目標：本文控制地區人口特性和社經條件，以台灣各縣市醫療資源分配相對於當地地理幅員和人口數為指標，探討相關因子與各地粗死亡率和事故傷害死亡率差異的關係。方法：本研究利用1998至2007年台灣地區23縣市的跨時跨區資料，以羅倫茲曲線和吉尼係數呈現醫療資源分佈的不均程度，並透過泊松迴歸(Poisson regression)方法分析醫療資源密集度、就醫交通可近性、人口組成、教育和所得等因素，對於各縣市粗死亡率、事故傷害死亡率的影響效果。結果：在影響兩項死亡率指標的相關因子中，以醫療的交通可近性、人口老化程度和社會經濟狀況三類影響因素最為重要，醫療資源相對於人口的密集度、原住民人口兩特性，則非主要因子。結論：在控制人口組成、社會經濟條件等因素後，台灣各縣市間粗死亡率和事故傷害死亡率的差異，可相當程度歸因於各地醫療資源、地理幅員所衍生之交通可近性；而且交通可近性條件對於事故傷害死亡率的影響更甚於對粗死亡率的影響。

Objectives: After controlling for population attributes and socio-economic factors, this article explored the allocation of medical resources among 23 counties in Taiwan and investigated the impact factors on the crude and accidental death rates. Methods: This study collected data from 23 counties in Taiwan for 1998 through 2007 and used the Gini coefficient and Lorenz curve to show the inequality of resource allocation among all counties. With Poisson regression, the crude
and accidental death rates were dependent variables and resource concentration, access to medical facilities, population attributes, level of education, and income were treated as independent variables. Results: Access to medical facilities, age, and socio-economic status were the most
influential and significant factors associated with the two death rates. Resource concentration and the aboriginal index were less important. Conclusions: After controlling for other factors, traffic access to medical resources was an important factor in the crude and accidental death rates for 23 counties in Taiwan, more for the accidental death rate than for the crude death rate.