Abstract
The 10-item Autism Spectrum Quotient (AQ10)1 is used internationally for autism screening, in line with the National Institute for Health and Care Excellence (NICE) guidelines.2 We have found a worrying discrepancy between the clinical cutoff recommended by NICE and the research informing their guidance.1,3 The NICE Guideline Development Group examined the suitability of the AQ10 for autism screening, on the basis of research indicating that a cutoff score of 6 or higher should inform referrals for specialist diagnostic assessment.1 However, NICE incorrectly recommend that a score above 6 (ie, 7 or higher) should be used for screening purposes. This cutoff of a score of 7 or higher was, in fact, examined and rejected by the NICE Guideline Development Group in favour of the 6 or higher value,3 which leads us to conclude that NICE have erroneously recommended the higher cutoff.
This discrepancy in the AQ10 cutoff scores is concerning because of its far-reaching effect on clinical practice and research. Screening accuracy is intrinsically related to cutoff values, whereby the 7 or higher cutoff set by NICE is less sensitive than the correct 6 or higher value. Because the AQ10 is used by general practitioners (typically the first to identify people who might have autism), the insufficiently sensitive implementation of this screening tool will be contributing to missed referrals, diagnoses, and opportunities for intervention. Some researchers are also using the incorrect 7 or higher cutoff,4 often misattributing this value to the original AQ10 research.1 Others use the correct 6 or higher cutoff, but mistakenly attribute this value to NICE guidance.5 Erroneous NICE guidance thus underlies several inconsistencies in the use and reporting of the AQ10, raising broader concerns about the robustness of recent research on autism and co-occurring psychiatric conditions.
In consideration of these issues, the NICE guidance on autism should be revised, emphasising the correct 6 or higher AQ10 cutoff. NICE recommendations are deeply embedded into resources for clinicians, particularly general practitioners, who should be made aware of the correct clinical cutoff. We hope that this Correspondence will initiate this process and, until there is revised NICE guidance on autism, clinicians should use the 6 or higher instead of the 7 or higher cutoff value in their practice. Furthermore, the extent to which the 7 or higher cutoff has resulted in the misinterpretation of research is unknown. Where required, researchers should reanalyse previously published data and future studies should use the correct AQ10 cutoff. It might, thereby, be possible to establish the extent to which the (mis)use of different cutoff values has been consequential, which could prompt corrigenda to published research. In sum, we call for the urgent review of and move towards the correct use of the AQ10 in clinical practice and research.
We declare no competing interests.