中文版意志量表之施測者間信度探討及關聯性分析

Abstract

背景及目的 精神障礙者經常有動機低落的現象，造成活動參與度不佳，影響其復健成效，本文採用中文版意志量表來評量個案參與活動的動機狀態–即意志高低，以及環境如何影響其意志高低。過去已有中文版意志量表的內在一致性、再測信度及建構效度之證據，由於此量表是觀察個案表現後，據以評估其意志高低之測驗，不同施測者所得之結果應具有一致性，因此本研究之目的為探討受訓過的臨床職能治療師使用此量表之施測者間信度，以期能建立一套模式，作為臨床職能治療師能客觀且有效地評量個案之參與活動動機的基礎，未來可用以作為衡量治療成效的指標之一。 研究方法 參與本研究之施測者共3位，執業年分別為18、14及11年，這些施測者在評分前需先閱讀中文版意志量表手冊之內容，及練習評估一位標準個案，再互相比較與討論，並進一步補充更詳細的評量標準，作為研究個案評分之準則。本研究共收集100位精神障礙者之資料，每位個案需參與一項自己選擇以及另一項由治療師指定之操作式活動，同時至少有兩位治療師觀察其活動表現並進行評分，之後個案需完成其他相關量表之填寫。研究者於民國103年3月到103年12月間於北台灣一間醫療中心之精神科病房進行收案，本研究也通過收案醫療院所的人體試驗委員會之審查。 本研究使用多面向羅序測量模式(Multi-Faceted Rasch Measurement Model)以及傳統測驗理論之組內間相關係數(ICC)和Multiple rater kappa來探討施測者間信度，資料分析是以FACETS軟體(3.71.4版)及SPSS軟體(17.0版)進行統計分析。 結果 100位個案中，平均年齡為47.8歲(標準差=12.6歲)；有59位男性個案(59%)；37%的個案診斷為思覺失調症；平均發病年為12.5年(標準差=9.8年)。中文版意志量表之項目分離值及離散信度分別為6.42及0.98，施測者間的給分完全一致之比例為89.1%，ICC為0.84 (95% CI=0.7~0.93，F=16.91，p=0.00)，14項題目之kappa值介於0.348~0.663之間，代表此量表具有可接受之施測者間信度，但施測者間的嚴苛度有顯著差異，不同診斷及活動類型會在意志量表項目難易度上有顯著的差異。 結論 中文版意志量表具有尚佳的施測者間信度，是一個可信且有效的測量意志高低之觀察式評估工具。關於治療師嚴苛程度之差異，須持續將評分內容敘述詳盡及具體，改善治療師的一致性，同時也需建立不同診斷及活動類型的特定項目難易度指標，作為未來使用之依據。本研究使用的活動種類有限(侷限於操作式活動)，未來可進一步驗證不同形式活動評分的一致性，建立施測者訓練的標準程序，改善使用本工具評估個案的信度。

Background and Purpose Many people with mental illness often appear to low motivation which result in poor activity participation and reduce the effectiveness of rehabilitation. In this study, we use the Chinese version of Volitional Questionnaire (VQ-C) to assess the client’s motivational status, the level of volition, of the activity participation, and to understand how the environment affects their volition. The evidence of internal consistency, test-retest reliability, and construct validity of the VQ-C had been studied previously. Since the VQ-C is conducted by assessing the level of volition after observing clients’ performance, the score results from different rater should be consistent. Thus, the purpose of this study is to examine the inter-rater reliability of the VQ-C conducted by well-trained clinical occupational therapists, and with a view to establish a model which can be the foundation for clinical occupational therapists to assess the clients’ participation motivation objectively and effectively, and can be expected to be one of the indexes of the evidence-based for clinical treatment in the future. Methods There are 3 raters in this study and the professional work experiences are 18, 14, and 11 years accordingly. Before the assessment, the raters need to read the manual of the VQ-C and practiced assessing a standard client followed by comparison and discussion with each other. Through complementing the standard of assessment more detail, it can be the criterion of scoring the research clients. Data of 100 participations with mental illness were collected in this study. Each client participated in one operating type activity chosen by self and the other one assigned by therapist. Meanwhile, at least 2 therapists conduct the score according to the client’s performance, and then clients need to complete other scales. All clients were recruited from psychiatric wards of a medical center in northern Taiwan. Data collection was conducted from March 2014 to November 2014. The study was approved by the ethics committees of the Tri-Service General Hospital. We use the Multi-Faceted Rasch Measurement Model, ICC, and multiple rater kappa to probe into the inter-rater reliability, and the statistical analysis was conducted by FACETS (version 3.71.4) and SPSS (version 17.0). Results The average age of the 100 clients is 47.8 years old (SD=12.6), 59 clients are men (59%), 37% of clients are schizophrenia, and the average number of year of onset is 12.5 years (SD=9.8). The separation index of the VQ-C is 6.42 and the separation reliability is 0.98. The rater consistency is 89.1%, the ICC is 0.84 (95% CI = 0.7~0.93, F = 16.91, p = 0.00), and the kappa of 14 items are between 0.348 to 0.663. It shows the VQ-C has acceptable inter-rater reliability, but still has a significant difference in rater severity. There are differential item functioning (DIF) of the VQ-C in diagnosis and activity type. Conclusion The Chinese version Volitional Questionnaire has fair inter-rater reliability. It is a reliable and effective observational assessment tool of volitional function. The difference of the severity between therapists in rating content will need to be described in detail and concrete continually to improve the consistency of the therapist, but as a basis for future use, it also need to establish a specific item difficulty index in different diagnosis and types of activities. Activity types used in this study is limited (limited to operating type activities), and it will need further verify the scoring consistency of different forms of activity, and establish the standard procedures of rater training program to improve the reliability of the VQ-C.