台灣中老年人失能過程因子、失能軌跡、健康與長期照護利用之研究

The Disablement Process Factors, Disability Trajectories, and Healthcare and Long-term Care Services Utilization among Middle-Old and Older Adults in Taiwan

游曉微(Hsiao-Wei Yu)

指導教授 ： 陳雅美

國立臺灣大學/公共衛生學院/健康政策與管理研究所/博士(2016年)

https://doi.org/10.6342/NTU201602336

摘要

台灣人口老化和老人失能(disability)人口增加的時程快速，有必要系統性瞭解中老年人失能本質，以增加健康與長期照護體系輸送效率。本研究套用失能過程模型(the disablement process model)和安德森健康照護利用模型(Andersen healthcare utilization model)雙理論架構，並導入失能軌跡(disability trajectory)概念，以瞭解台灣中老年人失能全貌。

本研究分為五個目的：第一、回顧台灣中老年人失能趨勢；第二、瞭解台灣中老年人不同的失能軌跡，以及失能軌跡和失能過程因子的關係；第三、依據失能過程模型，探索失能過程因子，包括：前傾特質、內在因子、外在因子，和台灣中老年人失能軌跡的關係；第四、依據安德森健康照護利用模型，探索台灣中老年人失能軌跡和健康與長期照護利用的關係；第五、整合失能過程模型和安德森健康照護利用模型，測試台灣中老年人失能過程因子、失能軌跡、健康與長期照護利用的關係。

目的一方法，是利用美國國家科學委員會評議老人失能調查資料庫品質指標，評價台灣中老年人調查資料庫。在中等至優等品質的調查資料庫中，進一步系統性回顧(systematic review)台灣中老年人近二十年的失能盛行率趨勢。目的二至五方法，研究樣本取自1996 - 2007年台灣中老年身心社會健康生活長期追蹤調查(Taiwan Longitudinal Study on Aging, TLSA)連續四次追蹤調查的中老年樣本(n = 3,429)。目的二至五統計方法包括三種：第一、潛在類別成長曲線模型(latent class growth curves modeling, LCGCM)瞭解台灣中老年人不同的日常生活功能(activities of daily living, ADLs)和工具性日常生活功能(instrumental activities of daily living, IADLs)失能軌跡；第二、成長混和模型(growth mixture modeling, GMM)測試失能過程因子和台灣中老年人失能軌跡的關係；第三、結構方程式(structural equation modeling, SEM)測試失能軌跡的中介效果，對於失能過程因子，包括：人口學因子、共病症數目、憂鬱程度、休閒活動(leisure activities, LAs)，以至於健康與長期照護利用的直接、間接關係。

失能盛行率趨勢的系統性文獻回顧發現：台灣中老年人Nagi功能限制和IADLs失能盛行率，雖然在過去二十年間呈現有升有降，但部分研究發現有下降的趨勢；ADLs失能盛行率趨勢則是一致上升的。失能軌跡研究發現：台灣中老年人分為三種不同的失能軌跡，包括：(1)失能數目在0 ~ 1的健康功能軌跡(n = 2,998, 87.43 %)、(2)失能數目從1陡升至超過10個失能的功能惡化軌跡(n = 363, 10.59 %)、以及(3)失能數目維持在7 ~ 8個失能的穩定失能軌跡(n = 68, 1.98 %)。

