臺灣版兩歲期自閉症篩檢工具簡版篩檢自閉症類群障礙症幼兒的正確率

Accuracy of the Taiwanese Version of the Screening Tool for Autism in Two-Year-Olds: Brief Form for Detecting Autism Spectrum Disorder in Toddlers

自閉症類群障礙症是先天、長期的神經發展障礙症。自閉症類群障礙症幼兒早期療育的效果彰顯早期篩檢的重要性，然而，大部患者都是在三歲以後才被診斷出來。本研究目的是發展臺灣版兩歲期自閉症篩檢工具簡版（TaiwaneseVersion of the Screening Tool for Autism in Two-Year-Olds: Brief Form, T-STAT:BF），透過兩個子研究，發展與驗證T-STAT: BF篩檢三歲以前自閉症類群障礙症幼兒的正確率。研究一招募生理年齡介於18-36個月的幼兒，參與者包括40名自閉症類群障礙症幼兒（平均生理年齡26.95個月，平均整體心理年齡17.2個月）與40名發展遲緩幼兒（平均生理年齡，25.85個月，平均整體心理年齡18.85個月）。使用T-STAT當作評估與篩檢工具，找出具鑑別力的題目作為簡版。依據不同標準可以取得五題與七題兩種簡版，分別稱為T-STAT: BF5與T-STAT: BF7。T-STAT: BF5以3分作為切截分數，敏感度與特異度分別是.93與.90；T-STAT: BF7以4分作為切截分數，敏感度與特異度分別是.93與.90。T-STAT: BF5與T-STAT: BF7篩檢正確率皆為優良。研究二參與者為生理年齡36個月以下的幼兒，包括126名自閉症類群障礙症幼兒（平均生理年齡26.71個月，平均整體心理年齡16.79個月）與118名發展遲緩幼兒（平均生理年齡24.69個月，平均整體心理年齡18.74個月）。依T-STAT: BF5進行分類，敏感度與特異度分別是.91與.84；依T-STAT: BF7進行分類，敏感度與特異度分別是.91與.86。T-STAT: BF5與T-STAT: BF7於大樣本之篩檢正確率皆良好以上，研究結果支持T-STAT: BF是正確率相當理想的自閉症類群障礙症幼兒早期篩檢工具，後續研究需以縱貫設計形式驗證T-STAT: BF的篩檢正確率的穩定度。

