Validity of the Social Responsiveness Scale to Differentiate HF-ASD From ODD/CD

Background:  

Autism spectrum disorders (ASD) are characterised by the presence of impairments in communication, reciprocal social interaction alongside with inflexible behaviour patterns, interests and activities (DSM-IV-TR; ICD-10). The international classification systems aim to establish the presence or absence of categorically defined symptoms, but research indicates, that autism is the extreme end of a continuously distributed trait. The Social Responsiveness Scale (SRS, Constantino & Gruber, 2005) is a 65-item questionnaire for parents and teacher that explicitly focuses on the measurement of autistic traits in children and adolescents which can also be used as a screening instrument. A factor analytic study resulted in a one factor solution (Constantino et al., 2004). A recent study in Germany reported a ROC of 98% when comparing ASD individuals with healthy typically developing controls. The ROC was much lower (81%), when compared to clinical controls who suffered from a mixture of child psychiatric disorders (Bölte et al., 2011).

Objectives:  

As no study has been performed to date on the validity of the SRS to differentiate between high-functioning ASD and other specific child psychiatric disorders, the present study aimed to assess its diagnostic validity in differentiating ASD from oppositional defiant disorder / conduct disorder (ODD/CD), which is also characterised by difficulties in reciprocal social interaction.

Methods:  

55 individuals with HFASD, 55 age, sex and IQ matched children with ODD / CD and 55 typically developing children were included in the study. Diagnosis was done by ADI-R / ADOS, K-DIPS (child psychiatric diagnostic interview), IQ-measurement, specific questionnaires and clinical observation.

Statistical analysis: Pairwise ROC-analysis between groups, cut-off values and sensitivity/specificity are calculated. The mean total SRS-score is compared between groups, influence of IQ, age, and gender are calculated by linear regression analysis. Correlation analysis will be done with algorithm scores of ADI-R, ADOS, and measures of general psychopathology (CBCL).

Results:  

ROC-values, sensitivity and specificity were high, when HF-ASD and typically developing controls were compared, but considerably lower, when HF-ASD had to be differentiated from ODD / CD. Mean SRS total scores of HF-ASD were highest, followed by ODD/CD, lowest in typically developing controls. Gender, but not IQ effects were observed. Correlation with ADI-R and ADOS algorithm scores was in the medium range, CBCL social problems and aggressive behavior scores also correlated with the SRS total score across groups.

Conclusions:  

The SRS has been developed as a measure of quantitative autism traits in the general population. Clinically, it is also used as a screening instrument for ASD. The current study replicates the good validity of the SRS to differentiate between ASD and typically developing controls, but alerts to the possibility of false-positive ASD diagnoses in children with ODD / CD by the SRS.

The study was supported by a grant of the JW Goethe University`s Medical Faculty Frankfurt am Main (Heinrich and Fritz Riese foundation).