失能軌跡的中介效果，對於失能過程因子和照護利用的結果發現：第一、功能惡化軌跡的中介效果，對於失能過程因子，包括：年齡較大(長期照護β = 0.005, SE = 0.000, p < 0.001；住院β = 0.007, SE = 0.001, p < 0.001；急診β = 0.007, SE = 0.001, p < 0.001；門診β = 0.004, SE = 0.001, p = 0.002)、教育程度較低(長期照護β = -0.002, SE = 0.001, p = 0.008；住院β = -0.003, SE = 0.009, p = 0.010；急診β = -0.002, SE = 0.001, p = 0.010；門診β = -0.002, SE = 0.001, p = 0.039)、共病症數目較多(長期照護β = 0.009, SE = 0.002, p < 0.001；住院β = 0.014, SE = 0.003, p < 0.001；急診β = 0.012, SE = 0.003, p < 0.001；門診β = 0.008, SE = 0.003, p = 0.006)，和晚年有較多的照護使用，具有正向關係；第二、穩定失能軌跡，對於憂鬱程度(長期照護β = 0.001, SE = 0.000, p < 0.001；住院β = 0.001, SE = 0.001, p = 0.046)和晚年較多的照護使用，具有正向的中介效果；第三、鼓勵中老年人積極從事休閒活動，不論是動態休閒活動(長期照護：功能惡化軌跡中介作用β = -0.006, SE = 0.003, p = 0.032；穩定失能軌跡中介作用β = -0.003, SE = 0.001, p < 0.001；住院：功能惡化軌跡中介作用β = -0.009, SE = 0.004, p = 0.034；急診：功能惡化軌跡中介作用β = -0.008, SE = 0.004, p = 0.035)或靜態休閒活動(長期照護：穩定失能軌跡中介作用β = -0.003, SE = 0.001, p = 0.003)，將透過避免發展失能軌跡的完全中介效果，間接地減少晚年照護利用。

從眾多失能過程因子中，釐清在年齡、性別、教育等先天(早年)決定因子，以及需要醫療介入的憂鬱症狀、慢性病管控因子之外，鼓勵中老年人積極從事休閒活動，是健康促進並減少晚年照護利用的有效策略。因此，本研究進一步分析台灣中老年人休閒活動軌跡(leisure activities trajectory)和晚年失能的關係。結果發現：第一、台灣中老年人有四種不同的休閒活動軌跡，包括：持續高度活動(n = 722, 22.66 %)、增加活動(n = 84, 2.64 %)、減少活動(n = 201, 6.31 %)、持續低度活動(n = 2,179, 68.39 %)休閒活動軌跡；第二、保持高度活動量(β = -0.99, SE = 0.13, p < 0.001)或持續增加休閒活動(β = -0.88, SE = 0.27, p < 0.001)的軌跡型態，有助於避免中老年人的晚年失能。

總結來說，本研究以「失能過程模型」和「安德森健康照護利用模型」雙理論為架構，突破過去片段式研究盲點，完整地串連台灣中老年人「失能過程因子、失能軌跡、照護利用」關係。台灣中老年人ADL失能盛行率趨勢是上升的，且台灣中老年人具有不同的失能軌跡型態。失能軌跡，對於失能過程因子和晚年健康與長期照護利用，具有中介效果。鼓勵中老年人積極地從事休閒活動，有助於促進健康，並且降低晚年照護利用。

關鍵字

失能軌跡 ； 休閒活動軌跡 ； 長期照護 ； 潛在類別成長曲線模型 ； 中老年人 ； 台灣

並列摘要

Taiwan has a fast-aging population with an associated increase in the number of older adults with disability expected. A strategic understanding of the nature of disability in middle-old and older adults makes the delivery of the healthcare and long-term care (LTC) system more productive and efficient. The study framework employed in this study is based on two theoretical models, the disablement process model and the Andersen healthcare utilization model, and integrates both models by applying the current knowledge about disability trajectory measurement in a sample of middle-old and older Taiwanese adults.

The study includes five aims: (1) to review the disability trends of middle-old and older adults in Taiwan; (2) to identify distinct disability trajectories and their associated disablement process factors in middle-old and older Taiwanese adults; (3) to understand, according to the disablement process model, how predisposing factors, intra-individual factors, and extra-individual factors influence disability trajectories in middle-old and older Taiwanese adults; (4) to examine, according to the Andersen healthcare utilization model, the association between disability trajectories and later use of healthcare and LTC services; (5) to understand, by integrating both the disablement process model and the Andersen healthcare utilization model, the relationships among disablement process factors, disability trajectories, and later use of healthcare and LTC services.

