您好：
您已成功加入此商城會員，後續您可使用以下帳密進行登入，如需變更資料可進入會員專區自行修改，感謝您的支持！
作家 / 早療協會 報導
孤独症儿童的早期发现：中国婴幼儿孤独症筛查量表
Early screening of children with autism: Checklist for Chinese Infants with Autism
邵智*1,2、张雅如1,2、赵琴1、郭磊2
Zhi Shao1,2, Ya-ru Zhang1,2,Qin Zhao1, Lei Guo2
1 重庆市第九人民医院重庆市儿童孤独症康复治疗中心，重庆 400700
2 西南大学儿童孤独症临床康复实训研究中心，重庆 400700
1 Chongqing Rehabilitation Center for Children with Autism,the Ninth People’s Hospital of Chongqing，Chongqing ,400700, China;
2 Rehabilitation Practice and Research Center for Children with Autism of Southwest University,Chongqing,400700, China
研究背景与目的：孤独症谱系障碍（autism spectrum disorders，ASD）的早期识别与及时干预能明显改善预后。与正常儿童比较，ASD高风险婴幼儿的早期征象在1岁左右就可以出现，表现出明显的社会行为差异。在我国，ASD的早期筛查与诊断技术仍比较滞后，尤其缺乏针对小年龄段，如12-24个月月龄的孤独症儿童。目前的临床工作中，常用的筛查量表，如M-CHAT等，仅适用于18个月龄以上的儿童，且由家长主观报告，可能与儿童实际情况出现偏差,影响客观性。因此，编制12-24个月龄婴幼儿的交互式ASD筛查量表，由专业人员直接观察儿童行为，有效筛出ASD高危倾向儿童，显得十分必要。本研究旨在编制一套适用于我国12-24个月龄儿童的ASD筛查量表，并对其进行信度、效度分析，为我国ASD的早期筛查和诊断提供有效可行的工具，使婴幼儿期ASD的早期发现、早期干预成为可能。
方法：本研究分为编制量表、筛查界限分确定、社区样本中运用三个阶段。首先编制婴幼儿ASD筛查量表；然后在混合样本中施测，确定筛查界限分和信效度；再在社区样本中实际应用；最后通过追踪随访，验证量表的应用价值。
第一阶段，编制量表。首先确立量表维度。通过开放式问卷调查等方式，确立了社交沟通、游戏技巧、感知觉与兴趣行为3个测评维度。然后根据文献资料、现有量表、临床经验等，分析了各维度的核心成分，根据核心成分编制题目。最后由发育行为儿科学、儿童发展心理学、心理测量学等领域专家对题目进行3轮次反复研讨与修改，最终形成了量表。量表共包含20道题目，选项按照题目所描述的行为特征,采用3点计分形式。
第二个阶段，确定筛查界限分。将编制好的量表在混合样本中施测。混合样本分为病例组和对照组，病例组来自国内华北（北京）、华南（深圳）、华西（西安）、华东（青岛）、西南（贵阳、重庆）地区6个城市三级综合性医院及妇幼保健院儿童保健科、儿童心理门诊就诊的儿童,共120例，月龄在13-20个月（平均月龄16.52+2.82月）之间；其中诊断为孤独症的儿童60例（平均月龄17.40+2.59月），发育迟缓的儿童（平均月龄15.67+3.28）60例。对照组由60例在上述医院做常规儿童保健的正常儿童构成，年龄在13-20个月之间（平均月龄在16.72+2.74月）。病例组和对照组均由两名经过专业训练的评估人员进行观察评定，给出分数。并于一周以后，对所有研究对象进行复测，以确定量表的重测信度。儿童年龄大于24个月以上时，对儿童进行随访。由两名副主任医师以上职称的医生依据DSM-5诊断标准进行诊断，将随访时的诊断结果作为校标，用于评价量表的校标效度。最终确定量表的界限分。
第三个阶段，将量表在社区样本中运用。社区样本选取上述6个城市中，7家三级综合性医院或妇幼保健院儿童保健科门诊接受服务的儿童4871名。
结果：第二个阶段，量表在混合样本中进行施测后，进行信效度分析，各评分者对量表及各个维度的评定分数相关系数均在0.90以上（P<0.01）。重测信度为0.96（P<0.01），内部一致性系数为0.901（P<0.01）。基于病例组灵敏度和特异度取值情况，最终确定量表中危标准和高危标准。确定为中危标准时，量表灵敏度为1，特异度为0.84，阳性预测值为0.65；确定为高危标准时，灵敏度为0.73，特异度为0.97，阳性预测值为0.94。
第三个阶段，在社区样本中应用，共筛出ASD倾向儿童135例，其中高于高危标准87例，高于中危标准48例。这135例儿童大于24个月时进行随访；另外，在未见ASD倾向的儿童中随机抽取的5%也进行随访（N=244）。随访时，87例高于高危标准的儿童，最终有82例确诊为ASD、5例确诊为发育迟缓；48例高于中危标准的儿童，有32例确诊为ASD、14例确诊为发育迟缓，其余2名为典型发展儿童；随机抽取的5%的正常儿童中，最终有1名诊断为ASD、4名诊断为发育迟缓，其余均为正常儿童（N=239）。所以在社区样本中，中危标准时，该量表灵敏度为0.97，特异度为0.94，阳性预测值为0.67；高危标准时，灵敏度为0.99，特异度为0.98，阳性预测值为0.94。
结论：该量表的结构符合量表设计的理论框架；由专业人员通过现场观察，对儿童行为进行分级评定；适用于12-24个月的婴幼儿，具有较好的信度。从社区样本中的应用情况可知，量表具有较好的效标效度，能有效筛出孤独症症倾向儿童；具有较好的推广使用价值。
关键词：中国婴幼儿孤独症筛查量表；孤独症；信度，效度
Background and objects Early diagnosis and in-time intervention can significantly improve the outcome of patients with autism spectrum disorder(ASD). As compared with typically developed children, infants and toddlers with high risk of ASD would manifest obvious social behavioral impairments as early as 1 year old. Currently, there are only few early screening and diagnostic tools available for autism spectrum disorders in China, especially for young children aged 12 to 24 months old, and among them the most frequently-used screening scales(such as M-CHAT) are only specific for children aged above 18 months old. Furthermore, the above-mentioned checklists are mostly subjectively reported by caregivers, which might impair their objectivity as there may be discrepancies between actual and reported results. Hence, it is essential to develop a reciprocal screening tool conducted by professional staffs for children with high risk of ASD aged between 12 to 24 month-old. This study aimed to establish an efficacious and feasible early screening and diagnostic checklist for children with autism spectrum disorder aged between 12 to 24 months old of China, and make early diagnosis and interventions possible for these young children.
Methods The study proceeded in three stages. The first stage involved the development of the checklist, pilot testing in a mixed sample of children, and confirmation of screening thresholds, reliability and validity. At the second stage, the checklist was applied in local community samples for further feasibility testing. The last stage involved the long-term follow-ups.
