The Signposting Questionnaire for Autism-2 (SQ-A-2): using input from autistic adults to inform questionnaire development

Background: Brief questionnaires help clinicians to identify people who may require in-depth assessment for Autism Spectrum Disorder (hereafter autism) and also facilitate research. However, many current questionnaires do not map onto the latest DSM-5 criteria for autism and were developed by researchers and/or clinicians, with limited input from autistic people. The Signposting Questionnaire for Autism (SQ-A) is a caregiver-report questionnaire for measuring autism signs in children, based on DSM-5 criteria. Here, we developed a revised lifespan version, the SQ-A-2, with input on wording from autistic adults. We tested the SQ-A-2 as a self-report measure in autistic and non-autistic adults, and assessed the impact of the adapted wording on both the psychometric properties and the acceptability of the measure. Methods: The SQ-A, containing highly discriminating DSM-5 items from a valid clinical interview, was revised to form the 18-item SQ-A-2, with self- and informant-report options. Following input from a team of four autistic consultants and six autism professionals, item wording on the SQ-A-2 was adapted, thus forming Original and Adapted versions. Subsequently, a sample of autistic and non-autistic adults (N = 302), matched on age, sex, and general cognitive ability, completed both versions of the new SQ-A-2, as well as the widely-used 10-item Autism-Quotient (AQ-10). Autistic participants also rated their preferences for Original versus Adapted items, and provided qualitative responses for their reasoning. Results: Overall, the self-report version of the SQ-A-2 demonstrated excellent internal reliability and convergent validity with the AQ-10, and autistic adults scored significantly higher than non-autistic adults. This group difference was greater for the Adapted compared to the Original version of the SQ-A-2, suggesting autistic consultation may facilitate the development of clinically-useful measures. Additionally, autistic participants preferred the wording of the Adapted versus Original items, both quantitatively and qualitatively. Limitations: Our findings may not generalise to autistic people with intellectual disability. Ethnicity data was not collected. Conclusions: The adapted version of the SQ-A-2 has the potential to be a valid, clinically useful questionnaire for measuring DSM-5 signs of autism across the lifespan. Our findings also highlight the utility of incorporating autistic input into measure development. Further psychometric testing of the SQ-A-2 in larger, more diverse samples, including the informant-report version, is now required.