孤独症谱系商儿童版-中文简化版的信效度研究

Reliability and Validity of the Chinese Version of the Short Abridged Autism Spectrum Quotient-Children’s Version

背景目前国内适用于孤独症儿童初级筛查的量表少。研究显示,孤独症谱系商儿童版-简化版适用于孤独症儿童的初级筛查。目的检验孤独症谱系商(AQ)儿童版-中文简化版的信效度,为我国孤独症儿童的初级筛查提供新的有效工具。方法选取2015年6-9月、2017年9-12月于北京大学精神卫生研究所门诊连续就诊的孤独症儿童101名为孤独症组(低功能孤独症者58名,高功能孤独症者43名),同期按照年龄与孤独症组相匹配的原则,选取天津市一所普通幼儿园及一所普通小学的儿童124名为正常对照组。采用AQ儿童版-中文简化版评估受试者孤独症症状。孤独症组、正常对照组均由最了解儿童情况的1名家长(多为儿童母亲)填写AQ儿童版-中文简化版,同时分别从孤独症组、正常对照组中各选取10名儿童,由了解其情况的另1名家长单独填写AQ儿童版-中文简化版。2周后,再次分别从孤独症组、正常对照组中各选取10名儿童,由首次填写AQ儿童版-中文简化版的家长(同1名家长)再次填写AQ儿童版-中文简化版。使用区分效度和效标效度进行效度分析;使用评分者信度、重测信度、内部一致性信度进行信度分析。结果低功能孤独症者AQ儿童版-中文简化版总分与高功能孤独症者比较,差异无统计学意义(P>0.05)。孤独症组AQ儿童版-中文简化版总分高于正常对照组(P<0.05)。以儿童精神科主任医师按照美国精神障碍诊断与统计手册第4版(DSM-Ⅳ)做出的诊断为金标准,绘制ROC曲线,ROC曲线下面积为0.97(P<0.01);AQ儿童版-中文简化版总分临界值为15.0分时,灵敏度为0.94,特异度为0.89。评分者信度分析结果显示,AQ儿童版-中文简化版总分的组内相关系数(ICC)为0.90(P<0.05),各条目的 ICC为0.60~0.86(P<0.05)。重测信度分析结果显示,AQ儿童版-中文简化版总分的ICC为0.97(P<0.05),各条目的 ICC为0.82~0.92(P<0.05)。AQ儿童版-中文简化版的Cronbach’sα系数为0.85(P<0.01)。结论 AQ儿童版-中文简化版的信效度较好,灵敏度、特异度较高,可用于孤独症儿童的筛查。

Background There are few scales for the screening of autism in children in primary care in China.Studies show that the Short Abridged Autism Spectrum Quotient-Children’s Version(AQ-Child-Short)is suitable for screening autism in children in primary care.Objective To test the validity and reliability of the Chinese version of AQ-Child-Short,so as to provide a new effective tool for screening autism in children in primary care in China.Methods 101 children with autism(58 of high-functioning,and other 43 of low-functioning)were enrolled during June to September,2015 and September to December,2017 from those consecutive outpatients of Peking University Institute of Mental Health.They were compared to agematched 124 normal children selected from an ordinary kindergarten and an ordinary primary school in Tianjin.The Chinese version of AQ-Child-Short was used to screen autism symptoms in all the children,which was administered by a parent who knows the child best(the child’s mother in general).At the same time,10 children were selected from each group,and their another caregiver who knows the child well also filled out the Chinese version of AQ-Child-Short.Two weeks later,10 children were chosen from each group,and received the survey again,the scale was administered by the same caregiver participating in the first round of survey.Discriminant validity and criterion-related validity were used for validity analysis.Inter-rater reliability,retest reliability and internal consistency reliability were used for reliability analysis.Results There was no significant difference in the mean total score of Chinese version of AQ-Child-Short between high and low functioning autism children(P>0.05).But the mean total score of Chinese version of AQ-Child-Short of the autism children was higher than that of the normal controls(P<0.05).The diagnostic accuracy of this scale was compared with the gold standard,namely,the diagnosis made by the chief physician of Pediatric Psychiatry Department,Peking University Institute of Mental Health following the DSM-Ⅳ.ROC analysis of the scale showed that,in diagnosing autism,the AUC of the scale was 0.97(P<0.01),with corresponding sensitivity of 0.94,and specificity of 0.89,when the critical value of the scale’s total score was determined as 15.The inter-rater reliability test results show that the intraclass correlation coefficient(ICC)of the total scale score was 0.90(P<0.05)and of the item scores was 0.60-0.86(P<0.05).The test-retest reliability results show that the ICC of the total scale score was 0.97(P<0.05)and of the item scores was 0.82-0.92(P<0.05).The Cronbach’s α of the scale was 0.85(P<0.01).Conclusion The Chinese version of the AQ-Child-Short has good validity and reliability,and high sensitivity and specificity.It could be used as an instrument for screening autism in children.