人類職能模式篩選量表之信效度探討

Abstract

研究背景 根據各項統計顯示，罹患精神疾患之人數逐年上升；在此族群當中，超過一半的個案為適齡工作者。精神障礙者復健的最終目標之一是重回職場並適應社會生活。然而，由於疾病的限制以及機能退化，許多精神障礙者之職業功能受限；因此，面臨長期失業或無法維持工作的局面。 多面向的評估工具可較完整地評量精神障礙者之工作潛能與工作限制。本研究翻譯以人類職能模式(the Model of Human Occupation)為基礎之多面向評估工具─人類職能模式篩選量表(the Model of Human Occupation Screening Tool, MOHOST)，並探討其應用於適齡工作精神障礙者之適用性及心理計量特性。 研究目的 翻譯人類職能模式篩選量表，驗證中文版人類職能模式篩選量表(MOHOST-C)應用於參與不同工作型態之精神障礙者之心理計量特性，包括：內在一致性、施測者間信度、建構效度及同時效度；並檢驗此量表對於性別的試題功能差異，以及不同工作型態個案之職能參與程度。 方法 本研究經歷兩次反翻譯產生中文版人類職能模式篩選量表。由101位介於18-65歲之精神障礙者參與研究。個案之工作類型分別為：臨床職能治療團體28位、職業訓練41位、庇護性就業20位及支持性就業12位。研究者使用簡短式智能評估(MMSE)進行篩選，當個案分數大於24分時，納入收案。共有4位職能治療師參與此研究，40位個案由兩位以上評分者同時進行觀察，並獨立進行人類職能模式篩選量表之評分。 研究者於自然情境中，觀察個案參與工作或職能治療活動之表現20-30分鐘。待工作結束後，與個案進行訪談，訪談內容為個案對於參與工作或活動之觀點，以及其對於工作或活動環境之看法。訪談結束後請個案填寫相關問卷。包含：中文版職能自我評估(OSA-C)、台大症狀量表(NTUHSC)。另外，研究者則根據觀察個案在工作或活動中的表現與訪談結果，進行中文版人類職能模式篩選量表(MOHOST-C)、中文版意志量表(VQ-C)及中文版溝通與互動技巧測驗(ACIS-C)之評分。 資料分析 本研究使用SPSS (11.5版本)與Facets (3.62.0版本)進行信效度驗證分析。 中文版人類職能模式篩選量表之分數經羅序模式(Rasch model)轉換為等距分數，再進行統計分析。以項目分離信度(Item separation reliability)檢驗內在一致性；以羅序分析檢驗施測者間信度及建構效度；並探討中文版人類職能模式篩選量表對於不同性別個案之結果差異。另外，本研究使用羅序模式轉換中文版人類職能模式篩選量表(MOHOST-C)、中文版意志量表(VQ-C)、中文版溝通與互動技巧量表(ACIS-C)及中文版職能自我評估(OSA-C)之分數，以皮爾森相關係數(Pearson’s correlation coefficient)分析其同時效度。以斯皮爾曼相關係數(Spearman’s correlation coefficient)分析中文版人類職能模式篩選量表與台大症狀量表(NTUHSC)原始總分之同時效度。另使用變異數分析(one-way ANOVA)，檢驗不同工作型態之個案於中文版人類職能模式篩選量表之得分差異。 結果 本研究結果顯示，人類職能模式篩選量表具有良好之內在一致性。建構效度方面，本量表可將職能參與區分為綜合自我能力與環境支持度兩種建構。同時效度方面，受試者在「職能動機」、「溝通與互動技巧」次量表之羅序分數與「中文版意志量表」、「中文版溝通與互動技巧量表」之羅序分數均達顯著相關。但中文版人類職能模式篩選量表與「中文版職能自我評估」之羅序分數雖呈現正相關，卻未達顯著；另與「台大症狀量表」之分數亦無顯著相關。施測者間信度方面，施測者符合羅序測量模式之建構，證實職能治療師可藉由閱讀「中文版人類職能模式篩選量表」手冊後，有效且正確地評分。本研究亦發現「姿勢與移動性」及「力量與費力度」兩題項對於男女受測者有不同之反應。另外，接受不同層級工作相關訓練之精障者，在中文版人類職能模式篩選量表之表現有顯著差異：接受職能治療團體之個案，其職能參與程度較接受庇護性就業及支持性就業之個案低；接受職業訓練之個案，其職能參與程度較接受支持性就業之個案低。

