Gender differences in parent-reported age at diagnosis of children with autism spectrum disorder

https://doi.org/10.1016/j.rasd.2018.02.003Get rights and content

Highlights

  • Girls are only diagnosed with ASD later than boys when diagnosed over age 5 years.

  • Toileting problems and temper problems are associated with earlier ASD diagnosis in both boys and girls.

  • Additional diagnoses are associated with later ASD diagnosis in both boys and girls.

  • Toileting problems, temper problems, and additional diagnoses are stronger predictors of diagnosis age in girls than in boys.

  • Parents and professionals should consider these characteristics as ‘red flags’ during ASD assessment in school age girls.

Abstract

Background

Autism spectrum disorders (ASD) are more commonly observed in boys than in girls. There is growing awareness of ASD in girls and recognition that under-diagnosis is common. The current study aimed to investigate any evidence of reduction in the average age at diagnosis for girls by assessing whether: 1) girls’ age at diagnosis has reduced, compared to boys’, across two age cohorts – children born between 1996–1999 and 2002–2005; 2) age at diagnosis differed between boys and girls diagnosed across childhood; 3) any characteristics are associated with earlier age at diagnosis in girls.

Methods

Data were available from large UK databases of children with ASD: The Database of Children with Autism Spectrum Disorder Living in the North East (http://daslne.org) and the Autism Spectrum Database–UK (www.asd-uk.com).

Results

There was no differential reduction of parent-reported age at diagnosis for girls over time. For children receiving their diagnosis at age ≥60 months, boys received diagnoses an average of one year earlier than did girls (98.2 months, SD = 31.6 vs. 109.1 months, SD = 36.4). For boys and girls, earlier diagnosis was associated with toileting problems and temper problems. Having additional diagnoses (e.g., dyslexia, dyspraxia, and epilepsy) was associated with later diagnosis.

Conclusions

Age at diagnosis has not decreased over time. Girls with ASD are diagnosed later than boys when aged 5 years or older. Health and education professionals would benefit from better understanding factors such as toileting problems, temper problems, and additional diagnoses that could potentially guide early identification of ASD in clinical practice for school-age girls.

Introduction

The reported prevalence estimates of autism spectrum disorder (ASD) are always greater in males than females. The male to female ratio is around 4:1 (Baird et al., 2006; CDC, 2014; Fombonne, 2009) in the absence of intellectual impairment (see Rivet & Matson, 2011 for a review). Increased rates of ASD diagnoses in children over the last 15 years or so may be due to the broadening of, and changes to, ASD diagnostic criteria and practice, improved identification, earlier age at diagnosis in the teenage years, and different methodologies used to estimate prevalence (Russell, Collishaw, Golding, Kelly, & Ford, 2015).

Cognitively able girls may be diagnosed with ASD significantly less frequently and at an older age than boys despite there being no gender differences in the age at which parental concern is expressed (Giarelli et al., 2010). However, with the exception of learning/intellectual disability, there is little evidence to suggest what factors might be associated with earlier age at ASD diagnosis in girls. Indeed, several studies have suggested no gender difference in age at diagnosis (Mussey, Ginn, & Klinger, 2017). In a cross-cohort comparison study, Russell et al. (2015) found no gender difference in age at diagnosis nor differences in gender ratios of diagnosed children aged 7 years assessed in 1998/1999 (n = 96) and 2007/2008 (n = 209). Furthermore, in a large UK cohort study, Brett, Warnell, McConachie, and Parr (2016) found that the average age at diagnosis was 67.3 months for boys and 72.1 months for girls and had not decreased over the decade from 2004 to 2014. Although some children are now diagnosed by age 2 years, Brett and colleagues found no reduction in age at diagnosis for children diagnosed under age 3 years in the UK (Brett et al., 2016). While gender was not a significant influencing factor, earlier age at diagnosis was associated with language regression, lower socioeconomic status, greater degree of support required, greater symptom severity, and greater parental concern about initial symptoms (Brett et al., 2016; Daniels & Mandell, 2014).

There has been growing interest in the identification of ASD in girls, and it is becoming recognised that under-diagnosis or later diagnosis may be common (Dworzynski, Ronald, Bolton, & Happé, 2012; Loomes, Hull, & Mandy, 2017; Van Wijngaarden-Cremers et al., 2014). One way to measure whether there has been better detection of ASD in girls in recent years is to show whether there is evidence of change in the average age at diagnosis for girls. While Brett et al. (2016) assessed factors that influenced age at diagnosis, they did not specifically examine factors that are associated with earlier age at diagnosis independently for boys and girls. Thus, the current study builds on that of Brett et al. (2016) to examine gender differences in parent-reported age at diagnosis in large UK databases of children with ASD.

We had the following aims: First, to assess whether girls’ age at diagnosis has reduced, compared to boys’, across two age cohorts – children who were born between 1996–1999 and 2002–2005. Changes in diagnostic practices have been shown to have a substantial effect on the increased prevalence of ASD (King & Bearman, 2009). Thus, this separation of birth cohorts was chosen to better control for this potential confound of changes over a decade, so that any evidence of reduced age at diagnosis in girls is less likely to be an artefact of general changes in diagnostic practices. If ASD diagnosis in girls has become more timely, we would expect a differential reduction in age at diagnosis, for girls compared to boys, between the age periods. Second, to investigate whether age at diagnosis differed between boys and girls diagnosed across childhood. We assessed gender differences in age at diagnosis across childhood by grouping the sample at the median age at diagnosis where any differences between boys and girls would most likely be salient (see below). Finally, to examine characteristics that might be associated with earlier age at diagnosis in girls.

