應用大規模臨床資料庫探勘骨關節炎之早期風險模式

Mining Early Disease Risk Patterns for Osteoarthritis from Large-Scale Clinical Databases

王懿淳(Yi-Chun Wang)

指導教授 ： 曾新穆

國立交通大學/資訊學院/資訊學院資訊學程/碩士(2018年)

摘要

骨關節炎(Osteoarthritis, OA)又稱退化性關節炎，因隨著時間，關節軟骨磨損、退化，導致骨頭之間互相摩擦，初期為疼痛，接著會關節僵硬，後期會關節變形、肢體無力。骨關節炎易發生於膝及髖關節，末期的退化性膝關節炎，關節損害程度和疼痛已嚴重影響生活，甚至無法行動，則需進行全膝人工關節置換術，且骨關節炎已不完全是老人的專利，有呈現年輕化的趨勢。然而，若骨關節炎可以在早期階段被診斷，則可透過療程來減緩退化的速度。本研究是第一個結合關聯規則分類技術和大型臨床資料庫應用在退化性關節炎上，透過資料預處理、危險因子探勘、模型驗證和統計分析，共四個階段。利用關聯規則分類探勘出骨關節炎發病前的風險模式和危險因子。接著，將危險因子分成三個等級，層級1的疾病，在過去醫師看診時，確實會特別注意這類病人在骨關節炎病情上的變化，而層級2和3則是過去醫師或病人較忽略的部份。最後，將探勘出的風險模式和危險因子進行病例對照研究，計算勝算比來驗證罹患疾病的危險性。由層級2和3的疾病所組合之風險模式，為本研究最重要的貢獻，將可提供醫師進行更深入的臨床研究，並可進一步提醒病患盡早進行檢查診斷及治療，減緩退化的速度。

關鍵字

臨床資料庫 ； 資料探勘 ； 病例對照研究 ； 骨關節炎

並列摘要

Osteoarthritis (OA) is also known as Degenerative Joint Disease. According to articular cartilage wears and degenerates over time, OA causes bones within cartilage to rub together. In the initial stage, patients start to feel pain and then stiff in the joints. Finally, it will become deformed. OA is more common in knee and hip joints. At the end stage of Knee OA, the joint will seriously affect patients’ life with damage and pain. Therefore, Total Knee Arthroplasty is required for them. OA is not completely patented by the elderly, and it has a tendency to the younger. However, if OA can be diagnosed in the early stages, the rate of degeneration can be slowed down through the course of treatment. Our study is the first to utilize Association-based Classification techniques and large-scale clinical databases to detect early risk factors for Osteoarthritis through four stages, namely data preprocessing, risk pattern mining, model validation and statistical analysis. The discovered risk factors are divided further into three levels. In the past, doctors paied special attentions to patients who have level 1 diseases. However, compared with level 1, physicians or patients were inclined to ignore level 2 and 3. Finally, the case-control study is conducted on the risk patterns and risk factors that are explored, and the odds ratio is calculated to verify the risk of the disease. The major contribution of this research lies in discovery of risk pattern combined with diseases of level 2 and 3. The result can provide physicians for more in-depth clinical researches. Moreover, they can be used to remind patients for relevant risk factors and to conduct suitable treatments as early as possible for slowing down the rate of degeneration.

並列關鍵字

Clinical Databases ； Data Mining ； Case-Control Study ； Osteoarthritis

參考文獻

1. 謝邦昌，鄭宇庭，資料採礦之技術及應用（Excel實例演練），初版，新陸書局，台北市，民國一零四年

2. Jiawei Han, et al.著，資料探勘，郝沛毅，李御璽，黃嘉彥譯，初版，高立圖書，新北市，民國一零三年

3. 王靜枝、楊月穎、劉美吟，「高齡長者之全人健康照護概念與介入措施」，護理雜誌，第65卷第2期，5-12頁，2018.04。

4. 陳崇桓，「退化性膝關節炎的治療」，高醫醫訊，第36卷9期，2017.02。

5. 葉文凌、吳濬哲、陳健煜，「膝蓋退化性關節炎年輕化」，常春月刊，第392期，22頁，2015.11。