弦外之音測驗臺灣版於思覺失調症患者之心理計量特性驗證

Abstract

研究背景：心智推理 (theory of mind) 是個體推論自己或他人心智狀態的能力。研究顯示，心智推理障礙為思覺失調症患者的特徵之一，影響患者的社會功能與社區生活的獨立甚劇。因此，有效評估患者心智推理能力對治療計畫擬定十分重要。Strange Stories Test是國外研究者常用來評估思覺失調症患者進階心智推理能力的評估工具，目前臺灣已發展中文版又名弦外之音測驗，其簡易、施測快速、適用臺灣文化，且在正常人上具良好再測信度等優點，使其相較其他評估工具更具研究、臨床上的使用潛力。然而，目前臺灣版的弦外之音測驗於思覺失調症患者的心理計量特性尚未被驗證。

研究目的：本研究欲驗證弦外之音測驗於思覺失調症患者的信度與效度，以瞭解此測驗用於思覺失調症患者心智推理評估的適用性。 研究方法：本研究募集共34位台大醫院日間病房的慢性思覺失調症患者參與研究，受試者共接受研究者2次施測，2次時間間隔為3周。每次的施測項目包括弦外之音測驗與社會認知篩檢問卷 (SCSQ)。此外於前測時，也會以蒙特利爾認知測驗 (MoCA) 評估受試者的認知功能，且由醫師以CGI-S評估其症狀嚴重度。受試者於第一次弦外之音測驗的回答在經其同意後，由研究者錄音以供施測者間信度的驗證。再測前，受試者先經CGI-S評分確認其症狀穩定無變化才進行再測。研究者於收案期間由病歷蒐集受試者的人口學資料，並向日間病房精神科醫師收集受試者最近一次評估的個人與社會功能量表 (PSP) 與活性與負症狀量表 (PANSS) 分數。本研究以弦外之音測驗的前後測結果進行資料分析，分析指標包括再測信度、隨機評估誤差、最小可偵測變化值、練習效應、內部一致性、施測者間信度、天花板與地板效應；再分別以PSP、SCSQ作外在效標，驗證弦外之音測驗之生態效度與收斂效度。本研究另比較患者與過去研究中正常人族群的弦外之音測驗分數差異，以檢驗測驗的區辨效度。 研究結果：結果顯示測驗具有可接受的施測者間信度、內部一致性、收斂效度。另外，測驗的再測一致性及生態效度不佳、隨機評估誤差與最小可偵測變化值較大，且具低至中度練習效應及明顯的天花板效應；區辨效度的結果顯示測驗對患者與正常人的區辨效度不好。 研究結論：弦外之音測驗的驗證結果顯示此測驗使用在慢性思覺失調症患者上的心理計量特性不佳，使得測驗的實用價值受限。未來，可能須改良此測驗或發展更具信效度之心智推理測驗以適用於思覺失調症患者。

Background The term theory of mind (ToM) refers to the ability to infer one’s own and other people’s mental states. A great deal of evidence shows that impaired ToM is a trait of schizophrenia which attributes to poor social function of patients. To address this problem, it is important for clinicians and researchers to use a good ToM test for schizophrenia patients to assess their abilities in ToM. Among several ToM tests, the Taiwanese version of the Strange Stories Test is one with many advantages because it is brief, easy to administer, and reliable in a normal group. However, the psychometric properties of this test in schizophrenia patients remain unknown. Objective The purpose of this study is to validate the reliability and validity of the Taiwanese version of the Strange Stories Test in patients with schizophrenia. Methods Two measures including the Taiwanese version of the Strange Stories Test and the Social Cognition Screening Questionnaire (SCSQ) were tested twice on 34 schizophrenia patients from National Taiwan University Hospital by the researcher, with a 3-week interval between assessments. The Montreal Cognitive Assessment (MoCA) was also administered on the participants in the initial evaluation to assess the cognitive function of the participants. After giving their consent, the participants’ verbal responses of the Strange Stories Test in the initial evaluation were recorded to examine inter-rater reliability. The researcher used scores of the Clinical Global Impression scale (CGI) to confirm that the clinical states of the participants remained stable before retest. In addition, participants’ scores of Personal and Social Performance Scale (PSP) and Positive and Negative Syndrome Scale (PANSS) were collected from two psychiatrists, and the demographic information of participants was collected from medical charts. The test-retest reliability, random measurement error, minimal detectable change (MDC), practice effect, internal consistency, inter-rater reliability and ceiling and floor effect were examined by analyzing the test and retest scores of the Strange Stories Test. The ecological and convergent validity were examined by analyzing the correlation between the scores of the Strange Stories Test and PSP and the scores of the Strange Stories Test and SCSQ respectively. The discriminative validity was analyzed by comparing the difference between the scores of the patients and the scores of the normal group in a previous study.

Results This measure had generally acceptable inter-rater agreement, internal consistency and convergent validity. The test-retest agreement was poor with non- negligible random measurement error and the MDC% of this measure exceeded the acceptable standard. This measure also showed a notable ceiling effect and small to moderate practice effects. Finally, the ecological validity and the discriminative validity were unsatisfactory. Conclusions Our results revealed that the Taiwanese version of the Strange Stories Test had insufficient reliability and validity. Therefore, we might consider a further revision of this test. It is also a useful way to develop a new ToM measure with both clinical usefulness and scientific soundness.