- 學位論文
運用健保資料庫探勘疾病之併發症與治療結果-以心臟及臟疾病為例
Analysis National Health Insurance Database-Focus on Comorbidity and treatment in cardiac and renal disease
摘要
前言：目前分析健保資料庫已有許多研究利用資料探勘的技術，尋找資料中有用的特徵以及關連性。而藉由此方式來分析健保資料庫更能夠完整找出資料庫中隱含的資訊，提供臨床研究更進一步的應用及研究。材料與方法：本研究首先利用小型的臨床資料庫結合資料探勘方式分析心律變異跟末期腎臟病患間的關係，接著結合資料探勘方式應用於全民健保資料庫來找出(1)Statin 治療感染性心內膜炎；(2)末期腎臟病患之細菌性肝膿瘍研究；(3)自體顯性遺傳之多囊腎疾病在特定族群和惡性腫瘤的相關。結果：小型臨床資料中發現心律變異與SDNN, NN50、rMSSD、QTc 及QT 有統計上的相關性；全民健保資料中發現：(1)Statin 對於治療感染性心內膜炎有一定的功效；(2)細菌性肝膿瘍在末期腎臟病患的死亡率偏高，且發現糖尿病、高齡、硬化及高血壓等可能是重要的影響因素；(3)自體顯性遺傳之多囊腎疾病與腎臟及肺臟癌症有高度相關。結論：雖然資料探勘的結果並不能提供直接的因果關係，但是以資料探勘的方式找出上述的治療與預防因子對於進一步的臨床試驗有其重要的參考價值。
並列摘要
Introduction: Data mining has been extensively applied in analyzing National HealthInsurance Database. It can be used to search the characteristics and related properties of data,and provide further information for clinical study. Material and Method: First, we useclinical data to analysis heart rate variability with end-stage renal failure patients, and weanalyse National Health insurance database using data mining technique to study: (1) Statin Therapy in Infective Endocarditis; (2) Pyogenic liver abscess in end-stage renal disease patients; (3) Polycystic kidney disease and kidney malignancy. Result: In clinical data we can find heart rate variability parameters (SDNN, NN50, rMSSD, QT, QTc) statistic significantly after hemodialysis. In National Health insurance database studies summarized as following:(1) statin therapy is associated with a reduced risk of in-hospital and subsequent mortality of IE; (2) In end-stage renal failure patients have a higher risk of Pyogenic liver abscess, diabetes mellitus, old age, cirrhosis, and hypertension were predisposing factors. (3) Our investigation found that Autosomal Dominant Polycystic kidney patient are prone to developing kidney and lung malignancies. Conclusion: Although data mining can’t support direct causal relationship, but we can provide treatment and predicting factor for further clinical study.