A Personalized Risk Tracking Website for Primary Prevention of Coronary Artery Disease

心臟病是長期潛在的慢性疾病，近年來連續五年在國人十大主要死因中高居第二，若能將冠心症相關的各健康檢查項目做長期追蹤，並使用這些檢查項目的結果做綜合的風險評估，可警示與篩檢出潛在的冠心症患者，及早進行自主健康管理，避免病情持續惡化，因而達到健康老化之目的。我們透過蒐集及分析臺灣各大醫院冠心症相關的健康檢查項目，設計兩種評估冠心症風險的方案，兩種方案使用的健檢項目有部分不同，我們並以臨床案例庫資料計算比較兩種方案評估冠心症風險的可信程度。本網站系統根據個人冠心症相關的各健康檢查項目的結果，以實例式學習法找出案例庫中健康檢查結果類似的案例，以統計這些類似案例中各健康檢查項目的結果，供使用者比較個人與類似案例中冠心症患者／非冠心症患者之間的健康差異，本系統同時提供功能以追蹤各冠心症相關健康檢查項目的趨勢，協助警示與長期追蹤個人冠心症的相關風險。
Heart disease has been the second leading cause of death over the last five years in Taiwan. Many sudden cardiac deaths were attributed to coronary artery disease (CAD). CAD is a chronic process with a long asymptomatic latent period, which provides a chance for early preventive interventions. Since the incidence of CAD is largely explained by modifiable lifestyle-related risk factors, a logical way for preventing CAD is to increase awareness and encourage people developing CAD to reduce risks through health-promoting diet and lifestyle. In this study, a website was developed to provide a series of functions for the users to realize their CAD risks. The functions include risk computation, risk tracking and risk alerting. The risk computation is based on an instance-based learning algorithm. The set of training instances contains information about demographics and multiple plasma biomarkers. The biomarkers include traditional and nontraditional risk factors of CAD. The tracking function of the website helps the user be aware of his/her CAD risk trend. The risk alerting function help the user notice how his/her health is different from those of the CAD patients/non-CAD individuals in the training instances in terms of CAD risk factors.