Rationale and Purpose: Autism spectrum disorder (ASD) is a lifelong neurodevelopmental disorder that typically manifests during early childhood. Individuals with ASD often exhibit deficits characterized by impaired social interaction and communication, as well as restricted and repetitive patterns of behaviors, interests, or activities. Multiple studies have indicated that early interventions can improve the outcomes and prognosis for toddlers with ASD and that such interventions can have long-lasting effects. These benefits underscore the importance of the early detection and diagnosis of ASD. However, despite such benefits, most children with ASD typically receive their diagnosis only after reaching the age of 3 years. In Taiwan, a scarcity of appropriate screening tools hampers the early detection and diagnosis of ASD in young children. This limitation presumably prevents young children with ASD from receiving appropriate care designed to address their core deficits and impairments. The Taiwanese version of the Screening Tool for Autism in Two-Year-Olds (T-STAT) is an interactive, autism-specific screening tool designed to differentiate between toddlers with ASD and those with developmental delays (DD) before the age of 36 months. The T-STAT is used to evaluate four key early social-communicative domains: play (two items), requesting (two items), joint attention (four items), and imitation (four items). According to previous studies, the T-STAT can identify ASD in young children with high levels of sensitivity and specificity. Clinicians can use the T-STAT to evaluate suspected cases of ASD and thus facilitate timely access to necessary treatments for children with ASD and their family. However, despite the benefits of the T-STAT, many clinicians in Taiwan usually face heavy workloads because of a large patient base. Therefore, a more efficient ASD screening tool is required to assist in the early identification and diagnosis of ASD in toddlers in Taiwan. The purpose of this study was to develop and evaluate the effectiveness of the Taiwanese Version of the Screening Tool for Autism in Two-Year-Olds: Brief Form (T-STAT: BF). Methods: Two studies were conducted to evaluate the utility and accuracy of the T-STAT: BF in detecting ASD in toddlers before the age of 36 months. Study 1 focused on establishing an algorithm for the T-STAT: BF, and Study 2 focused on validating the accuracy of this tool. This study was approved by the Ethical Committee for Human Research of Chiayi Christian Hospital. Informed consent to participate was obtained from the parents of young children aged between 18 and 36 months with suspected ASD or DD. Children with genetic disorders or severe sensory or motor impairments were excluded from this study. The diagnostic process involved clinical assessments, observations of children’s free play, parent interviews, and Autism Diagnostic Observation Schedule (ADOS) assessments. The final diagnosis was established in accordance with the criteria outlined in the fifth edition of the Diagnostic and Statistical Manual of Mental Disorders by a senior psychiatrist or a clinical psychologist with a PhD. All measurements included the Mullen Scales of Early Learning (MSEL), the T-STAT, and the ADOS. Cognitive ability (mental age) was evaluated using the MSEL, and early autism symptoms were evaluated using the T-STAT and ADOS. The T-STAT, consisting of 12 items across four domains, was used to evaluate and screen all the participating children. The T-STAT: BF was developed based on the discriminative power of each item. T-STAT and ADOS assessments were conducted by separate examiners blinded to the results of all other research measures. In Study 1, 40 toddlers with ASD (mean chronological age = 26.95 months, mean overall mental age = 17.20 months) and 40 toddlers with DD (mean chronological age = 25.85 months, mean overall mental age = 18.85 months) were assessed. The ASD and DD groups each contained 32 boys and eight girls. In Study 2, participants included 126 young children with ASD (108 boys and 18 girls) and 118 young children with DD (83 boys and 35 girls). The average chronological age of the ASD group was 26.71 months, and this group had a mental age of 16.79 months. The average chronological age of the DD group was 24.69 months, and this group had a mental age of 18.74 months. Results/Findings: In Study 1, significant differences were observed in 7 of the 12 T-STAT items between the ASD and DD groups (ps < .004). Five items (ps < .001) exhibited even more significant differences when stricter significance levels corrected for multiple comparisons were used. Two short forms of the T-STAT, namely T-STAT: BF5 and T-STAT: BF7, were established. Receiver operating characteristic (ROC) curve analysis revealed that the T-STAT: BF5 had a cutoff score of 3, with high sensitivity and specificity scores of 0.93 and 0.90, respectively. Similarly, ROC curve analysis revealed that the T-STAT: BF7 had a cutoff score of 4, with the same sensitivity and specificity scores as those of the T-STAT: BF5. These results confirmed the excellent sensitivity, specificity, and accuracy of both the T-STAT: BF5 and T-STAT: BF7 in differentiating between toddlers with ASD and those with DD before the age of 36 months. In Study 2, the accuracy of the T-STAT: BF5 and T-STAT: BF7 was evaluated with a larger sample size. An analysis of covariance, with mental age controlled for, revealed a significant difference between the ASD and DD groups regarding the total scores of the two versions of the T-STAT: BF. At a cutoff score of 3, the T-STAT: BF5 exhibited sensitivity and specificity of 0.91 and 0.84, respectively. For the T-STAT: BF7, Study 1 suggested a cutoff score of 4, yielding sensitivity and specificity of 0.91 and 0.86, respectively. Both the T-STAT: BF5 and T-STAT: BF7 demonstrated excellent sensitivity and high specificity in differentiating between toddlers with ASD and those with DD. Conclusions/Implications: Overall, the results of this study confirm the effectiveness of the two short forms of the T-STAT, namely the T-STAT: BF5 and T-STAT: BF7. These versions exhibit excellent accuracy in differentiating between young children with ASD and those with DD. Therefore, these findings highlight the T-STAT: BF5 and T-STAT: BF7 as promising screening tools for differentiating between toddlers with ASD and those with DD before the age of 36 months. As level 2 autism-specific screening tools, the T-STAT: BF5 and T-STAT: BF7 are effective for the early screening of toddlers with ASD. Furthermore, regarding practicality, these two versions require shorter administration times compared with the original T-STAT. Overall, this study provides clinicians with an alternative approach to detect ASD in toddlers, particularly those belonging to high-risk populations. In addition, the results support the utility of the T-STAT: BF in Taiwan for the early detection of ASD in toddlers. Further longitudinal studies are required to evaluate the short- and long-term predictive accuracy of the T-STAT: BF for the early screening and diagnosis of ASD.