日本社會於2007 年邁入超高齡社會。在此階段，平均壽命延長、高齡人口增加、國民醫療費與長期照護費用攀升等多重現象同時進行。本研究基於「健康 轉換」的概念，選取慢性疾病、衰老疾病、健康壽命、平均壽命、高齡化與長期照護費用作為代理變數，構建一個由六個方程式組成的長期照護需求計量模型。 研究使用2000 年至2020 年的資料，分別採用多元線性迴歸（OLS）與廣義矩估計法（GMM）進行估計。結果顯示，調整後的決定係數達0.9，主要解釋變數，均 具有統計顯著性（P<0.01）。其中，平均壽命與衰老疾病被確認為直接影響長期照護需求變動的因素。此一研究成果不僅對日本長期照護政策的制定具有參考價 值，同時亦能為正邁向超高齡社會的台灣，在未來長期照護需求的估計與制度設計上，提供重要的啟示。

Japan entered the stage of a super-aged society in 2007. In such a society, multiple phenomena have been progressing simultaneously, including the extension of life expectancy, the increase in the elderly population, and the rising costs of national medical care and long-term care. This study, based on the concept of "health transition," constructs an econometric model of long-term care demand comprising six equations, with proxy variables such as chronic diseases, senescence-related diseases, healthy life expectancy, average life expectancy, aging, and long-term care expenditures. Using data from 2000 to 2020, estimations were conducted through multiple linear regression (OLS) and the generalized method of moments (GMM). The results showed that the adjusted coefficient of determination reached 0.9, and the main explanatory variables were statistically significant (P < 0.01). In particular, average life expectancy and senescence-related diseases were identified as direct determinants of changes in longterm care demand. These findings not only contribute to the formulation of long-term care policies in Japan but also provide important implications for the projection of future long-term care demand and the design of relevant systems in Taiwan, which is also transitioning into a super-aged society.