자폐스펙트럼장애 영유아 대상 한글판 Q-CHAT-10의 신뢰도 및 타당도 연구 = Reliability and Validity of the Korean Q-CHAT-10 for Infants and Toddlers with Autism Spectrum Disorder

국문 초록 (Abstract)

본 연구는 자폐스펙트럼장애 영유아 선별도구인 ‘Q-CHAT-10 (Quantitative Checklist for Autism in Toddlers 10)’의 한국판 표준화 연구의 일환으로 신뢰도와 타당도에 관한 예비적 연구를 진행하여 ‘한글판 Q-CHAT-10’의 유용성을 검증하고자 하였다.
연구대상은 현재 국가적 차원에서 개발 중인 ‘영유아 자폐성장애의 통합적 조기선별 도구 서비스 개발’로 진행된 연구에 참여한 14개월에서 48개월 사이의 영유아와 부모 131명이다. 연구대상은 두 집단으로 구분되었는데 자폐스펙트럼장애 집단은 50명으로 소아정신건강의학과 의사로부터 자폐스펙트럼장애 진단을 받았거나 전문가로부터 자폐스펙트럼장애로 진단받고 지역 아동임상센터에서 치료받고 있는 영유아이다. 공인 신뢰도가 검증된 평가자에 의한 ‘자폐증 진단 관찰 스케줄-2(ADOS-2)’ 실시결과 자폐 우려 수준이 ‘경도-중도의 우려’, ‘중도-고도의 우려’ 또는 ‘자폐증’, ‘자폐스펙트럼’에 속한다. 일반 집단은 81명으로 부모의 발달력 보고에서 특별한 결함이나 장애로 진단받은 경험이 없고 전반적인 영역에서 정상 발달이 보고된 영유아이다. 공인 신뢰도가 검증된 평가자에 의한 ‘자폐증 진단관찰 스케줄-2’ 실시결과 자폐 우려 수준이 ‘약간-전혀 우려되지 않음’ 또는 ‘비스펙트럼’에 속한 21명, ‘유아행동평가척도 부모용(Child Behavior Checklist: CBCL 1.5-5)’의 문제행동 총점T 점수가 60점 미만으로 정상범위에 속하는 56명, ‘영유아 중 아동 자폐스펙트럼 검사(Childhood Autism Spectrum Test: CAST)’ 점수가 15점 이하인 4명으로 구성되었다.
본 연구의 측정도구는 원저자로부터 허가를 받은 한국어로 번역된 ‘한글판 Q-CHAT-10’, ‘유아행동평가척도 부모용’, ‘아동기 자폐증 평정척도(CARS)’가 사용되었다.
자료 분석은 ‘한글판 Q-CHAT-10’의 기술통계와 문항 간 상관관계를 분석해 문항별 특성을 확인하였다. 내적 일치도와 검사-재검사 신뢰도를 통해 신뢰도를 검증하였으며, 구성 타당도 검증을 위해 탐색적 요인분석을 실시하고, 공인 타당도 검증을 위해 ‘유아행동평가척도 부모용’와 ‘아동기 자폐증 평정척도’ 간의 상관분석을 하였다. 척도의 효과성을 검증하기 위해 민감도와 특이도를 확인하고 절단값을 산출하였다.
본 연구 결과는 다음과 같다. ‘한글판 Q-CHAT-10’은 10번 문항을 제외한 문항 간 상관이 0.26~0.86(p<0.01)이었고, 문항 간 내적 일치도인 전체 Cronbach’s α지수는 0.87(p<0.01), 검사-재검사 신뢰도 지수는 0.91(p<0.01)로 높은 신뢰도가 검증되었다. ‘한글판 Q-CHAT-10’의 요인구조를 확인하기 위해 탐색적 요인분석을 실시한 결과 10번 문항을 제외한 모든 문항이 1요인으로 나타났다. ‘한글판 Q-CHAT-10’과 ‘유아행동평가척도’ DSM 진단척도 중 전반적발달문제 하위척도와의 상관지수는0.69(p<0.01), ‘아동기 자폐증 평정척도’와의 상관지수는 0.55(p<0.01)로 공인타당도가 검증되었다.
‘한글판 Q-CHAT-10’의 2점 득점방식은 총점인 2.5점을 기준으로 했을 때 민감도 0.72, 특이도 0.91로 적절한 수준으로 나타났다. 따라서 2점 득점방식에서 본 도구의 목적인 자폐스펙트럼장애 선별 목적에 맞게 민감도와 특이도가 가장 높은 통계적 절단값(cut-point)은 3점으로 판단되었다.
본 연구는 국내 자폐스펙트럼장애 영유아를 대상으로 한 선별검사 도구가 매우 제한적인 실정에서 ‘한글판 Q-CHAT-10’의 신뢰도 및 타당도를 국내에서 처음으로 검증한 것에 의의를 두며 이 도구가 앞으로 임상현장에서 유용하게 사용될 것을 기대한다.

