| Autism spectrum disorder（ASD）has become the focus of the international community due to its high incidence and heavy disease burden.Effective screening tools can improve the efficiency of early identification and diagnosis of ASD children,and are also an important factor affecting effective intervention of ASD children.The Diagnostic and Statistical Manual of Mental Disorders Fifth Edition（DSM-5）published in 2013 has made an important update to the diagnostic criteria for ASD,but the commonly used ASD screening tools in China still lack of update and revision.The Gilliam Autism Rating Scale Third Edition（GARS-3）is an ASD screening tool with good psychometric characteristics developed based on DSM-5.It includes six subscales of restricted/repetitive behaviors,social interaction,social communication,emotional response,cognitive style and maladaptive speech.Of these,the last two subscales only applicable to individuals who can speak.The purpose of this study is to revise the Chinese version of GARS-3 and evaluate the psychometric characteristics of GARS-3used in the Chinese cultural context,with the hope of providing a DSM-5 based ASD screening tool for China.The Chinese version of GARS-3 was developed through multiple steps such as scale translation and back translation,cultural adjustment and pre-test feedback.The tool was then used to assess 326 individuals with ASD,50 individuals with other types of disorders,and 71 individuals with typical development.The findings were as follows:1.Item analysis: The critical ratio values of 7 items in the subscale of cognitive style,failed to reach 3.0,and the correlation coefficients between the score of each item in the subscale and the total score of the scale were extremely low,showing no statistical significance.The critical ratio values of all other items were all greater than 3.0,and the correlation coefficients between each item and the total score of the scale were all greater than 0.3,showing statistical significance.The results indicated that 7 items in the subscale of cognitive style were not suitable enough and need to be deleted,and the rest items could be retained.2.Validity test:（1）construct validity.Exploratory factor analysis showed that the four-factor model could explain 60.05% of the total variation,and the four-factor items were consistent with the four subscales items corresponding to the source scale.The five-factor model could explain 64.61% of the total variation,and the five-factor items were also consistent with the five subscales items corresponding to the source scale.The correlation coefficients between each subscale ranged from 0.235 to 0.637,and the correlation coefficients between the score of each subscale and the total scores of four subscales,the total scores of five subscales ranged from 0.593 to 0.855.The scores of each subscale and the total scores of four subscales and five subscales in the ASD group were significantly higher than those in the typical developmental group and other disorders group.（2）Content validity.The items of the scale were directly related to the diagnostic criteria of ASD,and the contents have been reviewed by experts,so the content validity was guaranteed to a certain extent.（3）Criterion validity.The correlation coefficients of the total scores of four subscales,five subscales in the Chinese version of GARS-3 and the total scores of ABC scale were 0.713 and 0.746.（4）Predictive validity.For four subscales,the sensitivity and specificity were 0.89 and0.86 when the cutoff was 63 points.For five subscales,the sensitivity and specificity were 0.81 and 0.93 when the cutoff was 79 points.Reliability test:（1）Internal consistency reliability.The Cronbach’s α coefficients of the four subscales and five subscales were 0.954 and 0.913,respectively.And the split-half reliability of the four subscales and five subscales were 0.972 and 0.961,respectively.（2）Retest reliability.The retest reliability of the four subscales and five scales were 0.869 and 0.828.（3）Rater reliability.The intraclass correlation coefficients of the four subscales and five scales were 0.841 and 0.839.The conclusions are as follows:1.The final Chinese version of GARS-3 contains 51 items,including 5 subscales of restricted/repetitive behaviors,social interaction,social communication,emotional response and maladaptive speech.The first four subscales need to be evaluated for all individuals,and the maladaptive speech subscales only need to be evaluated for individuals who can speak.2.The Chinese version of GARS-3 has good reliability and validity,sensitivity and specificity.It is a reliable and effective screening tool for ASD,and can be used in the screening and auxiliary diagnosis of ASD in China. |
PDF Full Text Request