目標：檢視台灣地區跳樓自殺的空間分布型態，並探討高樓可得性對於自殺者跳樓傾向的影響。方法：自殺死亡資料取自衛生署1994-2007年的死因統計檔(n=50,705)，鄉鎮市區建物資料取自2000年戶口及住宅普查原始檔(n=358)。先使用探索式空間資料分析來檢視跳樓自殺比例與高樓比例的空間型態，續以多層次迴歸模型檢驗鄉鎮市區之高樓可得性對自殺者選擇跳樓傾向之影響。結果：跳樓自殺比例呈現出有系統性的地理變異與空間聚集趨勢(Moran’s I=0.552,
p<0.001)，其熱區主要分布在基隆-台北-桃園、台中、高雄等都會區。高樓比例也同樣具有顯
著的地理變異型態(Moran’s I=0.747, p<0.001)，其熱區的空間分布型態與跳樓自殺比例十分吻合。多層次分析結果顯示自殺者的方法選擇(跳樓vs.其他)有10%的變異數來自總體層次。在控制了個人層次之性別、年齡、婚姻、就業、季節、以及時期、與總體層次之社會碎化、經濟剝
奪、以及都市化程度等因素之後，自殺者所居住的鄉鎮市區之高樓可得性愈高，則愈傾向於使用跳樓方式來自殺。結論：跳樓自殺比例在鄉鎮市區之間有明顯的差異，且自殺者選擇跳樓的傾向與高樓可得性有顯著的正向關聯，因此社區自殺介入策略應該考慮此區域差異現象，在高樓比例較高的區域，更應該著重跳樓自殺的防治。

Objectives: The aim of this study was to explore the spatial patterns of suicide by jumping in Taiwan and to examine the effects of the availability of high-rise buildings on suicide completers’ propensity to jump from heights. Methods: Data on all completed suicides during the period 1994-2007 (n=50,705) were derived from the cause of death database provided by the Department of Health, Executive Yuan, Taiwan. The proportion of high-rise building at the township-level (n=358) came from the 2000 census of population and housing. Exploratory spatial data analysis methods were used to investigate the spatial patterns of proportions of suicide by jumping and high-rise buildings. Multilevel logistic regressions were employed to assess the association between the township-level availability of high-rise buildings and the individual’s choice of suicide methods. Results: The proportion of suicide by jumping showed a significant spatial dependence (Moran’s I=0.551, p<0.001). Hot spots were detected in Keelung-Taipei-Taoyuan, Taichung, and Kaohsiung metropolitan areas. Such a spatial pattern coincided with the proportion
of high-rise buildings. Results from multilevel models indicated that 10% of the variance in the choice of suicide methods was between townships. After controlling for gender, age, marital status, occupational status, season, period, social fragmentation, economic deprivation, and level
of urbanization, we found that the deceased who lived in a township with a greater availability of tall buildings had a significantly higher likelihood of committing suicide by jumping. Conclusions: Since there are significant township-level differences in the proportions of suicide
by jumping, community suicide intervention strategies should take these spatial variations into account. Programs to prevent suicide by jumping should be targeted to areas with more high-rise buildings.