다국어 초록 (Multilingual Abstract)

This study was intended to verify effectiveness of the Korean Quantitative Checklist for Autism in Toddlers & 10 items (Q-CHAT-10) by a preliminary study on its validity and reliability as part of a study for Korean standardization of Q-CHAT-10.
Subjects were 131 infants and toddlers aged 14 to 48 months and their parents participating in the government-led study to develop an instrument for integrated early screening of infants and toddlers with autism spectrum disorder (ASD). They were divided into two groups: general and ASD. The ASD group included 50 infants and toddlers diagnosed as ASD by child psychiatrists and child mental health specialists and treated in local child clinics. They belonged to mild-to-moderate, moderate-to-severe, autism, and autism spectrum ranges of concern of the Autism Diagnose Observation Schedule, Second Edition (ADOS-2) administered by an expert with verified reliability. The typical development group of 81 children were never diagnosed as special deficits or disorders and found to show normal development in overall areas based on the development history reported by their parents. Out of the 81 children, 21 children belonged to the none-to-minimal or non-spectrum range of ADOS-2 administered by an expert with verified reliability and 56 children were classified in the normal range with less than 60 of the total T problem behavior score by Child Behavior Checklist (CBCL) 1.5-5 for parents. Four children scored less than 15 in Childhood Autism Spectrum Test (CAST).
Instruments were Q-CHAT-10 translated into Korean under the permission of the authors, CBCL for parents, and Childhood Autism Rating Scale (CARS). Data analysis was conducted to check item characteristics by inter-item correlation analysis and technical statistics of the Korean Q-CHAT-10. Reliability was verified by internal consistency and test-retest reliability. Construct validity was checked by exploratory factor analysis and concurrent validity was verified by correlation analysis with CBCL for parents and CARS. Effectiveness of the Korean Q-CHAT-10 was verified by sensitivity and specificity and the cut-off point was calculated. The findings are as follows:
The inter-item correlation excluding item number 10 of the Korean Q-CHAT-10 was 0.26~0.86 (p<0.01), the total Cronbach’s α, which is inter-item internal consistency, was 0.87 (p<0.01), and test-retest reliability was 0.91 (p<0.01), proving high reliability of the instrument. Factor structure of the Korean Q-CHAT-10 was checked by exploratory factor analysis, finding that all items excluding item number 10 loaded onto one factor. The correlation coefficient between the Korean Q-CHAT-10 and the overall developmental problem, a sub-category of CBCL DSM was 0.69 (p<0.01) and the coefficient between the Korean Q-CHAT-10 and CARS was 0.55 (p<0.01), proving concurrent validity of the Korean -CHAT-10. Sensitivity and specificity of the Korean Q-CAHT-10, which scores zero or one point for each question item, were 0.72 and 0.91 respectively based on the total score of 2.5, standing at appropriate levels. Therefore, the statistical cut-off point having the highest sensitivity and specificity was judged as three points.
This study is very meaningful in that it is the first study to verify reliability and validity of the Korean Q-CHAT-10 when screening tools available for young children with ASD are very limited in Korea. It is expected that the Korean Q-CHAT-10 will be used effectively in the clinical scenes.

목차 (Table of Contents)

• Ⅰ. 서론·························································································1
• 1. 연구의 필요성 및 목적···························································1
• 2. 연구 문제·············································································3
• Ⅱ. 이론적 배경··············································································4
• 1. 자폐스펙트럼장애··································································4
• 1) 자폐스펙트럼장애 개념·······················································4
• 2) 자폐스펙트럼장애 특징·······················································5
• 3) 자폐스펙트럼장애 영유아 특성·············································7
• 2. 자폐스펙트럼장애 조기선별 및 선별도구···································8
• 1) 자폐스펙트럼장애 조기선별의 중요성····································8
• 2) 자폐스펙트럼장애 선별도구·················································9
• 3) 자폐스펙트럼장애 선별도구로서의 Q-CHAT-10···················15
• Ⅲ. 연구 방법················································································19
• 1. 연구 대상···········································································19
• 1) 연구 대상 선정································································19
• 2) 연구 대상의 인구사회학적 특성·········································20
• 2. 측정 도구···········································································21
• 1) 한글판 Q-CHAT-10························································21
• 2) 유아행동평가척도(CBCL 1.5-5) ·······································23
• 3) 아동기 자폐증 평정척도(CARS)·········································24
• 3. 연구 절차···········································································25
• 4. 자료 분석···········································································27
• Ⅳ. 연구결과·················································································29
• 1. 한글판 Q-CHAT-10의 기술통계와 상관································29
• 1) 한글판 Q-CHAT-10의 기술통계·······································29
• 2) 한글판 Q-CHAT-10의 문항 간 상관·································29
• 2. 한글판 Q-CHAT-10의 문항별 특성······································32
• 3. 한글판 Q-CHAT-10의 신뢰도 분석······································34
• 1) 한글판 Q-CHAT-10의 내적일치도····································35
• 2) 한글판 Q-CHAT-10의 검사-재검사 신뢰도························35
• 4. 한글판 Q-CHAT-10의 타당도 분석······································36
• 1) 한글판 Q-CHAT-10의 탐색적 요인분석·····························37
• 2) 한글판 Q-CHAT-10의 공인 타당도···································38
• 5. 한글판 Q-CHAT-10의 효과성 분석······································41
• 1) 한글판 Q-CHAT-10의 민감도와 특이도·····························41
• 2) 한글판 Q-CHAT-10의 로지스틱 회귀분석··························45
• Ⅴ. 논의 및 결론···········································································47
• 참고문헌······················································································52
• ABSTRACT·················································································63