一般認為醫療費用隨著年齡增加而增加，因此人口老化將加重醫療費用上漲壓力。然而，部分研究指出老年人的平均醫療費用比年輕人高的關鍵不只是年齡較大，還有老年人較年輕人更接近死亡，因此主張討論人口老化與醫療費用的關聯時，必須納入「死亡距離」的考量。本文從死亡距離角度分析臺灣不同年齡人口之死亡距離別醫療費用的分佈，以檢視年齡、死亡距離與醫療費用三者的關聯。最後結合人口推計結果，比較年齡與死亡距離兩種取向下未來全民健保整體支出的趨勢。研究結果顯示，個人醫療費用的高峰集中於死亡前一段時間，而死亡年齡越高，死亡距離別平均醫療費用越低，隱含壽命延長一方面因增加高齡人數而對整體醫療費用帶來壓力，另一方面也對整體費用的上漲有遞延作用。此外，從歷年資料來看，平均每人醫療支出開始快速增加的時間與死亡時間的距離逐漸擴大，顯示罹患疾病的時間拉長，又加大醫療支出。綜合而言，總體醫療費用將隨著人口老化而逐漸成長，但是費用推計的結果顯示年齡模型的估計值明顯較死亡距離模型估計值來得高，顯然單從年齡取向探討人口老化與醫療費用的關聯，容易高估人口老化對醫療費用的影響。

The effects of ageing on health care spending are uncertain. Some argue that health care expenditures increase substantially with age primarily because mortality rates increase with age, and expenditures increase with closeness to death. The main reason why healthcare costs increase with age is that older people are closer to death than younger people. The pattern of health expenditure by age is considerably influenced by the concentration of expenditure in the final years of life. The greater expenditure on the elderly is a consequence of the heavy weight of so-called death costs. In addition, increases in longevity may be expected to lead to postponement of these costs of the final years of life, and declines in age-specific mortality may be expected to lead to declines in age-specific costs because declining mortality reduces the proportion of those near death. If we overlook this aspect, we envisage a misleading scenario of health expenditure increase.Incorporating the concept of time to death, this paper reexamines the relationship of population aging and healthcare expenditure in Taiwan. Results indicate that different death age groups have similar average expenditures patterns against time to death, but expenditures are higher for younger people than for older people. These imply that increases in longevity delay the high costs associated with the final stage of life. Projections failing to take time to death into account might result in upward biased simulations, while total expenditures will still be raised by the absolute number of elderly people.