In the second of two blog posts by researchers from University College London, Ellie Roberts explores the advantages and challenges of using the Millennium Cohort Study to better understand the mental health of autistic young people.
In the last blog post, we shared how the Mental health and Wellbeing in Autistic Young people (M-WAY) study uses a mixed-methods approach – combining qualitative research, longitudinal cohort data, and lived experience perspectives – to better understand the factors shaping the mental health of autistic young people.
One theme emerged especially strongly throughout this process: the importance of social experiences.
Feelings of belonging, social connection, participation, and alienation were repeatedly highlighted as key influences on autistic young people’s mental health.
That’s why it’s important we have lived experience teams, who are people that share their experiences of the topics we are researching.
Discussions with these teams emphasised the importance of autism identification itself – particularly the timing of diagnosis and how being identified earlier or later in life may shape social experiences, support and longer-term outcomes.
Below we describe some of our studies using the Millenium Cohort Study (MCS) that further examine these themes and their implications for the wellbeing of autistic individuals.
The Millenium Cohort Study
The Millenium Cohort Study (MCS) is an ongoing longitudinal cohort survey tracking the development of approximately 19,000 individuals born in England, Wales, Northern Ireland and Scotland between 2000-2002. To date, there have been eight waves of data collection, ranging from age 9 months to 23 years.
The wide variety of social and psychological data available make the MCS ideal for autism research in young populations. Few cohort studies have collected such a range of data over time in a large youth sample, in addition to information on autism diagnosis.
We therefore chose to use data from the MCS to answer many of our research questions, which often focus on development and predicting outcomes.
1. How do social relationships and mental wellbeing in adolescence vary by autism status and gender?
Our recently published study investigated whether experiences of social relationships (social support and alienation) and mental wellbeing differ depending on autism diagnosis and gender in adolescents.
We found that autistic adolescents and girls had lower mental wellbeing.
The importance of social support and alienation for mental wellbeing was the same for autistic and non-autistic adolescents.
Considering the poor mental wellbeing of autistic adolescents, it is critical to improve their social experiences.
Interventions targeting autism awareness for non-autistic individuals and implementations of autistic peer support programs may subsequently be beneficial in promoting mental wellbeing for autistic adolescents.
2. Are mental health difficulties associated with the timing of an autism diagnosis in children?
This study looked at how emotional, behavioural and social difficulties are associated with the timing of an autism diagnosis in children.
It was observed that children who were diagnosed earlier had more emotional, behavioural and social difficulties. However, children diagnosed later had a faster increase in these difficulties over time, and by adolescence experienced greater difficulties than children who were diagnosed earlier.
While early-diagnosed children may have more obvious difficulties, late-diagnosed children may be overlooked and lack the support to prevent exacerbation of their difficulties.
It may therefore be useful for parents and teachers to be aware that such difficulties may not always appear severe, but should be acknowledged to reduce the likelihood of a late diagnosis and lack of suitable support.
3. What are the determinants of an autism diagnosis in childhood and adolescence?
Here we sought to identify factors that could predict the timing of an autism diagnosis.
We found that being diagnosed after starting primary school (age 5 years) compared to after starting secondary school (age 11 years) was predicted by living in poverty and the absence of any parental concerns.
Being diagnosed after starting secondary school was also predicted by typical-range intelligence (in other words, scoring within a normal range on measures of vocabulary, problem-solving, and spatial awareness).
Strategies promoting earlier identification of autistic children could help those more likely to be diagnosed later. For example, making teachers and parents aware of factors associated with later diagnosis to prevent these individuals being overlooked.
We also have some more recent findings using cohort datasets like the MCS that have not been published yet but are available to view online.
Challenges
While using large cohort datasets can provide longitudinal and representative data, this can come with drawbacks when researching the experiences of autistic individuals.
As mentioned, findings and interpretation of secondary data may not reflect the lived experiences of autistic individuals. Data on important topics such as autonomy, identity and sensory phenomena are sparse.
Understanding the context provided by lived experience is especially necessary to accurately interpret data on topics such as those which may be affected by masking, like social relationships, which may hide true associations.
Other technical challenges also arise with cohort data.
Measures validated in non-autistic samples may not behave the same in autistic individuals. For example, domains of mental wellbeing and social preferences may differ for autistic individuals, such as valuing solitude and alternative forms of social connection.
Interpretation of individual items within measures may also vary, as differences in communication, literal thinking, and alexithymia in autistic individuals may result in scores which appear identical but do not represent equivalent experiences in non-autistic individuals.
Subsequently, scores derived from these measures may not be comparable between autistic and non-autistic samples, leading to reduced construct validity and potentially inaccurate conclusions.
Autistic representation
In a longitudinal cohort study like the MCS, data completeness, both at each timepoint and over time, may be biased towards autistic individuals with lower support needs, creating a gap in data for those with more complex needs.
The requirement of long-term engagement and lengthy survey responses may lead to reduced participation of these individuals and their families, as well as heightened risk of attrition.
These individuals are therefore underrepresented and it is uncertain how well findings from the general autistic population may translate to such subgroups.
Furthermore, diagnostic heterogeneity and changing diagnostic criteria throughout the duration of longitudinal studies may create inconsistencies in case identification and complicate interpretation.
The collection of longitudinal data by studies such as the MCS and the Avon Longitudinal Study of Parents and Children (ALSPAC) span periods of substantial change in autism awareness and diagnostic frameworks.
Earlier data collection may have overlooked individuals with less stereotypical autistic traits, particularly females and ethnic minorities.
It is important to consider these challenges when working with cohort data, in the formulation and interpretation of research regarding autistic individuals, and to involve lived experience researchers wherever possible.
Summary
Longitudinal cohort datasets such as the MCS contain highly useful information that can be used to understand factors that influence development and outcomes in autistic young people. However, potential challenges of using such datasets should be considered when conducting autism research.
Nevertheless, the collection of data is continuously improving, with greater recognition of differences in autistic presentation, which will contribute to the production of more inclusive research.
Meet the author
Ellie Roberts is a PhD student at University College London conducting research on social relationships and mental health in autistic adolescents using longitudinal cohort datasets.
Her research interests revolve around early adversity and the interaction between social, biological, and psychological factors in the development of mental health difficulties in adolescence.
Connect with Ellie on LinkedIn.
Comment or question about this blog post?
Please email us!