心臟病電腦診斷與決策輔助系統建立之研究－－冠狀動脈心臟病為例

A Study of Building Heart Disease Computer Diagnosis and Decision Support System-- A Case Study of Coronary Artery Heart Disease

本文提出應用人工智慧中專家系統的技術為基礎建立一套『冠狀動脈心臟病之診斷決策輔助專家系統』之研究。研究的目的在於建立一套決策輔助系統來協助心臟病專科醫師得以早期評估病人心臟病致病因子及嚴重性，以提供迅速並有效的治療方式，如此不僅可以避免病人致病所帶來的痛苦，且可以降低醫療支出的浪費。本研究的方式為藉由文獻探討及匯集醫師專家之專業知識並彙整來提供做為建立專家系統(rule-based expert system)推論引擎(reasoning engine)之基礎，經專家系統的推理，達到互動式(interactive)查詢的功能，選定由Java語言構成且執行效率良好的JESS (Java Expert Systems Shell)專家系統開發工具並做為系統推論的引擎，並運用以冠狀動脈心臟病醫療知識為基礎來建立此一以冠狀動脈心臟病為例之心臟病電腦診斷 與決策輔助專家系統。本研究的實驗結果顯示電腦心臟病輔助診斷及專家系統確實可以達到與專科醫師一樣或更佳的診斷水準，並且藉由本系統的協助，專科醫師可更快速且廣泛從系統及系統所匯集的知識庫推論出相關輔助診斷及診療指引，不僅可突破醫師本身的專業知識限制並減少記憶繁瑣之負擔，早期認知冠狀動脈心臟病之危險因子，掌握初期癥狀，提供早期治療的契機，更可使患者得到適當的醫療保健服務，來減少冠狀動脈心臟疾病的發生所帶來的痛苦。

This paper presents a research of building a computer diagnosis and decision support system for coronary heart diseases by applying the expert system technology in artificial intelligence. The objectives of this research aim to build a decision support system for coronary heart diseases to aid the heart-disease physicians able to evaluate the disease entity at the earlier stage and decide on the severity of the disease so that a prompt and better treatment can be provided, thus can reduce the unnecessary medical expenditure whilst avoid the pain of get ting illness of patients. In this research, we firstly review medical journals and collective survey data of the physician expertise knowledge of coronary artery heart disease to facilitate the foundation of building the rule-based expert system and reasoning engine. Through the logical reasoning of the expert system, enriching the inquiry function to an interactive level, we can institute an interactive program for the diagnosis and decision-making expert system for coronary heart diseases. The Java language and the JESS (Java Expert System Shell) are selected as the system development tools and system reasoning engine by applying the diagnostic knowledge and treatment guidelines of coronary artery heart disease to build the system. The experimental results show this computer diagnosis and decision support system truly can provide same as and even better diagnostic level than a single physician of coronary heart disease. Through the aid of this system, the physicians can break through their restrictions of narrow professional knowledge, reduce their tedious memory\ies, and gain more broad information to enhance their quality of cares. This system also may enable early identification of risks, provide more aggressive treatment and reduce the appearance of symptoms like angina, heart failure etc. which may result in the expensive treatment modalities, as well as to lessen the incidence of coronary heart disease.