透過您的圖書館登入
IP:3.227.239.9
  • 學位論文

護理之家照顧服務品質之多層級分析_以中部地區四縣市為例

Quality Of Care For Nursing Homes In Middle Taiwan: A Multi-Level Analysis

指導教授 : 鄭讚源
若您是本文的作者,可授權文章由華藝線上圖書館中協助推廣。

摘要


隨著老人人口快速成長,長期照護問題已是世界各國非常重視的議題,在許多已開發國家中已將提升長期照顧服務品質列為最優先政策目標,而國內機構式照護場所快速增加,其照顧服務品質的提升是相當重要的,而品質是多面向、多層級且動態的,無法以單一指標衡量機構品質。本研究以Donabedian提出結構、過程及結果三面向的照護品質模式為基礎,結合Scott組織理論模型中開放系統的觀點,並加入Stufflebeam發展之CIPP評鑑模式概念(包含環境脈絡、輸入、過程、成果變項)來建構本研究架構。 研究目的為瞭解護理之家住民特性、組織特性、組織脈絡特性、環境特性與服務品質的現況,並釐清品質的多層面特性,以及分別以單一層次及多層次分析的角度,檢視住民特性、組織特性、組織脈絡特性、環境特性對服務品質的影響。 本研究採橫斷性多層級研究設計,使用全民健康保險醫療院所申報資料及全國第一次護理之家評鑑資料進行次級資料分析,以中部地區(台中縣市、彰化縣及南投縣)的一般型護理之家為研究對象(共78家機構、6210位住民)。服務品質研究是具有多層面多層次現象,所以本研究針對多層面(multi-dimensions)品質採因素分析,多層次(multi-level)品質採HLM(Hierarchical Linear Modeling)階層線性模式分析。 本研究結果發現:當服務品質係為單一層次時,因素分析可抽離為九大層面,當吾人以其中第二層面(住院醫療品質因素)為依變項時,床數、評鑑結果為合格、醫療可近性、占床率、城鄉因素、機構平均住民疾病嚴重度、機構平均住民身心障礙比率等為其重要影響變項,解釋力達50.94%。其次,當服務品質為多層次概念時,用因素分析可抽離為五大層面,本研究分別以住民、組織、環境、組織脈絡特性分別加以分析,經由不同模式組合後最後發現的模型中,在個體層次的解釋變項方面,其中失能程度對五大層面住民服務品質皆具有影響,另住民總住日數、年齡、身心障礙程度、疾病嚴重度、女性、有重大傷病者,則對各層面的住民服務品質有不同的影響力;在組織層次方面:部分之組織特性(機構與醫療院所同棟樓層或位於隔壁)及組織脈絡特性(住民入住天數、疾病嚴重度、失能、身心障礙)對住民服務品質具有脈絡效果;簡而言之,個體層次與組織層次對住民服務品質皆會有影響。而個體層次變項解釋住民服務品質13.45%~36.47%的組內變異,組織層次則解釋 9.59%~52.22%的組間變異,透過實證研究護理之家住民巢套於組織環境下,部分的組織特性(組織屬性為附設型、經營型態為私立)、組織脈絡特性(平均住民總入住天數、平均失能率、機構住民失能變異、機構住民身心障礙變異)、環境特性(每萬人口西醫家數)對個體層次解釋變項與服務品質的關係具有跨層級交互作用(調節效果)。 另在複迴歸分析中有顯著差異的解釋變項,在多層級分析中卻呈現無顯著差異,顯示採用多層級分析是較佳的選擇,當資料有巢套關係(nested)時,透過多層次分析將可彌補傳統單一層次模型分析所可能違反誤差獨立性假設的問題,並可以做為正確推論。實證分析得到個體層次與組織層次變項對個體層次服務品質的影響力。 藉此實證研究在學術理論的貢獻為驗證社會科學領域階層結構資料鑲崁、巢套(nested)的性質,分析個體(微觀)層次與組織(鉅觀)層次變項對微觀層次品質的影響力。從研究中可對品質更加釐清,因其可以展現於個體、組織、環境不同層次中,而且變項結構間亦有動態關係存在。同時也驗證了CIPP模型脈絡的優勢,運用脈絡、輸入、過程和成果評鑑,提供主管機關及機構去檢視照顧服務品質,若僅運用Donabedian的結構過程結果模式,未考量脈絡重要因素,則在社會科學研究中略屬不足。另結合照顧相關理論、組織理論及評鑑模型等建立品質多層面、多層次、動態等概念的研究,實證回應理論之多層次的重要性。 在方法學上的貢獻為運用多層級分析提供我們在個體層次與組織層次的關注。以往因方法學上無法突破,故較少同時處理個體層次、組織脈絡層次的資料,而本研究真正從方法學上(多層級分析) 釐清服務品質受到個體層次、組織層次的影響,研究中亦考量組織脈絡特性、環境特性等變項,將可強化品質研究涵蓋面不足。 在政策管理的貢獻為透過個體層次、組織層次、組織脈絡等釐清,瞭解哪些組織、環境、脈絡特性直接影響品質,進而協助機構經營者或管理者加強內部管控,協助組織維護及提升照護品質,同時透過評鑑的外控機制來確保品質提升。另相關主管機關對長照機構之各層面的品質應加強管理與監控。未來在不同層次之品質管理政策亦應有所不同,以提升長照住民服務品質。