Background Increasing numbers of people around the world are developing mental illness. Of the people comprising this large group, approximately half are adults, who should be producers or workers in their current life stages. A major consequence of psychiatric symptoms and functional degeneration in clients with psychiatric disorders is unemployment. A multidimensional assessment could be used to evaluate the holistic work participation of clients with psychiatric disorders and then to help determine their work potential and limitations. In this study, we chose the MOHOST as the assessment tool. We translated it into Chinese and then examined if the MOHOST-Chinese version (MOHOST-C) was an appropriate tool for evaluating the work participation of clients with psychiatric disorders in work or task-oriented situations. In addition, we examined the psychometric properties of the MOHOST-C. Methods The MOHOST-C was developed through translation of the MOHOST. It was back-translated twice and checked by a committee of experts to ensure that the meaning of the original was maintained. The translated version was applied to 101 clients with psychiatric disorders aged between 18 to 65 years old. Clients were from 4 different groups: 28 from task-oriented occupational therapy groups, 41 from vocational training programs, 20 from sheltered employment, and 12 from supported employment. Researcher used the MMSE to screen the primary cognitive abilities of clients, and those with MMSE scores above 24 were included. Four occupational therapists were involved in this study. The forty clients included were rated by at least two raters simultaneously. Clients were observed for 20-30 minutes when participating in work or a task-oriented group and interviewed according to the MOHOST-C interview guidelines. In addition to the interview, the clients were administered related questionnaires, including the National Taiwan University Hospital Symptom Checklist (NTUHSC) and the Chinese version of the Occupational Self Assessment (OSA-C). The researcher combined the results of observations and filled out the MOHOST-C, the Chinese version of the Volitional Questionnaire (VQ-C), and the Chinese version of the Assessment of Communication and Interaction (ACIS-C). Data analysis We used the SPSS (version 11.5) and Facets (version 3.62.0) to do data analysis. Scores on the MOHOST-C were transformed to interval data by Rasch measurement model. Item separation reliability was used to examine internal consistency; Rasch analysis was used to examine inter-rater reliability, construct validity and differential item function (DIF) on gender. Additionally, Pearson’s correlation coefficient was used to examine the concurrent validity between the VQ-C, ACIS-C, OSA-C, and the MOHOST-C. Spearman’s correlation coefficient was used to examine the concurrent validity between the NTUHSC and the MOHOST-C. In addition, one-way ANOVA was used to examine the score differences for the four work types. Results The MOHOST-C was found to have excellent internal consistency. The items on the MOHOST-C could be separated into the Self aspect and the Environment aspect, and both aspects represented unidimensionality. The MOHOST-C and the VQ-C showed positive correlation (r=0.568, p<0.000), as well as the ACIS-C (r=0.767, p<0.000). However, the correlations of the MOHOST-C with the OSA-C (r=0.169, p=0.091 in self aspect; r=0.116, p=0.249 in environment aspect) and the NTUHSC (r=0.010, p=0.922) were not significant. Most raters fit the Rasch measurement model, and the results indicated that raters could learn the MOHOST-C by simply reading the manual. Two items (posture & mobility; strength & effort) belonging to the motor skills aspect in the MOHOST-C were detected with DIF in the Rasch measurement model. These items were easier for the males than for females. Furthermore, there were significant differences of the MOHOST-C scores between different types of work.