目標：花蓮縣有三個山地鄉，分別為秀林、萬榮、卓溪，透過對山地鄉及非山地鄉民眾的死亡率趨勢探討分析，可進一步瞭解民眾之健康狀況及差距。本研究目的為呈現1986～2005年花蓮縣山地鄉及非山地鄉的死亡率及重要死因，並比較健保實施前後十年之變化，期能提出相關之政策建議。方法：研究資料來自「衛生署死因統計檔」及「台閩地區人口統計」，以花蓮縣2000年人口組成為標準人口，用直接標準法計算標準化死亡率，並比較趨勢。結果：花蓮縣山地鄉之死亡率從1986年每十萬人口1575人到2005年1509人，非山地鄉從1986年每十萬人口1054人到2005年765人，不同區域均有下降之情形；兩區域之死亡率相對比值在1986年為1.49，到2005年為1.97，其差距反而拉開；事故傷害、慢性肝炎及肝硬化及肺結核之死亡率均有下降，但山地鄉仍高於非山地鄉。結論：花蓮縣山地鄉與非山地鄉之死亡率呈現下降趨勢，但其死亡率差距卻拉開，未來應針對造成差距拉大的原因進一步釐清。

　　　　　Objectives: Hualien county includes three aboriginal areas and can be used as a good example to illustrate the health differences between aboriginal and non-aboriginal regions. The aim of this study is to demonstrate the mortality trends between these two populations in their respective regions in Hualien from 1986 to 2005, and to compare the mortality rate before and after the launching of national health insurance in 1995. Methods: Mortality data comes from the Mortality Registry conducted by the Department of Health, from 1986 to 2005. Census data are based on the Taiwan-Fukein Demographic Fact Book from the Ministry of the Interior. We used the 5-year demographic distribution of Hualien in 2000 as the standardized population to calculate the direct standardized mortality rate. Results: The mortality rate in aboriginal areas decreased from 1575 (per 100,000) in 1986 to 1509 in 2005. The same trend was also evident in the non-aboriginal area (1054 persons/per 100,000 in 1986 to 765 in 2005). But the relative ratio of the mortality rate between aboriginal and non-aboriginal regions rose from 1.49 in 1986 to 1.97 in 2005. The mortality rates due to accidents, cirrhosis of the liver, and tuberculosis declined, but aboriginal areas still had a higher rate than the non-aboriginal area. Conclusions: The overall mortality rate in Hualien decreased from 1986 to 2005, irrespective of whether a region was classified as aboriginal or non-aboriginal. But the mortality trends between the two populations in their respective regions were significantly different. An increasing health disparity was evident in this study during the study period. This indicates that additional studies to determine the risk factors that affect health in aboriginal areas are needed in the future.