隨著醫院權屬別不同，醫院享有的資源也會有所不同，因此，本文的研究目的為以肝癌病患做為研究對象，來檢驗醫院權屬別和醫療結果的關聯性。本文的資料來源主要為全民健保資料庫1996-2007年的 「2000年承保抽樣歸人檔」 ，並透過門診費用檔(CD)、住院費用檔(DD)、承保資料檔(ID)、醫事機構主檔(HOSB)和重大傷病檔(HV)找出肝癌病患的個人特性變數、醫院特性變數等變數，並以肝癌病患的1個月內死亡率、6個月內死亡率和1年內死亡率做為衡量醫療結果的指標。在實證分析時，由於肝癌病患可能依據自我疾病嚴重程度來選擇醫院，進而造成樣本自我選擇的問題，因此本文以兩階段最小平方法做為實證模型，工具變數則選擇肝癌病患距離最近的公立醫院距離、最近的非營利醫院距離、治療肝癌醫院家數和男性人口數四個工具變數。 本文的實證結果得到，在使用工具變數調整過後，和營利醫院相比，不管是公立醫院或非營利醫院，對1個月內、6個月內和1年內死亡率都沒有顯著的影響，顯示肝癌病患的死亡率並不受到醫院權屬別影響，表示肝癌病患的死亡率可能受到本身疾病嚴重程度的影響，甚於在醫院獲得的治療。This paper analyses the relationship of hospital ownership and health outcome on liver cancer patients in Taiwan. We use the National Health Insurance Bureau claims data from 1996 to 2007. According to the claim data, we extract the hospital characteristics like hospital accreditation, ownership status, and hospital location and demographic covariates such as patient’s gender, age, income, and health status as condition for the following process. Then we use one-month, six-month and one-year mortality rates as index to measure patient’s health outcome. Because patients may select admitted hospital based on their illness severity, we employ the two-stage least square model to account for this problem. And the instrument variables we use in this paper are the following four, the nearest distance to public hospital, the nearest distance to nonprofit hospital, male population by county and treating liver cancer patients’ hospital numbers by county. From the results of two-stage least square model estimation, compare to for-profit hospital, either public hospital or nonprofit hospital, we find that there is no significant influence on one-month, six-month and one-year mortality rates in statics. Based on the static result above, we will make a conclusion that the health outcome is affected by patient’s illness severity more than hospital care does.