Stage I Initial checklist development Three measurement dimensions-social communication, play skills, perception and behavior of interests were determined based on an open-questionnaire survey. Core components of each dimension were extracted after analyzing available references, scales and clinical experiences, based on which testing items were designed. The checklist was finally confirmed after three times of repeated discussions and modifications by specialists in the fields of developmental behavioral pediatrics, children developmental psychology and psychometrics. 20 items were comprised in the scale, and three-point scoring method was adopted to rate the behavioral characteristics of subjects according to descriptions of each item.
Stage II Threshold confirmation Pilot testing was then conducted in a mixed sample consisting of patient and control groups. 120 patients average aged between 13-30 month-old(mean age=16.52+2.82 months) were recruited as the patient group from children health care and children psychological clinics of several Level III general hospitals, and maternity and childcare hospitals of six cites of north(Beijing), south(Shenzhen), west(Xi’an), east(Qingdao), and southwest(Guiyang, Chongqing) of China, among them 60 were with autism spectrum disorder(mean age=17.40+2.59 months) and 60 with mental retardation (mean age=15.67+3.28 months). Meanwhile 60 control samples of typically developed children were recruited from aforementioned clinics(mean age=16.72+2.74 months). Both groups were evaluated and scored by two trained staffs. Re-tests were conducted a week after to determine the test-retest reliability. Follow-ups were carried out when subjects reached the age of 24 months old or above, and definite diagnosis was made by two senior specialits. Diagnostic results were considered as criterion validity index, and intermediate and high risk cutoffs were final determined afterwards.
Stage III Application in community samples Community samples consisting of 4871 children from 7 clinics of the six above-mentioned cities received the checklist evaluation at the final stage.
Results Results of the second stage investigation indicated high interrater reliability coefficient of above 0.90(P<0.01), excellent test-retest reliability(0.96, P<0.01) and good internal consistency(0.901, P<0.01). Intermediate and high risk cutoffs were determined based on the sensitivity and specificity values of the patient group. When the cutoffs were set as the intermediate and high risk standardization. the sensitivity turned to be 1 vs. 0.73, the specificity 0.84 vs .0.97, and the positive prediction rate 0.65 vs. 0.94 respectively.
135 children with high risk of ASD were screened out during the last stage of study, and among which 87 scored above the high risk and 48 above the intermediate standards. 244 randomly selected children(5%) without risks of ASD and the 135 children mentioned above all received follow-up evaluations. At follow-up visits, 82 children were diagnosed as ASD, and 5 as mental retardation out of the 87 subjects scored above high risk standard. 32 children were diagnosed as ASD, 14 as mental retardation and 2 as typically developed children among the 48 subjects who scored above the intermediate standard. Among the randomly selected children, there were 1 diagnosed as ASD, and 4 as mental retardation. The sensitivity of the checklist was 0.97 vs.0.99, the specificity 0.94 vs.0.98, and the positive prediction rate 0.67 vs.0.94 when the cutoffs were set as intermediate and high risk standards respectively among community samples.
Conclusions The checklist meets well with the structural criterion of scale designing with good reliability and criterion validity. Subjects are rated by professional staffs through observations, and children with high risk of ASD can be efficaciously screened out. The checklist was recommended for extensive use clinically.
Keywords: Checklist for Chinese infants with autism; Autism spectrum disorder; Reliability, Validity
