本研究針對台灣地區十二年間(1974~1985 )選擇性之六項死因──自殺、缺血性心臟病、結核病、肝硬:化、肺癌、糖尿病之鄉鎮別年齡標準化死亡率進行區位分析。研究結果顯示此六項死因死亡率之鄉鎮差異相當大。其中自殺與結核病及肝硬化之死亡率呈正相關，而與糖尿病、缺血性心臟病及肺癌死亡率呈負相關。進一步對區域社經與死亡率關係的探討發現，社經因素對「自殺－肝硬化－結核」死亡率結集之解釋力較大，其中又以「地方粗死亡率」對男性肝硬化及自殺之正向相關聯較為明顯。至於「糖尿病－缺血性心臟病－肺癌」死亡率之區域社經分析，僅「地區教育程度」和「糖尿病」死亡率呈較明顯之正向關聯。至於以每千人西醫數代表之醫療資源一項，對六項死因死亡率未顯示任何統計上之解釋力。此發現指出區域社會流行病受非醫療因素影響甚鉅，區域性社經結構因素之分佈特別值得注意。此外，目前社區健康服務政策上所採之「中央統一劃一性決策」、及「個人個別化約性之健康服務」特性之不足。故建議社區健康服務應予以「地方化」，根據需求予以規劃，制定更能符合地方上需求之策略。

This study analyzes age standardized mortality rates over twelve years (1974 1985) for six causes of death - suicide, liver cirrhosis, tuberculosis, diabetes, ischemic heart disease, and lung cancer in each Hsiang. The results indicate considerable areal variation in mortality. Among the causes of death, suicide, liver cirrhosis, and TB are positively correlated, and these three causes of death are negatively correlated with diabetes, lung cancer, and ischemic heart disease. Further analysis of the relationship between socio-economic factors and the area's mortality reveals that these factors explain much regional variation in the suicide-TB-liver cir­rhosis mortality cluster. Among these factors general health status is the most powerful explanatory variable for the suicide- TB-liver cirrhosis mortality cluster, while regional educational resources best explain lung cancer-diabetes-ischemic heart disease mortality variation. Somewhat surprisingly, the medical resources factor (as indicated by the number of physicians per thousand in the population) has no explanatory significance. This suggests that non-medical influences - especially the socio-economic structure of the locality - account more for the areal variation in specific health problems. Furthermore, the centralized and universal hea1th policies as well as the individual oriented health services are not adequate to ameliorate area-specific problems. Instead, more localized community health service are suggested to meet local needs.