並列摘要


Recent trends have seen the rapid growth in elderly population and the resulting emphasis in long-term care arrangements and policy-making, especially in developed countries. Among research issues in long-term care, institutional care and its quality have always been the major concern and also most investigated. Previous studies have focused on various aspects in institutional care quality. However, few research has treated the quality of institutional care as both a multi-dimensional as well as a multi-level phenomena. Thus, taking a multi-dimensional and a hierarchical approach, this study tries to investigate the possible effects of the following variables on the quality of nursing homes: residential characteristics, organization characteristics, contextual effects, environment characteristics. The quality of care in the nursing homes was approached from a conceptual model deriving from the integration of arguments from various sources: Donabedian's structure-process-outcome model, organizational theory, and CIPP model in evaluation. Then factor analyses, traditional regression analysis and HLM are used to test the empirical data. This study used secondary data. The outpatient and inpatient reimbursement claim data of the National Health Insurance in 2009 was retrieved. The effects of independent variables at individual and organizational levels on nursing home quality were analysed using multilevel models(Hierarchical Linear Model) .A total of 6210 residents from 78 nursing homes were selected from middle Taiwan in a cross sectional design. Initial analysis shows that resident characteristics are predictors of the outcome variable at the individual level. When accessed at a single level, factor analysis resulted in nine dimensions for the quality of care in nursing homes. Among them, when the second dimension (quality of inpatient medical care) was used as the dependent variable, a qualified value on the number of beds and the evaluation results was revealed while the medical accessibility of, occupancy rate, urban and rural factors, organizational residents’ average illness severity, organizations residents’ average disability rate served as important variables with an explanatory power of 50.94%. Besides, predictor slopes are influenced by organization characteristics, contextual variable, environmental characteristics. The quality of care in nursing homes are then factor-analyzed into five dimensions when accessed as a multi-level phenomena.HLM analysis result revealed that the one of the individual-level explanatory variable, the degree of disability, is influential in all five dimensions of the quality of residential care.Also, total duration of residential occupancy, age, degree of disability, disease severity, gender (women), and catastrophic illness each holds influences to different dimensions of the quality of residential care. Organizational characteristics (when institute is in close proximity to the medical organization) and organizational contextual characteristics (the total duration of residential occupancy, disease severity, disability, and mental or physical disorders) also holds contextual effects on the quality of residential care. Besides, HLM results also show interactive effects of the independent variables. In all, the present study’s original proposition that quality of care should be treated as both a multi-dimensional and a multi-level phenomena is basically empirically validated. From a theoretically point of view, this research will extend our understandings in the fields of organization theory, quality of care, evaluation, and even sociology by assuming and then validating a multi-dimensional and multi-level social construct. And the findings are also important in other two fronts: policy/management, and methodology. Policy makers and managers of long-term care institutions could learn from this study that individual variables from both the individual and organizational can have effects on the quality of care, and thus they can formulate and implement relevant policies or managerial regulations. The use of HLM also distinguishes statistical effects from different levels correctly, an advantage over traditional regression analysis.

參考文獻


劉麗雯(2002)。《老人長期照護機構網絡建立評估之研究》。內政部社會司委託研究。
劉佳鑫(2008)。《以多層次分析探討個人和醫師層次特質對新診斷無併發症糖尿病病患發生相關併發症之影響》。高雄醫學大學健康科學院公共衛生學系碩士班流行病學組碩士論文。
詹婷婷(2007)。《探討地區因素對基隆市民眾篩檢行為之影響—應用多層次分析結果》。中山醫學大學公共衛生學系碩士論文。
戴桂英(2007)。《中風出院病人後續醫療照護之研究》。台北:台灣大學衛生政策與管理研究所博士論文。
王卿蘭(2004)。《老人長期照護機構服務品質與生活品質之研究-以台北縣為例》。台北:臺灣大學醫療機構管理研究所碩士論文。

延伸閱讀