If a person thinks they might have autism, they will usually see their general practitioner as the first port of call. The GP will use a screening test to help them decide if the person should be referred to a specialist for a formal diagnosis.
However, we recently discovered, by chance, an error in the clinical guidelines doctors use when making their initial autism assessment. That guideline is published by the UK’s National Institute for Health and Care Excellence (Nice), and it also informs clinical guidelines internationally.
The guideline on autism recommends that doctors use the autism spectrum quotient (AQ10) to measure autistic traits. AQ10 contains ten statements about autistic traits, such as, “I find it difficult to work out people’s intentions”, which reflects research showing that autistic people find it difficult to understand what other people are thinking. The maximum score is ten, and higher scores represent more autistic traits.
Nice recommends that people who score seven or more should be referred for a full diagnostic assessment by their doctors, but the correct value should be six or more. Nice examined the suitability of the AQ10 based on research in 2012 showing that a value of six or more was optimal, but the final guidance appears to be incorrectly published with the “seven or more” recommendation.
A score of six or more is the optimal screening value because this makes the AQ10 sensitive enough to help identify autistic adults but also specific enough to help rule out people without autism.
The difference between six and seven might not seem large, but a one-point difference on a test with only ten statements has a big impact. The incorrect seven or more cut-off recommended by Nice makes the AQ10 far less sensitive. This is because many people who score six are likely to be autistic but may not have been referred for clinical assessments as needed.
Autism diagnoses may have been delayed as a result, and some people may not have been diagnosed at all. Many people will not have received the correct support at the right time and are likely to have experienced additional mental health difficulties, such as anxiety, which commonly accompany autism.
A decade of incorrect screening
To make matters worse, the Nice guidelines have been in place since 2012, so incorrect autism screening has been occurring for almost a decade. Many researchers have also used the incorrect Nice guidance, for example, by incorrectly recruiting study participants based on their AQ10 scores.
It is impossible to put a number on how many autism diagnoses have been affected by the incorrect Nice guidance. By highlighting this error, we hope to enhance the discussions people have with their doctors, rather than undermine this important relationship.
We have now informed Nice about the issue – they are yet to reply, and their guidance needs to be reviewed. Until then, doctors and researchers should use the correct six or more cut-off score, in line with the original research.
Several new autistic personality trait tests are emerging, creating a better future for understanding autism. They include a shortened version of the AQ10 and a test that uses a different scoring method.
There have also been improvements in detecting the signs of autism in children, as well as using films rather than surveys to guide doctors. While these developments are yet to work their way into the Nice guidelines, they will improve autism screening, diagnosis, and research in the coming years.
Lucy Waldren receives funding from the Economic and Social Research Council.
Punit Shah receives or has received funding from the UKRI Medical Research Council and Economic and Social Research Council.
Rachel Clutterbuck receives funding from the Economic and Social Research Council.