修訂版學步期自閉症檢核表（M-CHAT）篩檢4歲以下自閉症類群疾患兒童效度探究

Using the Modified Checklist for Autism in Toddlers (M-CHAT) for Detecting Children with Autism Spectrum Disorders under 4 Years Old

楊景琦(Ching-Chi Yang)

指導教授 ： 吳進欽

高雄醫學大學/人文社會科學院/心理學系碩士班/碩士(2015年)

https://doi.org/10.6832/KMU.2015.00066

摘要

研究背景與目的：自閉症類群疾患（autism spectrum disorders, ASDs）患者的發展與行為障礙在2歲前即出現，但是正式診斷卻延遲到3歲以後，不利於患者早期介入。因此，需要發展具效度的早期篩檢工具，改善診斷延遲的問題。修訂版學步期自閉症檢核表（Modified Checklist for Autism in Toddlers , M-CHAT）是廣被使用的篩檢工具，國際上不少研究探究其篩檢效度，但是台灣缺乏篩檢效度的研究。基於早期篩檢需求，本研究探究M-CHAT篩檢4歲以下ASDs兒童的效度。
方法：受試者為236名生理年齡介於18-47個月大的兒童，包含ASDs兒童113名，發展遲緩兒童123名。以23題M-CHAT為篩檢工具，家長在兒童完成診斷評估前，填寫M-CHAT。採用過去研究建議的指標23題任3題失敗、關鍵6題任2題失敗及關鍵7題任2題失敗，鑑驗指標篩檢效度，並尋找是否有適合台灣臨床樣本的篩檢指標。
結果：過去篩檢指標的效度，23題中任3題失敗，敏感度.88，特異度.53；關鍵6題任2題失敗，敏感度.46、特異度.94；關鍵7題失敗2題，敏感度.57、特異度.89。分析結果與過去研究不一致，過去篩檢指標未達良好篩檢效度的水準。然而，23題中任4題失敗敏感度.77，特異度.72，則符合良好篩檢效度的水準。以判別分析與共變數分析找到14題版，失敗3題為最佳切截分數，其敏感度.71，特異度.82，符合良好篩檢工具的水準。將受試者區分成18-30個月與31-47個月兩年齡組，各指標篩檢結果如下：18-30個月組，23題中任4題失敗，敏感度.79，特異度.74；14題版失敗3題，敏感度.74，特異度.86。31-47個月組，23題中任4題失敗，其敏感度.75，特異度.70；14題版失敗3題，敏感度.68，特異度.77。
結論：M-CHAT篩檢台灣4歲以下臨床樣本，敏感度與特異度表現屬於可接受的範圍；細分不同年齡組，M-CHAT篩檢年幼組（18-30個月）的表現較年長組（31-47個月）略佳。M-CHAT中有5題在不同關鍵題目皆出現，包括：社交興趣、分享式手指指示、展示、模仿及跟隨手指指示等早期社會互動表現，支持ASDs早期社會互動障礙是早期篩檢的重點。台灣臨床場域缺乏ASDs層次二篩檢工具，本研究結果支持將M-CHAT做臨床使用，幫助台灣ASDs早期篩檢工作進行。

關鍵字

修訂版學步期自閉症檢核表 ； 自閉症類群疾患 ； 早期篩檢 ； 敏感 ； 特異度

並列摘要

Background and purposes: Children with autism spectrum disorders (ASDs) show developmental and behavioral impairments before the age of two. However, the diagnosis won’t be made until three years old or later. This situation will bring negative impacts to early intervention. Therefore, dealing with the issue of delayed diagnosis, we need to develop an early screening tool with good validity. The Modified Checklist for Autism in Toddlers (M-CHAT) is a well-known and used screening tool; there are a lot of studies discovering its validity. However, there are no related validity studies in Taiwan. Considering needs of early diagnosis, the present study aims to discover the validity of using M-CHAT to detect children with ASDs less than 4 years old in Taiwan.
Methods: There were 236 children aged from 18 to 47 months old recruited in this study, including 113 ASDs children and 123 developmental delay (DD) children. We invited parents to fill up M-CHAT Chinese version while waiting for diagnostic assessments. Analysing the screening validity of criteria suggested by former researchers, including any 3 items of the 23 items, failing 2 items of 6 discriminate items, and failing 2 of 7 critical items. Besides, we would like to discover if there’s a new or better criteria for screening clinical sample in Taiwan.
Results: The validity of criteria suggested by former researchers, including (1) any 3 items of the 23 items, yielded a sensitivity of .88 and specificity of .53; (2) failing 2 items of 6 discriminate items yielded a sensitivity of .46 and specificity of .94; (3) failing 2 of 7 critical items yielded a sensitivity and specificity of .57 and .89, respectively. These results were incongruent with former researches; screening validity didn't reach the standard of a fine screening tool. Additionally, we found that using any 4 of the 23 items yielded a sensitivity of .77 and specificity of .72. This result reached the standard of a fine screening tool. Meanwhile, through discriminated analysis and analysis of covariance (ANCOVA) we found a short version with 14 items overlapped between these two methods, and using the cutoff of 3 point, yielded a sensitivity of .71 and specificity of .82. This result also reached the standard of fine screening tool. Furthermore, we divided our subjects into two different age groups: 18-30 months old and 31-47 months old. For the 18-30 months group, using the criterion as any 4 of the 23 items yielded the sensitivity and specificity of .79 and .74, respectively; using a short version with 3 failing items in 14 critical items yielded the sensitivity and specificity as .74 and .86, respectively. For 31-47 months group, using criterion any 4 items of the 23 items, yielded the sensitivity and specificity of .75 and .70, respectively; using the short version with failing 3 items in 14 critical items yielded the sensitivity and specificity as .68 and .77, respectively.
Conclusion: The validity of using M-CHAT to screen clinical samples less than 4 years old in Taiwan seemed acceptable. Distinguishing different age groups, we found that M-CHAT had better performance screening those within the age of 18-30 months than 31-47 months. There as were 5 items overlapping among different criteria, including item 2 (social interest), item 7 (pointing for interest), item 9 (showing), item 13 (imitation), and item 15 (following pointing), supporting early social interaction impairments are key points to ASDs early screening. The clinical field in Taiwan lacked ASDs level two screening tools, so the results of this study supported the utility of M-CHAT screening clinical samples and support ASDs early screening works in Taiwan.

並列關鍵字

M-CHAT ； ASDs ； early screening ； sensitivity ； specificity

參考文獻

吳進欽、姜忠信、虞燕婷（2010）。自閉症類幼兒社會注意力的探究。中華心理學刊，52(1)，57-74。

余麗樺、郭乃文、潘秀琴（1998）。〈盧-尼神經心理測驗組中文修訂版的信效度研究〉。《台灣精神醫學》，12，365-378。

吳進欽、朱慶琳、侯育銘、林姿伶、姜忠信（2014）。T-STAT評估18~24個月自閉症類嬰幼兒效度探究。中華心理衛生學刊，27(4), 621-644。

吳進欽、朱慶琳、侯育銘、姚淑芬（2014）。克氏行為量表篩檢4歲以下自閉症類疾患兒童效度探究。中華心理衛生學刊，27(1)，131-161。

吳進欽、姜忠信、侯育銘（2013）。兩歲前自閉症類及患嬰幼兒篩檢工具：文獻回顧。中華心理衛生學刊，26(4)，549-568。