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

Diagnoses were self-reported through an online panel and were not verified against clinical records, so we do not know for certain who in the autistic group had been diagnosed or whether anyone in the comparison group had been missed. Only 5% of the autistic group had an intellectual disability, and the authors say people with above-average cognitive ability are likely to have been over-represented, so the results may not carry over to autistic adults with greater support needs or to people who cannot fill in a questionnaire themselves. People who preferred the rewritten wording tended to have higher general cognitive ability, which the authors say suggests people with lower IQ may have different or fewer preferences about the language. It is not known whether the questionnaire tells autism apart from conditions that commonly occur alongside it, such as ADHD or social anxiety. The authors say establishing this matters if the questionnaire is to be used to decide who gets referred. No cut-off score, sensitivity or specificity has been worked out, so the questionnaire cannot yet be used to make referral or diagnostic decisions. The authors state the study was underpowered for the fuller statistical work needed, including confirming that the questionnaire measures the same thing in men and women, in children and adults, and at different levels of intellectual ability. The sample was UK-based and largely from England, no ethnicity data were collected, and the authors note that wording preferences are likely to reflect the participants' wider cultural context. The four extra items in the 18-item version have only been tested in adults filling it in about themselves; whether they work in children, or when a family member fills it in, is untested. The two groups could not be matched on gender: 15 autistic participants described their gender as different from their sex at birth, and no non-autistic participant did.

The only funding information in the text supplied names the Waterloo Foundation and the Leverhulme Trust, both research charities. The text contains no author declaration about competing interests, and it does not say whether the funders had any role in designing the study, analysing the data or writing it up. Four autistic consultants and two autistic pilot participants were paid for their input.

Do not read this as a screening test you can start using with adult clients. The questionnaire was compared in a volunteer online sample whose diagnoses were self-reported, no cut-off score exists, and there are no figures on how often it correctly identifies autism or how often it wrongly flags people who do not have it. It has also not been shown to tell autism apart from ADHD or social anxiety. The authors themselves say further testing is needed before it is used in practice.