Towards understanding the under-recognition of girls and women on the autism spectrum

It is only in recent years that research has begun to focus on gender differences in males and females on the autism spectrum. The first International Conference in the United Kingdom addressing this was in 2008, and this was a result of women themselves questioning their often late diagnosis. In Kanner’s original group of children, there were 11 males and 3 females (Kanner, 1971). Asperger (1944) only referred to males but later recognised that there were females with his pattern of behaviour. In our original epidemiological study of autism, we came up with a male/female ratio of 2.6 to 1 (Wing and Gould, 1979). However, the children all had intellectual disability, and possibly, their symptoms were better defined as more classically resembling Kanner’s original descriptions. Only a few children fitting Asperger’s description were included in the study as all the children had special needs and were in special education. Children in mainstream education were not included.

There is now an increasing awareness that we are missing women and girls on the autism spectrum and the assumption has been that there are more males with autism or Asperger syndrome. The questions needing to be asked are as follows:

Three of the papers in this special edition attempt to address these questions.

Dean et al. (2017) suggest that girls on the spectrum mask or camouflage their symptoms, but in many ways, girls are not unlike their typically developing female peers. The study explores in detail how the girls interact compared with the boys. Observations were made comparing autistic boys and girls with their typically developing peers in playground activities.

Gender differences in social behaviour were evident throughout the sample. Typically developing boys tended to play differently from typically developing girls. The boys were more likely to play organised games, with the boys on the spectrum tending to play alone. The social challenges of boys with autism spectrum disorder (ASD) were more obvious than the social challenges of girls with ASD. From a distance, girls with ASD ‘looked like’ typically developing girls. They spent a significant amount of time talking and engaging and weaving in and out of groups. Untrained observers were less likely to pick up the difficulties shown by the autistic girls. The girls’ social challenges were concealed from the playground attendants but importantly were not hidden from their peers. The girls were not able to maintain mutual engagement in activities requiring social synchronisation and were not able to adjust their behaviour to align with group norms. In conclusion, using camouflage to mask social challenges makes girls vulnerable and less likely to receive intervention. The point made is that an understanding of how boys and girls with ASD fit into the social landscape at school is necessary when planning social interventions. Currently, there is a need to recognise that the social domain is where the girls struggle most. A social curriculum designed to recognise the difficulties girls experience should be part of our National Curriculum (Gould and Ashton-Smith, 2011). Dean et al.’s study emphasises that more research is needed to examine the social behaviours and acceptance of boys and girls with ASD throughout primary and secondary school. The subtleties of social interaction become more complex in adolescence and young adulthood; therefore, the problems are more likely to be picked up in younger girls.

Interestingly, it is pointed out that more research is needed to examine social behaviours and social competence comparing autistic children with typically developing children with varying degrees of competence.

The second paper by Lai et al. (2017) attempts to operationalise camouflaging in autistic adults by comparing males and females. The authors suggest that a measure of camouflaging can be made by examining the quantitative discrepancy between the person’s ‘external’ behavioural presentation in social–interpersonal contexts and the person’s ‘internal’ status. They found that this operationalised camouflaging measure was not significantly correlated with age or intelligence quotient (IQ) but, on average, autistic women had higher camouflaging scores than autistic men, but there was variability in groups. Their findings call into question the clinical impression that females who camouflage their social difficulties require considerable cognitive effort leading to possible stress, anxiety and depression.

Although recognising that camouflaging may be an integral part of the female phenotype of autism, that is not to say that it does not occur in males. The core difficulties of social interaction, social communication, social imagination and repetitive behaviours are evident in males and females. Furthermore, neither women nor men consistently conform to gender stereotypes.

In both the papers of Dean et al. (2017) and Lai et al. (2017), it is pointed out that the heightened tendency to camouflage difficulties in females in both social interaction and social communication may not be picked up by teachers, primary care workers or unenlightened diagnosticians, making an ASD diagnosis less likely. Improving our understanding of camouflaging and identification of masking symptoms is important to enhance timely diagnosis and support for both males and females.

The way autism affects individuals is highly variable. Lai et al.’s paper states that the distributions of camouflaging scores overlapped substantially between autistic men and women. On average, there was a sex/gender difference, but they say that this should be viewed as a phenomenon reflecting individual differences in social coping rather than a diagnostic behavioural pattern distinguishing females versus males at an individual level. Neither did they find a significant relationship between camouflaging and measures of anxiety symptoms in either sex or gender. However, this was an adult population and there may be higher risk of anxiety in younger ages. Also, interestingly, there was a positive association with depressive symptoms in men. The authors speculate that men may be more susceptible to the burden of camouflaging than autistic women. It is suggested that perhaps women have had more practice and might be better adapted to implementing camouflaging due to gender-related social experience and demands.

I would speculate that both males and females with good verbal skills and good intellectual ability tend to learn the social rules through their intellect rather than by instinct or intuition, and maybe, this is more evident in females for the reasons given in Lai et al.’s paper.

The paper by Duvekot et al. (2017) looks at the factors influencing the probability of a diagnosis comparing girls versus boys. This is a large-scale study derived from a multicentre sample of referred children aged 2.5–10 years old. The question asked is why referred girls are less likely to be diagnosed with ASD than boys.

There were several gender differences regarding diagnosis in the total sample – girls were on average older, had higher levels of average IQ scores, more interrelationship problems as reported by parents and lower levels of autistic symptoms reported by teachers. A significant interaction effect with gender was found relating to repetitive behaviour, such that higher levels of restricted and repetitive behaviour tended to be less predictive of an ASD diagnosis in girls than in boys. Girls were more likely to be diagnosed with ASD when they had higher total levels of behavioural problems.

A very important finding was that girls were less likely to receive an ASD diagnosis based on the standardised diagnostic instruments. Diagnostic instruments and/or manuals need to be adapted to provide examples that are more characteristic of girls. This is particularly the case for special interests and routines. Boys tend to be more active and have more interests in technical hobbies and facts, and girls are more passive and collect information on people rather than ‘projects’. The interests of girls on the autism spectrum are often similar to those of typically developing girls, for example, animals, horses and classical literature. It is not the special interests that differentiate them from their peers but the quality and intensity of these interests and the time spent on them (Gould and Ashton-Smith, 2011). The repetitive behaviours in girls with ASD are not adequately captured by most of the current diagnostic instruments and are less likely to be recognised by clinicians as being characteristic of ASD.

Evaluation of sensory symptoms is particularly important for the evaluation of ASD in girls. As yet, there are few studies exploring sensory differences between males and females. The Diagnostic Interview for Social and Communication Disorders (DISCO) (Wing et al., 2002) is a semi-structured interview which enables a clinician to systematically collect information which not only can lead to a diagnostic label but, more importantly, can also give a detailed profile of the person. For clinical purposes, this is more helpful in defining the needs of the individual than just the diagnostic label itself.

Using the DISCO, emphasis is placed on the different ways in which behaviour is manifested in females for all aspects of social interaction, social communication, social imagination and the ways repetitive behaviours and routines may be different in females. A dimensional approach to diagnosis rather than using cut-offs is more effective in the diagnosis of females on the autism spectrum.

From clinical experience, gathering information from the individual gives us insight into the way the person thinks and how they perceive themselves. Relying on information gathered from informants is part of the process, but as has been demonstrated in all three papers, exploring the cognitive and behavioural processes underlying both the strengths and difficulties gives us a better understanding of this very complex condition.