魏氏成人智力測驗第三版四因素短式測驗組合應用於台灣封閉創傷性腦傷與顳葉癲癇病人之效度研究

Abstract

研究背景：台灣版魏氏成人智力量表第三版為臨床上常用的評估工具，但臨床施測時間較長，實作時常受限於不同狀況（如臨床時間限制、患者生理狀況等），故需發展短式測驗組合常模協助臨床工作者以最短時間得到個體整體認知表現。目前已有適用於台灣正常成人魏氏成人智力量表第三版短式測驗組合之常模，但尚未發展適用於台灣臨床患者的短式測驗組合。目的：驗證台灣版魏氏成人智力量表第三版四因素短式測驗組合應用於兩種台灣臨床常見診斷--封閉型創傷性腦傷與顳葉癲癇病人的適用性。方法：本研究採回溯性分析，以台大醫院神經科2005至2011年常見疾病診斷中，患者族群較為年輕的封閉型創傷性腦傷與顳葉癲癇，並完成台灣版魏氏成人智力量表第三版的兩組患者為研究對象（分別為81人與51人），以各種心理計量特質、時間限制，以及比例與線性方程式的質性估計程序比較四因素短式測驗組型應用於此兩種患者的適用性。首先檢驗國內目前已有適用於台灣正常人樣本的五種短式測驗組合，進一步檢驗其他三十一組短式組合。結果：本研究顯示兩種估計方式應用於兩組患者的測驗組合分析中，有一半以上可產生百分之七十以上的估計準確度，亦均會產生明顯膨脹的變異數。比較發現兩種估計方式無明顯優劣，因此同時將兩種估計方式納入分析中。在檢驗適用於台灣正常成人樣本的五種短式組合中，並未發現適用於此兩種患者的組合，進一步分析其他短式測驗組合。創傷性腦傷患者的「常識-圖形設計-算數-符號尋找」組型，以及顳葉癲癇患者的「詞彙-矩陣推理-算數-數符替代」和「類同-圖形設計-記憶廣度-符號尋找」組型，整體而言是相對較佳的估計組型。結論：本研究提供相對的心理計量特質，協助我們面對此兩種患者時，可選擇不同的魏氏成人智力量表第三版的短式測驗組合。此外，即使是良好心理計量特質的短式測驗組合，仍有明顯的錯誤分類率，且受限於回溯性分析，患者資料不齊全，使用者需注意以短式測驗組合，來估計患者全量表智商，僅作為篩選之用途。

Background: The Wechsler Adult Intelligence Scale-III (WAIS-III) is one of the most commonly used measures, but its administration time for the full battery is long due to various practical constrains (such as limited time or individual’s fatigue and inattention). Therefore, a short form of the WAIS-III could be used to shorten the administration time when only a brief intelligence screening is required. Currently, there are five available tetrads suitable for normal individuals in Taiwan. However, the issue of whether these tetrads can be fully generalized to patient populations remains unclear. Objective: To assess the appropriateness of factor-based short forms of the WAIS-III for Taiwanese outpatients with traumatic brain injury and temporal lobe epilepsy. Methods: The present study was retrospective in nature. The sample consisted of 132 individuals from the outpatient clinic at the National Taiwan University Hospital from 2005 to 2011. Two patient groups, 81 individuals with a history of TBI and 51 with TLE, were recruited. Various psychometric procedures, time constrains, and qualities of estimation among prorating and linear equating procedures, were compared among all possible factor-based short form combinations for patients with traumatic brain injury and temporal lobe epilepsy. We examined the applicability of five available tetrads for normal Taiwanese to both patient groups, and then further investigated psychometric qualities of the remaining 31 tetrads. Results: More than half of the estimates based on both estimation procedures for patient groups showed inflated variances, although they had available estimation accuracy. Results of both estimation procedures were about the same; thus, we considered both estimated procedures simultaneously in our study. Five tetrads appropriate for normal individuals were not feasible for use with our patient groups. With regard to application to clinical patient groups, the Information-Block Design-Arithmetic-Symbol Search form was recommended for the traumatic brain injury patient group, and the Vocabulary-Matrix Reasoning-Arithmetic-Digit Symbol Substitution form and the Similarities-Block Design-Digit Span- Symbol Search form for the temporal lobe epilepsy patient group, due to their good psychometric properties. Conclusion: Results of this study helps clinical practitioners choose feasible factor-based tetrads for both patient groups by providing their relative psychometric quality. Users are cautioned to use these short forms only for screening purpose because even preferred tetrads had substantial misclassification, and medical information of our patients was not completely collected due to retrospective study.