Study Of Structuralization For Electrocardiography Diagnostic Report

Electrocardiographic (ECG) examination, one of the general examinations in a hospital as well as a frequently-undertaken multidisciplinary examination with a lot of derived subgroups, serves as a major diagnostic evidence of cardiovascular disease. Among one of the most significant components of clinical information system, electrocardiographic information system is an indispensable part of hospitals in China and thus is in urgent demand. However, existing electrocardiographic information systems can only support for data collection, diagnosis and browsing, etc. The lack of structuralization imposes restriction on the full utilization of information. Therefore, the structuralization of data is a key prerequisite of developing next-generation electrocardiographic information system. Structured ECG report can provide data of a universal standardized and structuralized format, which is suitable to be automatically processed and analyzed by computer and can support experts to make comprehensive application on it.The research aims at proposing an approach to structuralize ECG diagnostic report. The thesis analyzed ECG diagnostic terms published by the American Heart Association (AHA), the American College of Cardiology (ACC), and the Heart Rhythm Society (HRS) in 2007. Based on the terms, the structuralization and standardization of ECG diagnostic report is implemented by the structuralized entry of information. Follow-up applications of the structuralized information were primarily discussed. However, approach that completely relies on structuralized entry is unable to fully satisfy requirements of physicians to express their meaning; narrative text needs to be added to allow for more expressiveness. Therefore, this thesis applied sub-language to structuralize free narrative text. Small-scale dictionary was constructed based on vocabularies and semantics of ECG diagnostic report. A large number of ECG diagnostic reports were analyzed and annotated while concurrent modals and constraints are discovered through statistics. Rule base was constructed via semantic-oriented Link Grammar. With the assistance of dictionary, rule base and parsing method, narrative text in ECG report is structuralized. The thesis analyzed both methods of structuralization and proposed an approach based on structuralized entry of ECG report and supplemented by Natural Language Processing method to handle free narrative text. The approach has been preliminarily applied to the electrocardiographic information system and ECG report is structuralized.The thesis proposes an approach to structuralize ECG report and applies it to electrocardiographic information system. It is the key of the next-generation electrocardiographic information system, and provides references to the research of structuralization of information in medical settings. The method to parse free narrative text based on sub-language contributes to the research of Natural Language Processing of Chinese.