全民健保實施至今已二十餘年,如何抑制醫療成本持續膨脹,減少醫療資源浪費是政府一直所重視的,因此部分負擔是全民健保制度設計之重要一環。本研究旨在探討2005年健保門診部分負擔大幅度調漲後,不同疾病(上呼吸道感染、糖尿病)患者中不同族群(所得、年齡)的醫療利用變化。本研究利用健保門診部分負擔調漲政策實施的前後一年(2004/07/15-2006/07/14)的看診資料來觀察上呼吸道感染與糖尿病病患中不同族群的醫療利用情形,並依照不同所得與不同年齡層作為分組依據來進行分析。我們使用序列logit迴歸模型(ordered logit regression model)來當作實證研究模型,並利用政策邊際效果與所得邊際效果來衡量此次政策的成果。 本研究的實證結果顯示,政策效果方面,上呼吸道感染患者的就醫層級選擇對部分負擔調漲政策有顯著影響;糖尿病患者中僅有低所得組、地區人口與農民受政策影響。所得效果方面,高所得族群較不受政策影響,對他們而言只要所得提高,就越有可能往高層級醫院就醫的誘因存在,但效果很小。 ;Since the implementation of National Health Insurance (NHI) on Taiwan for more than 20 years, it is important for the government to control the expansion of medical costs and reduce waste of medical resource. Therefore, copayment is a significant part of the design of the NHI system. The purpose of this study is to exam the changes in medical use of different patient groups (income, age) with different diseases (upper respiratory tract infection, diabetes) after a substantial increase in the outpatient copayment in 2005. This study used the observation data of the one-year (2004/07/15-2006/07/14) implementation of the NHI outpatient copayment increase policy to observe the medical utilization of different groups in upper respiratory tract infection and diabetes patients, and the analysis is based on different incomes and different age groups. We use the ordered logit regression model as an empirical research model, computing the marginal effects of the policy and the marginal effects of income to measure this copayment policy. The results of this study show that in terms of policy effects, the choice of hospital level for patients with upper respiratory tract infection has a significant impact on the copayment increase policy; while the diabetes patients, only the low-income group, the regional population and farmers are affected by the policy. On the other hand, in the terms of income effects, the high-income groups are less affected by the policy. For those people, as long as the income increases, the more likely they are to go to a high-level hospital for medical treatment, but this effect is small.