目標：本文探討2005年健保局調高部分負擔對民眾醫療利用和就醫選擇影響。方法：本文使用2005年健保百萬歸人檔中的西醫門診資料來分析政策效果。由於2005年部分負擔調高僅侷限於醫院，政策效果理應依民眾所屬鄉鎮使用診所高低而有差別。本文因而將病患經常就醫鄉鎮按平均診所就醫比例區分三組：低於65%（分組Ⅰ），65%-75%（分組Ⅱ），高於75%（分組Ⅲ），採障礙模型(Hurdle model)估計部分負擔調高後對各組就醫機率與次數影響。結果：部分負擔調整顯著減少就醫機率及門診次數。整體而言，調漲後半年門診次數減少0.34次（或6.1%），其中醫院次數為0.07次，高層級醫院為0.01次。更重要的是，各分組間政策效果存在相當差異。處於使用醫院最高比例鄉鎮（分組Ⅰ）民眾，降低醫院（和高層級醫院）利用；相反的，處於使用醫院比例最低鄉鎮（分組Ⅲ）民眾，增加醫院（和高層級醫院）利用。這估計結果即使將樣本時間延長至政策前後一年仍相同。結論：部分負擔調高的確減少就診次數，降低民眾醫療利用。另外，政策效果會依民眾經常就醫地點而有顯著差異。一個可能解釋為醫院服務的容量限制(capacity constraint)。因醫院存在容量限制，使得鄰近醫院的分組Ⅰ民眾較易接受醫院服務，但是當部分負擔調漲使得分組Ⅰ民眾降低醫院就醫，此時就醫距離較遠鄉鎮民眾（分組Ⅲ）填補原先醫院門診空缺，使得分組Ⅲ民眾醫院利用反而在部分負擔調高後有所增加。

Objectives: The aim of this study was to evaluate the impact of the increase in NHI copayments in 2005 on the choice and use of health care. Methods: We analyzed the outpatient utilization of one million NHI enrollees between 2004 and 2006. Because the policy increased only the copayment for hospital visits, it was plausible to expect that the effect would be smaller for individuals residing or working in towns with a lower propensity for visiting hospitals. Therefore, based on the average percentage of clinical care in the town of residence, the sample was separated into three groups: less than 65% (group Ⅰ), between 65 and 75% (groupⅡ), and more than 75% (group Ⅲ). We then used the Hurdle regression model to examine the effect of the new policy on the probability of health care utilization, and the number of outpatient visits for each group. Results: Our results showed that the increase in copayment significantly reduced the probability and the number of outpatient visits. We estimated that outpatient use six months after the policy change decreased by 0.34 visits (or 6.1%), of which 0.07 were visits to hospitals, and 0.01 were visits to regional hospitals or above. More importantly, the effect differed substantially across groups. Individuals residing or working in towns with a higher propensity to visit hospitals (group Ⅰ) reduced their visits to hospitals after the policy change. On the other hand, individuals residing or working in towns with a lower chance of visiting hospitals (group 3) increased their visits. These findings were robust even after extending the sample period from six to 12 months. Conclusions: Our results confirmed that a copayment increase in 2005 reduced the use of outpatient services. Nonetheless, the effect differed substantially across groups with various propensities to make hospital visits. One explanation might be that there is a capacity constraint that limits outpatient services offered by hospitals. While individuals residing or working in towns near hospitals reduced their hospital visits due to the price hike, those residing or working in more distant towns increased their hospital visits.