台灣的存活曲線自2006年開始逐漸呈現矩形化，也就是目前台灣人口死亡時間多發生在老年時期，而且越到生命後期，存活率下降越明顯；因此，若能找出影響老人死亡關鍵因素，將有助改善老人健康與延長平均餘命。本文以台中榮總高齡醫學中心1248名追蹤資料為例，探討老人健康折舊與死亡之關聯性。在健康折舊因子方面，包含9項區間變數、17項名目變數及4項控制變數，利用兩個獨立母體差異性檢定及二元邏輯斯迴歸等實證方法。主要發現心肌梗塞、血中白蛋白、迷你營養評估及血紅素之增減，對死亡邊際機率，分別為0.072、-0.037、-0.01與-0.009；結果顯示老人死亡風險增加，以心肌梗塞影響效果最大，但血中白蛋白、迷你營養及血紅素增加，可顯著降低老人之死亡風險。綜合以上，針對老人整合性照護體系死亡率的改善，應預防老人心肌梗塞、營養不良、貧血與低白蛋白血症。

Survival curve of Taiwan gradually became rectangularization since 2006. It means that most death of people occurs in the old stage, and the survival rate declines significantly getting toward life end. So researching factors related to elderly death will improve the health and life expectancy. Sample of this study is obtained from the Center for Geriatrics and Gerontology of Taichung Veterans General Hospital including 1248 patients. This study aims at investigating the relationship among health depreciation and mortality. The health depreciation factor includes 9 intervals variables, 17 nominal variables and 4 control variables, and utilize empirical method of two independent populations (death or not), statistical hypothesis test, and binary logistic regression (regression coefficient, odds ratio and marginal probability). To mortality, the marginal probability of myocardial infarction, serum albumin, mini nutrition assessment and hemoglobin are 0.072, -0.037, -0.01 and -0.009. It shows that myocardial infarction most significantly increases mortality risk, but serum albumin, mini nutrition assessment and hemoglobin decrease mortality risk. In conclusion, prevention of myocardial infarction, malnutrition, anemia and hypoalbuminemia will improve mortality of elderly integrated health care system.