本篇論文使用行政院衛生署家庭計畫研究所執行的「民國八十八年台灣地區中老年身心社會生活狀況長期追蹤調查」資料，我們希望瞭解獨居與非獨居老人的醫療利用行為是否存在差異性；研究議題包含了比較獨居與非獨居老人利用醫療服務的機率，以及在已就醫情況下，比較獨居與非獨居老人的醫療利用量多寡。由於存在不可觀察的因素，如老人本身不可觀察的健康狀況，同時影響老人的醫療利用行為與居住安排決策，因此本文為了修正醫療利用模型中可能存在的內生性問題，以「子女特性變數」做為工具變數，以幫助確認醫療利用模型中居住安排的估計係數。 本研究的實證結果顯示，獨居老人使用住院服務及居家護理服務的機率顯著高於非獨居老人，而且在已就醫情況下，獨居老人的住院天數以及利用居家護理服務的次數顯著高於非獨居老人。由於住院服務需要使用密集的醫療資源，居家護理服務為長期性的醫療負擔，因此本文建議透過政策減少老人獨居的機會，如協助老人再婚、鼓勵子女與父母同住，或是透過政策讓老人集體居住於特定社區，輔導老人組織互助團體，如此有助於降低社會的醫療財務負擔。 The purpose of this paper is to investigate if there exists different effects between two living arrangements --the elderly living alone or living with their family--on the medical utilizations of the elderly. Based on the 1999 Survey of Health and Living Status of the Elderly in Taiwan, we applied a two-part model to our study. In the first part, we compared the effect of living arrangements on the probability of using medical services among the elderly; in the second part, we compared the effect of living arrangements on the quantity used by the elderly conditional on having used medical services. Because it possibly exists unobservable factors that both affect the medical utilization and living arrangements of the elderly-- such as the unobservable healthy status, we use the characteristics of their children as instrumental variables to correct this kind of endogeneity problem. Findings show that the elderly living alone are more likely to use hospitalization and homecare services than those living with their family. While conditional on having used medical services, the elderly living alone apparently used more than those living with their family.