Section snippets

Methods

Data were available from two large representative UK databases: The Database of children with autism spectrum disorder living in the North East of England (Daslne, established in 2003; http://daslne.org) and the Autism Spectrum Database – UK (ASD–UK, established in 2011; www.asd-uk.com). Daslne covers six areas around Newcastle upon Tyne, whilst ASD-UK covers the rest of the UK. By 2017, the databases held data from over 4000 families, including information on children’s ASD and other medical

Results

Parent-reported age at diagnosis ranged from 7 to 213 months (mean = 67.0, SD = 37.2; median = 54, interquartile range = 46); only 5 parents reported an age at diagnosis below 12 months of age, therefore all parent-reported data were included. Chronological age at data collection ranged from 18 to 214 months (mean = 101.0, SD = 45.2; median = 94, interquartile range = 72). Table 1 shows descriptive statistics of chronological age and age at diagnosis of boys and girls with ASD by Year of Birth

Discussion

This large study found that, compared with boys, girls diagnosed with ASD at 5 years of age or older received their diagnosis an average of one year later. Whilst this was a small effect statistically, this difference can have a significant impact for children, families, and clinicians. This delay reduces opportunities for understanding girls’ difficulties, and accessing community support and interventions, educational support and strategies at school and for older young people, the workplace.

Declaration of conflicting interests

The authors declared that they had no conflicts of interest with respect to their authorship or the publication of this article.

Author contributions

A. Petrou analysed the data and J. R. Parr, and H. McConachie conceptualised the study and contributed to conceptual analysis. A. Petrou, J. R. Parr, and H. McConachie co-wrote the paper, and had complete access to the study data that support the publication.

Acknowledgements

We are grateful to the parents and children participating in Daslne and ASD–UK, and the child health teams who support the success of the databases through recruitment (see www.asd-uk.com for a list of participating NHS Trusts). We would also like to thank colleagues who have worked on the databases and members of the Steering Committee, Research and Parent Committees. We are grateful to the UK autism research charity, Autistica, for funding Daslne and ASD–UK (see www.autistica.org.uk).

References (28)

  • A.M. Daniels et al.

    Explaining differences in age at autism spectrum disorder diagnosis: A critical review

    Autism

    (2014)
  • E. Fombonne

    Epidemiology of pervasive developmental disorders

    Pediatric Research

    (2009)
  • J. Gould et al.

    Missed diagnosis or misdiagnosed? Girls and women on the autism spectrum

    Good Autism Practice

    (2011)
  • D. Hambrook et al.

    Empathy, systemizing, and autistic traits in anorexia nervosa: A pilot study

    British Journal of Clinical Psychology

    (2008)
  • Cited by (18)

    • Relationship between cognitive ability and predictors for age at the time of autism spectrum disorder diagnosis

      2022, Psychiatry Research
      Citation Excerpt :

      Similarly, some studies have found that female sex was associated with an older age at ASD diagnosis (Begeer et al., 2012; Petroua et al., 2018), while others found no such association (Mishaal et al., 2014; Twyman et al., 2009). These inconsistent findings might be attributable to methodological differences between studies (Table 1), including differences in statistical approaches (Bickle et al., 2015; Brett et al., 2016; Hrdlicka et al., 2016), data collection (Brett et al., 2016; Bickle et al., 2015), diagnostic tools (Hrdlicka et al., 2016; Mishaal et al., 2014; Petroua et al., 2018), and participant age (i.e., study population characteristics) (Brett et al., 2016; Mishaal et al., 2014; Wang et al., 2018). However, cognitive ability is the most robust dichotomizing factor in ASD, segmenting the population into those with and without ID.

    • Sex differences in the diagnosis of autism spectrum disorder and effects of comorbid mental retardation and attention-deficit hyperactivity disorder

      2022, Journal of the Formosan Medical Association
      Citation Excerpt :

      Culture-based sex differences have been found to influence the acquirement of a diagnosis of ASD and subsequent intervention. In relation to ASD, studies exploring the influence of sex worldwide have revealed inconsistent findings: some have revealed no association between sex and age at diagnosis of ASD,10,11 while others have reported a significantly later age of diagnosis for female ASD patients.12–14 Factors related to different manifestations of ASD in males and females include stereotypes of biological vulnerability and developmental trajectories across different sexes in light of cultural beliefs, and might lead to different ASD diagnostic behaviors between the sexes.15

    • The effects of sex on prevalence and mechanisms underlying neurodevelopmental disorders

      2020, Handbook of Clinical Neurology
      Citation Excerpt :

      At this stage, there are no sex differences in core symptoms, cognitive measures, or adaptive behavior (Andersson et al., 2013; Postorino et al., 2015; Fulton et al., 2017). In fact, sex is not associated with earlier diagnosis (Petrou et al., 2018; Zwaigenbaum et al., 2019). While there are some findings of later diagnosis in girls (Shattuck et al., 2009; Begeer et al., 2013; Petrou et al., 2018), the most common finding is no difference in age at diagnosis in boys and girls (Daniels and Mandell, 2014; Loomes et al., 2017).

    View all citing articles on Scopus
    View full text