The Trend of Chronic Kidney Disease Turn into End Stage Renal Dialysis：Growth Model Application

目的：瞭解台灣地區慢性腎臟病隨著時間因素、個人因素以及跨層次交互作用影響轉為末期洗腎。 方法：採縱貫式研究設計，次級資料分析。時間層次變項隨著時間改變的變項分別為時間變項、時間平方變項以及合併症變項，個人層次變項為個人特性，與依變項洗腎與否，對象為慢性腎臟病患者。個人特性包含性別、職業、所屬健保分局、投保薪資等級、居住地區醫院洗腎病床數加權洗腎人口以及居住地區洗腎診所數加權洗腎人口。資料來源為全民健保資料庫，其診斷碼符合ICD-9-CM為慢性腎臟病之診斷碼者。統計分析採用多層次方法中的階層線性模式之成長模式分析，包含隨機係數模型、截距預測模型。 結果：刪除遺漏值後，納入本研究之慢性腎臟病患者共有5,391人，直至2008年底共有132人轉為末期洗腎。初步研究顯示，慢性腎臟病患者轉為末期洗腎，有67.7% 的差異來自於個人層次(ICC)。經由階層線性模式分析，發現時間變項如時間、時間平方與合併症和末期洗腎有相關。而在控制時間層次後，個人層次的部份，於截距預測模式中有達統計上顯著水準的個人變項與末期洗腎呈現負向的相關，顯示居住於高屏地區、職業為低收入戶與非固定工作者以及投保薪資等級較低者，皆代表慢性腎臟病患者之經濟地位較低者易轉為末期洗腎；但於時間、時間平方以及合併症與個人變項之交互作用時卻是相反，即經濟地位較低者較不易轉為末期洗腎。 結論：不同的時間層次變項以及個人層次的變項對於慢性腎臟病患者轉為末期洗腎有著直接的影響。Objective: To explose the effect of time variables, individual features and economis status, and cross level in paients who have Chronic Kidyne Disease (CKD) to become End Stage Renal Dialysis (ESRD). Method: The study design adpated retrospective study analysis. All variables were abtained from National Health Insurance Database.Time variables meant the vaiables'' status were changed with timet hat were including time, time squared, and CCI. Individual variables were including gender, occupation, living area, salary, and the accessing of health care resource. The CKD samples were patients who the disease was confirm with ICD-9-CM code. Growth model of Hierarchical Linear analysis was used in the study, including random coefficient model, and intercept as outcome model. Result: Excluding the missing data, the CKD samples were 5931. Until 2008, 132 of them were become ESRD. In hierarchical linear analysis, discovering time variables were related to ESRD. When controlling time variables, individual variables were related to intercept as outcome model. It showed that living in south area, proverty and unstable work, and low income were related to ESRD. Conculsion: Time variables and individual variables would cause the patient who has CKD to become ESRD