應用邏輯斯迴歸於冠狀動脈心臟病之研究

Using Logistic Regression to establish a coronary heart disease diagnostic model

蔡孟潔(Meng-Jie Tsai蔡孟潔)

指導教授 ： 江行全

元智大學/工程學院/工業工程與管理學系/碩士(2006年)

https://doi.org/10.6838/YZU.2006.00059

摘要

心臟病長期以來一直是國人的前三大死因之一，其中冠狀動脈心臟病為最常見的猝死病因，目前臨床診斷僅有心導管手術較能正確判斷此疾病，但進行開刀手術不僅耗費龐大的醫療資源，且對病患的體力也是一種負擔。本研究最終成果是要建立心臟病之診斷預警系統，以提供專業醫師與心臟病患在病症發作前察覺其預兆。

對此，本研究以某醫院心臟科的315位冠狀動脈心臟病患者及32位正常無冠狀動脈心臟病之資料，分成兩部分進行探討，第一部分針對居家自行量測的生理指標(收縮壓、舒張壓、空腹血糖和身體質量指數)建構一個居家預警模型，作為是否需到醫院檢查的居家照護警訊；第二部分則是利用居家自行量測的生理指標和需到醫院量測變項(總膽固醇、三酸甘油脂、高密度脂蛋白和低密度脂蛋白)建構一個醫務預警模型，提供醫療人員做為是否需要做進一步檢查的參考(例如:心導管手術)。

利用邏輯斯迴歸分析得到較顯著之重要因子為身體質量指數(BMI)、空腹血糖(AC)、總膽固醇(CHO)、高密度脂蛋白(HDL)與收縮壓(STP)等5項，可提供醫療人員對病患診治參考，且其預測機率可說明發生心臟病的危險程度，有利於病患居家照護時的重要依據，未來可做為心臟病醫生診斷之參考。

關鍵字

冠狀動脈心臟病 ； 冠狀動脈心臟病危險因子 ； 邏輯斯迴歸

並列摘要

Coronary heart disease (CHD) is one of the major causes of death in Taiwan. Cardiac catheterization is the only way to identify this disease. The objective of this research is to construct an early warning system for heart disease diagnosis. The information obtained from this system could provide signs for professional doctors and heart disease patients to be aware of having relative symptoms of heart disease.

The study samples were collected from the Department of Cardiology of a local hospital. This study used several CHD risk indexes to establish a diagnostic model and divided the collected risk indexes into two parts. For the first part, the risk indexes could be measured at home, such as body mass index (BMI), ac sugar (AC), systolic pressure (STP) and diastolic pressure (DTP), to establish a home care system. For the second part, the risk indexes measured in the hospital, such as total cholesterol (CHO), low density lipoprotein cholesterol (LDL), high density lipoprotein cholesterol (HDL) and triglyceride (TG), could be integrated with home care model in the first part to construct a health assessment system.

There were several significant factors found as BMI, AC, CHO, HDL and STP by using logistic regression method. The factors provided useful information to diagnose and monitor patients' condition and the result indicate the recurrence probability. Finally, it could apply remote control to integrate and to analyze the collected data to complete the system. This integrated system can be served as a reference diagnosis consultation for heart disease diagnosis.

並列關鍵字

coronary heart disease (CHD) ； coronary heart disease risk factor ； logistic regression

參考文獻

張薰文，李源德，林瑞雄，許秀卿，蘇大成，簡國龍，陳文鍾，黃玉真，宋鴻樟著，活動量與血脂、血壓的相關探討─金山居民的研究，中華衛誌，18, 413-22，民89。

Coronary Heart Disease/Stroke Task Force Report

李博智，資料探勘在慢性病預測模式之建構，元智大學資訊管理研究所碩士論文，民91。

Abdolmaleki, P., Buadu, L. D. and Naderimansh, H. (2001). ”Feature Extraction and Classification of Breast Cancer on Dynamic Magnetic Resonance Imaging Using Artificial Neural Network,” Cancer Letters, Vol. 171, No. 2, pp.183-191.

Agrawal, R., Imielinski, T., & Swami, A. (1993). ”Mining association rules between sets of items in large database,” Proceedings of the ACM SIGMOD Conference on Management of Database, pp. 207-216.