The Research On Full-Text Retrieval Of Computer-based Patient Record

Compared with the research status of developed countries, the research on CPR in China is at the beginning nowadays. The management information system of CPR is not a separate unit but a small part of HIS in mass hospitals. When we want to get a patient's information, we must provide his ID to query system and the system returns the information from structured database based on the patient's ID. We can't make the words in unstructured information (for example, the symptom of patients) as key words, so the important information can not be used by researchers. Therefore the query mode of management information system of CPR should change to satisfy the need of doctors and researchers, and lay a good foundation for the management information system of CPR becoming an independent branch of HIS.This paper designs an intelligent full-text retrieval system of CPR and realizes its prototype, and then does a function expansion of semantic retrieval based on thesaurus, which could mine the deep information of CPR warehouse to support the clinical diagnosis, researching and CPR management. First, we give a survey of the theory and technology of full-text retrieval, focusing on Apache Lucene, an important full-text retrieval tools library, then provide an exemplification --- project SOUGUO; Second, we change the CPR from original format to XML based on HL7, appling the technology of full-text retrieval into CPR retrieval, then realizes a prototype. The current version of this prototype system is working well now, which could resolve relative problems in clinical and medical researching to a certain extent.