Autism tests don’t add up
Time to learn: A new approach to autism testing is long overdue. Picture: Susan Holt Simpson
Autistic girls and women could be missing out on diagnosis because the most commonly used test isn’t designed for us.
The AQ50, which assesses people’s ‘autism quotient’ from their answers to 50 statements, was created more than 20 years ago by Professor Sir Simon Baron-Cohen, director of the Autism Research Centre at the University of Cambridge, and his colleague Dr Sally Wheelwright, neither of whom is autistic.
The questionnaire was based on their observations of male presentation because, thanks to early studies’ focus on boys, autism was thought to be a predominantly masculine condition.
Extreme male bias
Professor Baron-Cohen even developed an ‘extreme male brain’ theory to explain away the few females receiving a diagnosis. According to this, autistic people of both sexes have a particularly ‘masculine’ way of thinking, marked by a predilection for data and systems, combined with low levels of ‘feminine’ traits such as empathy and imagination.
That’s why the AQ50 includes questions about noticing vehicle number plates and other strings of information, being fascinated by dates and numbers, remembering phone numbers and birth dates, and classifying things like cars and trains.
I have a formal diagnosis of autism, but when I took the test, I gave what are considered to be non-autistic answers to most of these.
A similar number of questions focus on autistic people’s supposed lack of imagination – we’re not meant to enjoy making up stories or reading fiction, for example – and, as an award-winning novelist, I lost several more marks by selecting the non-autistic options.
Falling through the net
Despite a significantly reduced score, I was still well within the autistic range, leading me to seek and receive an official diagnosis.
However, every autistic person is different. Many girls and women – as well as some men and non-binary people – will find losing marks on these questions puts them below the magic 26 out of 50 threshold at which diagnosis is considered likely.
This could mean they, or their GP, wrongly rule out autism as the cause of their problems, depriving them of a much-needed route to understanding and support, and the opportunity to build a happier and healthier life.
The NHS still maintains there are around ten autistic males to every female. More up-to-date thinking puts the ratio at between three-to-one and one-to-one. But until we have accurate testing, it’ll be impossible to know the true number.
Female-friendly testing
The good news is that researchers in the Netherlands are developing a new questionnaire, the AQ-f, that should be much more suitable for girls and women. As the name suggests, the format will be similar to that of the AQ50.
However, in their trials, while males taking the original AQ50 scored considerably higher than females, there was no significant difference between the sexes with the female-friendly AQ-f.
The researchers, who’re based at the universities of Amsterdam and Groningen, say: ‘The AQ-f seems to provide a more accurate reflection of the experiences and symptoms of women.’
They’ve promised once their research is complete, the new test will be made freely available on the internet.
Until then, if you’re a girl or woman taking the AQ50 as a step towards possible diagnosis, you could:
Reframe the data and systems questions to focus on things you’re unusually interested in, such as a particular celebrity, activity or type of collecting, or
Add a few extra points to your final score to counterbalance those questions, or even
Discount the problematic questions altogether, turn your score out of the remaining maximum into a percentage and divide by two to give a more accurate total out of 50.
These solutions may not be ideal, but they’re better than missing out on a potentially life-enhancing diagnosis because of an outdated and misogynistic test.