中文版工作者角色會談之心理計量品質研究—以台灣精障者為例

Abstract

前言：工作者角色會談根據人類職能模式發展而來，為半結構化之會談。當中包含會談與標準化之評分表格。工作者角色會談被用來評估傷病或長期失能之工作者，以了解工作者重返職場之社會心理與環境支持度。工作者角色會談之評分表一共有16項，分別涵蓋個人歸因、興趣、價值、習慣、角色與環境六個內容範疇。 研究目的：將英文版工作者角色會談中文化，並用於台灣精障者以驗證其信、效度。 研究方法：本研究採用正翻、反翻與委員會的方式對工作者角色會談進行翻譯，以建立中文版工作者角色會談(Worker Role Interview-Chinese version, WRI-C)。評分定義的部分包含項目與評分標準，共112項。第一次正翻後修正的項目共有26項，佔所有項目之23.21％。經過反翻後，再次對11項進行修訂，佔所有項目之9.82％。之後，進行第二次反翻並確定文意相似度，建立中文版初試版本。為了建立WRI-C之效度，由一位職能治療師對100位年紀18至65歲之精障者進行台大症狀量表、世界衛生組織生活品質量表—臺灣簡明版、職能自我評估、貝氏憂鬱量表第二版、WRI-C、加拿大職能表現測驗、人類職能模式篩選量表與意志量表之施測。另外，前40位受測者於施測後6~13天內，再次接受WRI-C之施測以建立再測信度。 資料分析：本研究使用WINSTEPS（3.56版）、AMOS（7.0版）與SAS（9.1版本）進行信、效度驗證分析。信度的部分，使用加權kappa係數對WRI-C各項目原始分數進行分析，以了解兩次量測的一致性（agrrement）。並使用WRI-C各範疇之羅序分數進行組內相關係數（Intraclass correlation coefficient, ICC）分析，以了解再測信度。建構效度的部分，使用羅序分析模式及驗證性因素分析進行檢驗。羅序分析除了可以驗證建構效度外，還可將量表之原始分數轉換為等距分數。同時效度的部分，使用WRI-C各範疇之羅序分數分別對相關評估工具進行皮爾森相關係數（Pearson’s correlation coeffincient）之分析。最後使用變異數分析（one-way ANOVA）檢驗不同工作能力之個案於WRI-C各範疇得分之差異，以驗證區辨效度。 結果：除「同事的看法」外，各項加權Kappa均為0.4到0.6之間，顯示具有可接受之項目一致性。ICC值的部分，WRI-C整體量表為0.76、意志範疇為0.69、習慣範疇為0.65與環境範疇為0.71，顯示具有可接受之再測信度。建構效度方面，研究結果顯示可將WRI-C分成意志、習慣與環境三種建構。意志、習慣範疇具有可接受之建構效度，環境範疇則較差。同時效度方面，WRI-C意志範疇分別與世界衛生組織生活品質問卷-臺灣簡明版的心理範疇、職能自我評估自我部分、加拿大職能表現測驗、意志量表及人類職能篩選量表有顯著低度到高度正相關；WRI-C習慣範疇與人類職能篩選量表、職能自我評估自我部分具有顯著中度正相關；WRI-C環境範疇與職能自我評估環境部分具顯著中度正相關；另外，三個範疇均與貝氏憂鬱量表具顯著中度負相關，與台大症狀量表不具顯著相關，表示WRI-C各範疇具有可接受至良好之同時效度。依工作能力分組，不同組間之WRI-C習慣與環境範疇之分數具有顯著差異，由事後檢定得知其主要差異來自於「進行職業訓練，合併就業輔導」與「進行庇護性之工作，合併任務導向之治療性團體活動」、「未接受工作訓練，接受任務導向之治療性團體活動」兩組的不同，顯示WRI-C的習慣與還就範疇具有良好之區辨效度。

Introduction: The Worker Role Interview (WRI) is a semi-structured interview rating scale developed based on the Model of Human Occupation. The WRI can be used to evaluate the psychosocial and environmental components related to return to work for those persons with injury or chronic disease or disability. The worker role interview contains six content areas including personal causation, interest, value, habit, role and environment, totaled as 16 items. Objectives: The purposes of this study Included: to translate the worker role interview into Chinese according to two-stage translation procedure; to examine the reliability and validity of the Chinese version of the worker role interview. Method: We adapted “Forward translation”, “Back translation” and “Committee approach” to arrive at the final Chinese version. For the test of the psychometric qualities of the worker role interview, one hundred subjects with mental illness were recruited to participate in the study. The National Taiwan Hospital Syndrome Scale(NTUSC), World Health Organization Quality of Life BREF Taiwan version(WHOQOLBREF-TW), Occupational Self Assessment(OSA), Beck Depression Invetory II(BDI-II), Canadian Occupational Performance Measure(COPM), Model of Human Occupation Screening Tool(MOHOST) and Volition Questionnaire (VQ) were used to investigate the concurrent validity of the worker role interview. A set of forty subjects within the whole sample were tested for the test retest reliability study. Data analysis: Comfirmatory factor analysis and Rasch analysis were used to examine the construct validity of the scale. Correlational analysis was applied for concurrent validity. Intraclass correlation coefficient and weighted kappa were utilized for test retest reliability. Finally, the analysis of variance was used for known group validity. We applied the SAS (version 9.0), Amos (version 7.0) and WINSTEPS (version 3.56) computer programs for data analysis. Results: The WRI-C contains volition, habituation and environment domains. Except for environmental domain which needs to be modified in the future to fit the measurement model, volition and habituation domains demonstrated qualitity of unidimentionality.The test-retest reliability of the WRI-C is acceptable. The concurrent validity of the volitional and habituation domains of the WRI-C were supported by its significantly moderate to high correlation with related masures. Finally, there were significant differences of the two domains of the WRI-C among three groups of persons with varying level of functioning which supported the known group validity of the WRI-C.