Predictability of Self-Report Questionnaires (RAADS-R-NL, AQ-28 AND AQ-10) in the Assessment of Autism Spectrum Disorders in Adults

Background:  

Although screening instruments for autism spectrum disorders (ASD) are widely used, none have been simultaneously evaluated in an outpatient setting in which the instruments are commonly used. This is unfortunate because the surge in referrals for ASD assessments in adults highlights the need for diagnostic tools in clinical practice.

Objectives:  

To assess if the reported high psychometric properties of screening instruments for ASD in adults also apply to the naturalistic outpatient setting.

Methods:  

We tested the Ritvo Autism Asperger Diagnostic Scale (RAADS-R) and two short versions of the Autism Spectrum Quotient (AQ), the AQ-28 and AQ-10 in 210 patients referred for psychiatric assessment and 63 controls. Six specialized outpatient settings in the Netherlands took part. Assessing clinicians were blind to the RAADS-R and AQ scores. At the end of the study, the clinical diagnosis (ASD or no ASD) was related to the RAADS-R and  AQ-scores collected prior to the assessments.

Results:  

Of the 210 patients, 139 received an ASD diagnosis and 71 received another psychiatric diagnosis. The scores in the clinical groups were higher than in the control group. In addition the scores in the clinical subgroup that was diagnosed with ASD were higher on all three instruments than in the subgroup not diagnosed with ASD. Despite these diffferences, the percentage correct diagnoses was rather low (RAADS-R 68%, AQ-28 and AQ-10 61%). Both the sensitivity and specificity of each of these instruments was insufficient, where the sensitivity of the RAADS-R was the highest (73%) and the AQ short forms had the highest specificity (70 and 72%).

Conclusions:  

None of these instruments have a sufficient predictive validity to be used to predict clinical outcome in out-patients settings.