中文版嬰幼兒感覺史量表之信效度研究

Abstract

研究目的 本研究目的為編譯「嬰幼兒感覺史量表」(Dunn, 2002)為中文版本，並建立中文版本之初步心理計量特性，以發展一套簡便、實用且具信度的篩檢0-3 歲嬰幼兒感覺處理障礙的量表。 研究方法 首先將嬰幼兒感覺史量表之 0-6 個月及7-36 個月分量表翻譯為中文並作跨文化的編譯及修訂。研究對象分成三組，有兩組為典型發展嬰幼兒，共有367 位0-6 個月及677位7-36 個月嬰幼兒家長填答中文版嬰幼兒感覺史量表 (0-6 個月組：平均年齡3.4 個月，標準差為1.8 個月。7-36 個月組：平均年齡22.3 個月，標準差為8.2 個月);另一組為43名發展障礙嬰幼兒（平均年齡28.4 個月，標準差為6.7 個)。內在一致性檢驗為使用所有的典型發展嬰幼兒樣本，間隔二至三星期之再測信度為取樣30 位7-36 個月發展障礙嬰幼兒（平均年齡30.2 個月，標準差5.8 個月）。建構效度是使用驗證性因素分析來 檢驗原版嬰幼兒感覺史量表的四因素模式是否適用於台灣孩童;區辨效度則以邏輯迴歸分析比較發展障礙與典型發展嬰幼兒在各分測驗及象限的分數。 結果 0-6 個月的中文版嬰幼兒感覺史量表具備差至可接受的內在一致性(Cronbach’s α= 0.42 -0.75)；7-36 個月的則具備可接受至好的內在一致性(Cronbach’s α= 0.53-0.87)。7-36個月中文版嬰幼兒感覺史量表總分具備良好的再測信度(ICC = 0.94)。除了視覺處理的分測驗分數的再測信度較低外(ICC = 0.56)，其他分測驗分數及因素分數亦有可接受至好的再測信度(ICC = 0.69-0.91)。驗證性因素分析結果皆能支持原本的四因素模式：0-6 個月量表( RMSEA = 0.05, SRMR = 0.08, CFI = 0.93 and NNFI = 0.92)及7-36 個月量表 ( RMSEA = 0.04, SRMR = 0.06, CFI = 0.95 and NNFI = 0.94)。邏輯迴歸分析結果亦顯示中文版嬰幼兒感覺史量表的分數可用來區辨正常發展幼兒與發展障礙之幼兒。 結論 本研究結果顯示，中文版嬰幼兒感覺史量表具有可接受的信效度基礎，可適用於評量出生至三歲台灣嬰幼兒的感覺處理功能。然而，由於發展障礙族群的樣本年齡皆限制於7-36 個月，此為在台灣較常轉介至臨床接受治療的個案年齡層，故未來尚需探討0-6 個月版之再測信度及其區辨效度之驗證，以確認量表於臨床及研究上的實用性。

Objectives: The aim of the study was to develop a Chinese version of the Infant Toddler Sensory Profile (ITSP-C) and investigate its psychometric properties when used in Taiwanese children. Methods: First, the ITSP was translated and cross-culturally adapted into Chinese, including subscales for children ages 0-6 months and 7-36 months. The participants consisted of three cohorts: 367 typically developing infants ages 0 to 6 months (mean age = 3.44 months, SD = 1.82), 677 toddlers ages 7 to 36 months (mean age = 22.33 months, SD = 8.15), and 43 children with developmental disabilities, including developmental delays and Autism (mean age = 28.4 months, SD = 6.7). Parents of all the children completed the ITSP-C. Internal consistency was determined using all the typically developing infants and toddlers. Test-retest reliability was examined using 30 children with developmental disabilities (mean age = 30.23 months, SD = 5.78) in a 2-3 week interval. Construct validity was examined using confirmatory factor analysis (CFA) to assess if the four-factor model of the original ITSP was applicable to Taiwanese data. To verify discriminative validity, the ITSP-C scores of the children with developmental disabilities were compared with those of the normally developing children using logistic regression analysis. Results: Internal consistency was poor to acceptable for children ages birth to 6 months (Cronbach’s α = 0.42-0.75), and acceptable to good for children ages 7 to 36 months (Cronbach’s α = 0.53-0.87). For the 7 to 36 month subscale, test-retest reliability was good for the total scores (ICC = 0.94), and acceptable to good for each section and each quadrant scores (ICC = 0.69-0.91), except the visual processing section scores (ICC = 0.56). The results of the CFA provided support for a four-factor model both for the 0-6 months subscale (RMSEA = 0.05, SRMR = 0.08, CFI = 0.93 and NNFI = 0.92) and the 7-36 months subscale (RMSEA = 0.04, SRMR = 0.06, CFI = 0.95 and NNFI = 0.94). Logistic regression analysis demonstrated that the ITSP-C were able to discriminate between children with developmental disabilities and normally developing children. Conclusions: The results of the present study supported the preliminary psychometric properties of the ITSP-C. That is, the ITSP-C has acceptable reliability and validity when used in Taiwanese children. However, only the test-retest reliability and discriminant validity of the 7-36 month scale assessed for children in this age range were most often referred to clinics in Taiwan. Thus, additional studies are needed to examine the test-retest reliability and discriminant validity of the 0-6 month subscale in order to establish its utility in both clinical and research settings.