為探討以戶籍地歸屬的縣市別機動車事故傷害死亡率問題，本研究以問卷回溯訪視民國83及84年花蓮縣台東縣居民死於機動車事故傷害者的家屬，詢問有關機動車事故發生地的相關訊息。本研究發現花蓮縣機動車事故傷害發生地在外縣市者之比例為17%(94/559)，台東縣為35%(113/323)。發生地在外縣市者(與發生地在縣內者比較)有較高比例是青少年與年輕人，較多使用小客車與貨車。在確知發生地與死亡地的樣本中，問卷所問出之發生地與死亡證明書所記錄之死亡地屬於相同縣市者，花蓮縣有548人(99%)；台東縣有301人(96%)。由於花東地區機動車事故傷害發生地在外縣市 的比例相當高，所以不宜以戶籍歸屬的縣市別機動車事故傷害死亡率推論該地區道路環境因素。本研究顯示，至少在花東地區，死亡地可作為發生地的替代。此外，機動車事故傷害發生地在外縣市者與發生地在縣內者特徵不同，所以防制計畫的標的團體也應不同。

　　　　　To assess the problems of regional mortality rates from motor vehicle injury (MVI) calculated according to place of residence, we retrospectively interviewed families of people who died from MVI during the years 1994 and 1995 in Hualien County and Taitung County. Seventeen percent(94/559) of MVI deaths occurred outside the resident county in Hualien County and 35% (113/323) in Taitung County. The characteristics of victims whose accidents occurred outside the county differed significantly in age distribu tion and pattern of road use. Among those in which the exact place of occurrence and place of death (recorded on death certificates) were known, in 99% (548/551) in Hualien County and 96% (301/315) in Taitung County, the place of occurrence and the place of death were in the same county. Because a high proportion of deaths occurred outside the resident county, the use of county-level regional mortality from MVIs calculated according to the place of residence should be evaluated very cautiously. At least in the Huatung area, the place of death recorded on the death certificate can be used as a proxy of the place of occurrence of the MVI. A control program targeting injuries which occur outside the resident county should be different from a program dealing with injuries which occur within the resident county.