發展職能治療專業人員之短版溝通技巧評估工具

Abstract

背景：溝通技巧為職能治療師必要的專業技能之一，有助於提升治療關係、決策效能及治療成效。完整、深入且精確地評估實習學生之溝通技巧，可協助臨床指導老師掌握實習學生之溝通能力及問題。「職能治療溝通技巧評估工具 (communication-skills assessment for occupational therapists, COT)」為目前極具潛力之工具。其完整包含所有溝通技巧向度且題數充足，可使臨床指導老師及研究人員充分掌握實習學生之能力水準及特定問題。然而，COT之題數過多，對評估者之負擔大，且其心理計量特性驗證闕如，嚴重影響可行性及限制結果之解讀。故COT亟待簡化。 目的：本研究目的有二：（一）發展適用於職能治療師之短版溝通技巧評估工具 (A short form of a communication-skills assessment for occupational therapists, SCOT)；（二）初步驗證SCOT之心理計量特性（含羅序信度、評估者間信度、收斂效度與反應性）。 方法：本研究分為二個階段：（一）以3步驟發展SCOT：(1) 大量施測COT於職能治療實習學生；(2) COT驗證：以多向度羅序分析COT題目，刪除不符羅序模型假設的題目；(3) 挑選題目以建構SCOT。（二）驗證SCOT之心理計量特性：使用參與前一階段之受試者資料，以模擬分析SCOT之心理計量特性。信度部份，研究者使用實習學生之能力估計SE計算SCOT之個別與平均信度，並使用不同位評估者之再次評估資料，以組內相關係數 (intraclass correlation coefficient, ICC) 驗證評估者間信度；收斂效度部份，驗證SCOT各向度分數與Gap-Kalamazoo Communication Skills Assessment Forms (GKCSAF) 相應效標題目之相關程度，另驗證SCOT與GKCSAF總分之相關程度；反應性部份，以配對樣本t檢定、效應值 (Cohen's d)、標準化反應平均值 (standardized response mean, SRM) 檢驗之，並與原版COT及GKCSAF比較。 結果：第一階段中，共270份實習學生與病人互動過程納入分析（包含150位實習學生及252位病人）。COT原始題目共78題，於刪除32題適配度不佳之題目後，剩餘46題適配度良好 (infit and outfit mean square = 0.52–1.33)。研究者自COT挑選難度與受試者平均能力相應之題目組成SCOT，題目共37題。第二階段心理計量特性驗證發現：SCOT八向度題目，除「接收」及「主導會談」外，其餘六向度具有可接受之團體層級信度（平均信度 = 0.70–0.82）以及個別信度（約60%–90%受試者的個別信度 ≥ 0.70）。二位評估者之ICC = 0.44–0.86，顯示中等至可接受之評估者間信度。SCOT各向度分數與GKCSAF相應效標題目之Spearman's ρs = 0.07–0.67，與GKCSAF總分之Pearson's rs = -0.04–0.64，此結果顯示SCOT具有不佳至良好之收斂效度。受試者於八向度之第一次與第二次分數差異皆達顯著 (p ≤ 0.009)，Cohen's ds = 0.26–0.67，SRMs = 0.26–0.63，顯示SCOT具有良好之反應性。 結論：初步結果顯示：SCOT可提供完整（八種治療師所需具備之溝通技巧）、有效（符合多向度模型），可呈現受試者溝通技巧改變之評估。SCOT具備潛力可應用於臨床教學與研究，以提升溝通技巧評估與效能。惟部份SCOT向度之評估者間信度及收斂效度不佳，待後續改良或再驗證。

Background: Communication skills are one of the core competencies for clinical practice, which affects the patient-therapist relationship, efficacy, and efficiency of occupational therapy. Thus, comprehensive, accurate, and precise assessments of communication skills are necessary for clinical supervisors to determine the communication difficulties and strengths of occupational therapy interns. The Communication-skills Assessment for Occupational Therapists (COT) has great potential for assessing interns' communication skills, because it assesses comprehensive domains and contains sufficient items. The COT can help supervisors capture the comprehensive scope and specific problems of interns' communication skills. However, the large number of items in the COT imposes time and administration burdens on raters, which severely limits its utility. Therefore, to simultaneously achieve a comprehensive, precise, and efficient assessment of communication skills, a short form measure developed by the multidimensional Rasch analysis appears to be a promising solution. Purposes: The purpose of this study was: (1) to develop a short form of a communication-skills assessment for occupational therapists (SCOT); (2) to preliminarily examine the Rasch person reliability, inter-rater reliability, convergent validity and responsiveness of the SCOT. Methods: This study contained two stages. First, the development of the SCOT was divided into three steps: (1) administering the COT on occupational therapist interns; (2) validating the COT: removing the items that did not fit the Rasch model; (3) selecting items from the COT to establish the SCOT. Then, simulations were performed to examine the psychometric properties of the SCOT. Regarding the Rasch person reliability, standard errors of interns' ability estimates were used to calculate the individual reliabilities and average reliability of the SCOT. The intraclass correlation coefficient (ICC) was used to examine the inter-rater reliability of the SCOT. We investigated Spearman's ρ between the SCOT and the corresponding items of the Gap-Kalamazoo Communication Skills Assessment Forms (GKCSAF), and Pearson's r between the SCOT and the total score of the GKCSAF to examine the convergent validity of the SCOT. The responsiveness was examined using the paired t test, Cohen's d, and standardized response means (SRM). The responsiveness of the SCOT was compared with the COT and GKCSAF using the bootstrap approach. Results: In stage 1, a total of 270 interactions between interns and patients were analyzed, involving 150 interns and 252 patients. The COT originally contained 78 items. After removing 32 misfit items, the remaining 46 items fit the Rasch model's assumptions (infit and outfit mean square = 0.52–1.33). From these, 37 items were chosen to constitute the SCOT based on matching the average ability of the interns. In stage 2, the results of the Rasch person reliability indicate that the SCOT has acceptable average reliability (0.70–0.82) and individual reliability (about 60%–90% of subjects had an individual reliability ≥ 0.70), except the domains of “Receiving Information” and “Guiding the discussion.” The results of the convergent validity indicate that the SCOT has poor to acceptable convergent validity (Spearman's ρs between the SCOT and the corresponding items of the GKCSAF = 0.07–0.67, Pearson's rs between the SCOT and the total scores of the GKCSAF = -0.04–0.64). The results of the responsiveness indicate that the SCOT has good responsiveness (Cohen's ds = 0.26–0.67, SRMs = 0.26–0.63). Conclusion: Our findings suggest that the SCOT provides comprehensive, valid, and responsive detection of interns' communication skills changes. The SCOT shows great potential for application in both clinical training and research settings to improve the efficacy of communication skills assessments. However, the inter-rater reliability and convergent validity of some domains are unclear. Further modification and validations of the SCOT are warranted.