Background: Increased rates of autism spectrum disorder (ASD) occur in a number of genetic neurodevelopmental syndromes. However, profiles of autism symptomatology may differ from those typically associated with ‘idiopathic’ ASD; ASD profiles also vary between syndromes. The Social Communication Questionnaire (SCQ) is a widely used screen to identify children at risk of ASD. The Autism Diagnostic Observation Schedule (ADOS and ADOS‐2) is a semi‐structured observational measure, often considered the ‘gold‐standard’ for aiding ASD diagnosis. The appropriateness of the SCQ and ADOS for identifying ASD in these groups remains unknown.

Methods: Social Communication Questionnaire and ADOS data were collected for 231 participants (age 4–59 years) with genetic syndromes: Down (DS, n = 68), Cornelia de Lange (CdLS, n = 59), Rubinstein–Taybi (RTS, n = 28), Fragile X (FXS, n = 51) and Cri du Chat (CdCS, n = 25). Vineland Adaptive Behavior Scales (VABS) scores were available for most (>90%) of the sample.

Results: Across the total sample, approximately 43% met criteria for ASD on both the SCQ and ADOS measures; 23% met criteria on neither. Of the remainder, 13% had scores at/above cut‐off on the SCQ but not on the ADOS, and 21% scored positively on the ADOS but not on the SCQ. Sensitivity of the SCQ in relation to ADOS/ADOS‐2 classifications was 68% and specificity 63%, both figures lower than those typically reported in studies of other autism cohorts. There was no consistent relationship between degree of concordance on these measures and participants' level of social, communication or daily living skills (as measured by the VABS). Agreement between SCQ and ADOS classifications (i.e. whether individuals met cut‐off for ASD) also varied between syndrome groups, with agreement highest in the FXS and DS groups and lowest for individuals with CdCS.

Conclusions: Caution is required when using standard ASD assessment instruments to identify autism in individuals with specific genetic disorders.