Objectives : The study aims to evaluate the effectiveness of a Pay-for-Performance (P4P) Programme introduced for type Ⅱ diabetes patients in Taiwan since 2001, by observing their risk of diabetes-associated complications as compared to their not enrolled counterparts. Methods : A retrospective study was conducted from 2001-2008, using National Health Insurance Research Database. Each sampled patient’s propensity score was computed based on their gender, diagnosed time, age at diagnosis, Charlson Comorbidity index and accreditation of healthcare providers. P4P and non-P4P patients were matched and selected by 5 → 1 digit Greedy match with a ration of 1:2 for analysis. Kaplan-Meier method and Cox Proportional Hazard Model were performed to measure cumulative survival rate and hazard ratio respectively of selected diabetes-associated complications facing P4P groups and non-P4P groups. Results : During study period, 41,988 patients were included in the study, of whom 26.41% being enrolled in the P4P Programme. The final sample included 10,730 P4P patients and 21,460 non-P4P patients. Analytic results showed significant increases on survival rates as time grows after the P4P program executed on complications of ischemic heart disease (IHD, P

目的： 2001 年 11 月，台灣中央健保局推行“全民健康保險糖尿病醫療給付改善方 案試辦計劃（Pay-for-Performance Programme）”嘗試從醫療過程入手，增加給付 方式，鼓勵醫療提供者給予更好的照護品質。本研究的主要目的是研究加入該試 辦計劃對糖尿病患者罹患併發症的長期影響，為該試辦計劃的政策改善提供參考。 方法： 本研究基於台灣國家衛生研究院 “全民健康保險研究資料庫 2005 年的百萬 抽樣歸人檔”，2001 年至 2008 年的回顧性研究。共有 41,988 位二型糖尿病患者 被納入研究，其中 26.41% 的患者參加了論質計酬（Pay-for-Performance）的給 付方式。研究採用傾向性分數（Propensity Score）1:2 配對的方法選出實驗組與 對照組，然後運用 Kaplan-Miere 分析和 Cox 比例危險模型分別比較兩組患者的 生存率的變化與發病風險比。 結果： 經傾向性分數配對后，選出 10,730 位曾加入論質計酬給付方式的患者（實 驗組），與對應的 21,460 位未加入該給付方式的患者（對照組）。Kaplan-Miere 分析結果顯示，在糖尿病確診之後，隨著時間的增長，實驗組的缺血性心臟病（P