本文利用民國88年921震災後，政府對災民發放921震災卡之免部分負擔的優惠為一自然實驗，以災區（南投縣及台中縣）原本即享有就醫免部分負擔的榮民榮眷、並具有災民身份者為控制組，同時以非災區（彰化及雲林縣）的榮民榮眷但非災民者、以及災區領有921震災健保卡之災民為兩個實驗組。將差異中之差異分析方法應用於非線性模型（負二項分配迴歸模型及二部模型），分析65歲以上老人之健保醫療利用的價格彈性。在控制時間趨勢與族群差異後，本文研究發現在相同的震災處置效果下，921震災健保卡之價格補貼顯著提升災民的門診醫療利用3.24次（男性樣本則為3.50次），門診價格彈性為－0.05（男性樣本為－0.06），但對住院醫療利用卻無明顯地影響，政府的免部分負擔政策期滿終止後，災民的門診醫療利用不再明顯地增加。在控制價格補貼的處置效果下，震災傷害提高老人門診醫療利用4.63次（男性樣本為4.55次），災後第一年的住院機率增加10.37%（男性樣本為12.58%），災後第二年增幅降為2.66%（男性樣本為3.29%），災後第三年則無顯著的變化，顯然震災傷害對老人醫療利用的影響大於價格補貼的效果。本文對老人醫療利用的估計結果，可提供政府對天然災害醫療救助補貼及控制老人醫療支出政策參考依據。

This study aims to analyze the effect of price changes on the medical care utilization of the elderly aged 65 and above. The 921 Chi-Chi earthquake occurred on September 21, 1999, and caused severe damage to the central part of Taiwan, especially Nantou and Taichung Counties. Immediately following the earthquake, Taiwan’s government provided medical aid to the victims via waiving the NHI copayments of ambulatory care for one year and of inpatient care for two years. We use the 921 earthquake as a natural experiment to investigate how the elderly reacted to the copayment exemption. By using the difference-in-difference in nonlinear regression specification, on one hand, we compare a group of veterans, who were victims and lived in the affected areas, with nonveteran victims, to identify the effect of price subsidization. On the other hand, we compared the victim veterans living in the affected area with the non-victim veterans living in the unaffected area to identify the impact of earthquake damage on the medical care utilization. Our empirical results show that the government copayment exemption policy increased the mean number of outpatient visits per person by 3.24 for all observations and 3.5 for the male sample, and had no effect on inpatient care either in the probability of hospitalization or length of stay after hospitalization. The earthquake resulted in an increase of 4.63 in the mean number of outpatient visits per person for all observations (4.55 for the male sample), and an increase of 10.37% and 2.66% in the likelihood of hospitalization roughly (12.58% and 3.29% for the male sample), respectively, in the first year and second year following the earthquake. The price elasticity of outpatient care for the elderly was about -0.05 for all observations and -0.06 for the male sample. Our empirical results provide some policy implications regarding the control of the growth of medical costs for the elderly.