index of electronic medical records and solve the problems of the single retrieval way and slow search speed. Then, on the Hadoop electronic medical records retrieval platform the retrieved medical record data is analyzed through the optimized Apriori algorithm in data mining and Map Reduce framework. This method can access to the correlation of property in electronic medical records. At last, the author completes the system integration and test on the Hadoop cluster environment in the laboratory. The results show that, this method not only realizes the storage, retrieval and analysis of electronic medical records, but also realizes a certain degree of improvement in retrieval function, efficiency. This method has a certain application value.