For aim 1, we applied referencing criteria established by the US National Research Council to assess the quality of several Taiwanese aging surveys. Among the aging surveys judged to be of high quality, we further systematically reviewed the disability trends of the past two decades among middle-old and older Taiwanese adults. For aims 2 to 5, we collected data from four successive waves of the Taiwan Longitudinal Study on Aging Survey (TLSA) in 1996-2007 (n = 3,429). The study methods for aim 2 to 5 included three statistical approaches: latent class growth curves modeling (LCGCM) was used to identify the distinct trajectories of activities of daily living (ADLs) and instrumental activities of daily living (IADLs) disabilities; growth mixture modeling (GMM) was used to examine the related disablement process factors and disability trajectories; structural equation modeling (SEM) was applied to examine the mediating effect of disability trajectory on disablement process factors, such as demographics, number of comorbidities, depression, and leisure activities (LAs), and on later LTC service use.

The results in reviewing disability trends of middle-old and older Taiwanese adults showed that the trends of Nagi’s functional limitations and IADLs disability were both mixed but exhibited declining prevalence; however, ADLs disability showed a trend of consistent increase in Taiwan during the reviewed years. Analysis of ADLs and IADLs disability trajectories in middle-old and older Taiwanese adults revealed three different subtypes: (1) a healthy function trajectory represents those middle-old and older participants who had less than one disability over the study period (n = 2,998, 87.43 %); (2) a progressive disability trajectory represents the older participants who reported an increase, from one at the beginning of the study to ten at the end in the number of disabilities they had during the study period (n = 363, 10.59 %); (3) a maintained disability trajectory represents the middle-old and older participants who had maintenance of seven to eight disabilities throughout the study period (n = 68, 1.98 %).

The results of the analysis of whether the disability trajectories of Taiwanese middle-old and older adults mediated the association between disablement process factors and later LTC service use showed that: (1) a progressive disability trajectory had a positive mediating effect on the association between disablement factors (older age, lower educational attendance, greater number of comorbidities) and greater use of healthcare and LTC services; (2) a maintained disability trajectory was found to be a positive mediator between depression and greater use of healthcare and LTC services; (3) in contrast, encouraging middle-old and older adults to engage in LAs reduced later use of LTC services through a significant mediating effect via a healthy function trajectory in middle-old and older Taiwanese adults.

In spite of unmodifiable risk factors, such as age, gender, and early-life educational attendance, and in spite of factors which require medical intervention, such as chronic disease and depressive syndrome, we found that encouraging middle-old and older adults to participate in LAs is an effective way to promote health and decrease healthcare and LTC service use in later life. Therefore, we further analyzed the relationship between LA trajectories and disability in middle-old and older Taiwanese adults and found that: (1) middle-old and older Taiwanese adults had four distinct LA trajectories, namely maintained high participation (n = 722, 22.66 %), increased participation (n = 84, 2.64 %), decreased participation (n = 201, 6.31 %), and maintained low participation (n = 2,179, 68.39 %) subtypes of LA trajectories; (2) middle-old and older people who were identified in the maintained high participation and increased participation subgroups of LA trajectories were less likely to be disabled.

In conclusion, ADLs disability in middle-old and older adults shows an overall trend of increase in the past two decades in Taiwan with heterogeneous patterns of disability trajectory found in middle-old and older Taiwanese adults. Disability trajectory had a mediating effect on the association between disablement process factors and later use of healthcare and LTC services. Encouragement of middle-old and older adults to participate in LAs was an effective way of promoting health and decreasing service use in late life.

並列關鍵字

Disability Trajectory ； Leisure Activity Trajectory ； Long-Term care ； Latent Class Growth Curves Modeling ； Middle-Old and Older Adults ； Taiwan

參考文獻

Andersen, R. M. (1995). Revisiting the behavioral model and access to medical care: does it matter? J Health Soc Behav, 36(1), 1-10.

Bentler, P. M. (1990). Comparative fit indexes in structural models. Psychological Bulletin, 107(2), 238-246.

Aday, L. A., Andersen, R. (1974). A Framework for the study of access to medical care. Health Services Research, 9(3), 208-220.

Agahi, N., Lennartsson, C., Kareholt, I., Shaw, B. A. (2013). Trajectories of social activities from middle age to old age and late-life disability: A 36-year follow-up. Age Ageing, 42(6), 790-793.

Al Snih, S., Markides, K. S., Ray, L. A., Freeman, J. L., Ostir, G. V., Goodwin, J. S. (2006). Predictors of healthcare utilization among older Mexican Americans. Ethn Dis, 16(3), 640-646.