2000-2007年台灣本島十歲以下兒童非蓄意性傷害死亡之時空變異與影響因子分析

Abstract

目的： 首先藉由空間統計分析瞭解臺灣地區十歲以下兒童非蓄意性傷害死亡之地區變異，包含偵測非蓄意性傷害死亡之群聚區域，以及比較不同年份之間群聚之時空變遷。其次探討鄉鎮層級之社會經濟與家庭因子，對於非蓄意性傷害死亡的影響。 方法：本研究資料資料以2000~2007年行政院衛生署死因統計檔為主。以未滿十歲兒童因非蓄意性傷害死亡者為研究對象。空間統計分析部分，用QUEEN鄰近規則為空間權重矩陣，以全域性空間自相關指標Moran`s I與區域性空間自相關指標LISA(Anselin Local Moran`s I)進行群聚分析比較。此外，將與非蓄意性傷害相關地區因子利用因素分析，萃取出四項地區社經指標，包括社會匱乏、家庭破碎、家庭貧窮以及失業。進一步利用迴歸分析，包括傳統迴歸模型與空間迴歸模型，將空間單元之社經因子納入模型當中，探討與傷害死亡相關的地區因子。 結果：空間分析的研究結果發現，2001-2006年兒童非蓄意性傷害死亡率在臺灣本島地區整體上具有明顯群聚現象，重複出現的區域多出現於新竹、苗栗、南投、花蓮、臺東縣等地區；而2000年與2007年之兒童非蓄意性傷害死亡率則屬隨機分布之情況。若以死亡機轉分類，則機動車交通事故、溺水、窒息死亡均有明顯群聚現象，而跌落或暴露於火或煙死亡則呈現隨機分布之情況。迴歸模型的部分，社會匱乏、家庭破碎、家庭貧窮與地區失業構念性因子均會對兒童非蓄意性傷害死亡率造成正向影響，也就是若所處地區醫療資源匱乏、社會匱乏比例越高、失業率越高、或家庭結構不完整，家庭貧窮較大，均會使得地區當中兒童因非蓄意性傷害而死亡的比例增加；除此之外，整體兒童非蓄意性傷害死亡率具有空間鄰近效應，機動車輛事故死亡率、溺水死亡率、與窒息死亡率也同樣有空間鄰近效應。 結論：居住在醫療資源匱乏、社會匱乏比例越高、失業率越高、或家庭結構不完整，家庭貧窮較大的群聚地區，會造成十歲以下兒童非蓄意性傷害死亡率的增加；而兒童非蓄意性傷害死亡率與特定機轉死亡率具有空間鄰近效應。

Objective：This study is to find out geographical variations of unintentional injury mortality rate aged below 10 in Taiwan by applying spatial analysis, including the detection of cluster(s) of high mortality rate and the comparison of the space-time transitions of high mortality rate cluster(s) among different years. Then, we try to assess the effects of town-level socioeconomic and family status variables for unintentional injury mortality rate. Methods：This study data are obtained from the death statistics of vital statistics data bank from 2000 to 2007 in the Department of Health, Taiwan. We recruit all children, under 10 years old with unintentional injury as the underlying cause of death. We explored the spatial cluster of children unintentional injury death by spatial analysis, including global spatial autocorrelation and local spatial autocorrelation with Queen rule as spatial weights matrix. We collect eleven town-level socioeconomic data and perform factor analysis. We find five principle components as regional socioeconomic indicators, including medical resource shortage, poverty, family fragility, family burden and unemployment. Furthermore, we use regression, including multiple regression and spatial regression, to determine the effect of town-level socioeconomic factors on childhood unintentional injury deaths for a variety of different injury mechanisms and for children of different genders and ages. Result：Our results indicate that there are some significant clusters of childhood unintentional injury mortality from 2001 to 2006 in Taiwan and those regions include Hsinchu, Taoyuan, Nantou, Huanlien and Taitung. However, the distribution of mortality in 2000 and 2007 was random. Some significant clusters are found in childhood injury death from traffic accident, drown and suffocation, but fall and fire/burns. Town-level socioeconomic indicators, including medical resource shortage, poverty, family fragility, family burden and unemployment, contribute to childhood unintentional injury death positively. Besides, our results showed the spatial autocorrelation in overall mortality rate, and mortality in traffic accident, drown and suffocation. Conclusion：High unintentional injury mortality rate is found among children live in regions with poor socioeconomic status, including higher index of poverty, family fragility, family burden, unemployment or medical reousrce shortage. Spatial autocorrelation is found in childhood unintentional injury death and in some injury mechanisms.