Objective: To develop and validate INCLEN Diagnostic Tool for Autism Spectrum Disorder (INDT-ASD).
Design: Diagnostic test evaluation by cross sectional design.
Setting: Four tertiary pediatric neurology centers in Delhi and Thiruvanthapuram, India.
Methods: Children aged 2-9 years were enrolled in the study. INDT-ASD and Childhood Autism Rating Scale (CARS) were administered in a randomly decided sequence by trained psychologist, followed by an expert evaluation by DSM-IV TR diagnostic criteria (gold standard).
Main outcome measures: Psychometric parameters of diagnostic accuracy, validity (construct, criterion and convergent) and internal consistency.
Results: 154 children (110 boys, mean age 64.2 mo) were enrolled. The overall diagnostic accuracy (AUC=0.97, 95% CI 0.93, 0.99; P<0.001) and validity (sensitivity 98%, specificity 95%, positive predictive value 91%, negative predictive value 99%) of INDT-ASD for Autism spectrum disorder were high, taking expert diagnosis using DSM-IV-TR as gold standard. The concordance rate between the INDT-ASD and expert diagnosis for 'ASD group' was 82.52% [Cohen's k=0.89; 95% CI (0.82, 0.97); P=0.001]. The internal consistency of INDT-ASD was 0.96. The convergent validity with CARS (r = 0.73, P= 0.001) and divergent validity with Binet-Kamat Test of intelligence (r = -0.37; P=0.004) were significantly high. INDT-ASD has a 4-factor structure explaining 85.3% of the variance.
Conclusions: INDT-ASD has high diagnostic accuracy, adequate content validity, good internal consistency high criterion validity and high to moderate convergent validity and 4-factor construct validity for diagnosis of Autistm spectrum disorder.