Will is the first author of this paper and sets the scene by clarifying what constitutes as ‘late diagnosis’ when talking about autism in children, before turning to the paper itself and providing a summary of what they looked at in this study.
Will talks us through the methodology used and shares an overview of the findings, including further insight into why some children get missed and the role of diagnostic overshadowing. Will also mentions that girls were overrepresented in the late diagnosis group and explores why this might be the case.
With this study showing that the late diagnosed group had milder problems in early childhood, presumably making them harder to identify, Will comments on what more could be done to find, assess and correctly diagnose children, before discussing what the implications are of his findings for CAMH professionals.
Please subscribe and rate our podcast from your preferred streaming platform, including; SoundCloud, iTunes, Spotify, CastBox, Deezer, Google Podcasts, Podcastaddict, JioSaavn,
Listen notes, Radio Public, and Radio.com (not available in the EU).
Will Mandy is a clinical psychologist and professor at University College London (UCL), where he is joint director of the UCL clinical psychology training course, which is the largest of its kind in Europe. His work aims to improve the recognition of autism, and to develop new interventions to help autistic people. He has a particular research interest in improving the identification and care of females on the autism spectrum, who are currently at high risk of going unnoticed and unhelped by clinical and educational services. He also studies sub-diagnostic autistic traits in non-clinical populations, and the role these can play in the development of a range of common childhood, adolescent and adult mental health problems. With colleagues at Great Ormond Street Hospital’s National Centre for High-Functioning Autism he has developed and trialled interventions to help children with autism transition from primary to secondary school, and to teach children about their autism diagnosis, with an emphasis on fostering their sense of self-worth and pride.
[00:30] Jo Carlowe: Hello, welcome to the In Conversation podcast series for the Association for Child and Adolescent Mental Health, or ACAMH for short. I’m Jo Carlowe, a freelance journalist with a specialism in psychology and I have with me Will Mandy, Professor of neurodevelopmental conditions at the research department of clinical, educational and health psychology at UCL London. Will is the first author of the paper ‘Mental health and social difficulties of late-diagnosed autistic children, across childhood and adolescence’, recently published in the Journal of Child Psychology and Psychiatry.
If you’re a fan of out In Conversation series, please subscribe on your preferred streaming platform, let us know how we did with a rating or review and do share with friends, and colleagues.
Will, welcome. Nice to meet you. Can you start with a brief introduction about yourself.
[01:28] Professor Will Mandy: Thanks very much for having me on the podcast. I’m Will Mandy. I’m a clinical psychologist by training and I work as you said, at UCL where I’m actually joint director of the clinical psychology training course. But I’m also an autism researcher. Been doing that for about 20 years now. And I have a particular interest in trying to better understand the needs and characteristics of what you might call under diagnosed autistic people.
So I’ve done quite a lot of work on autistic girls and women, but also other groups of people who often get missed, including those who first present with mental health difficulties actually. And so this work kind of grew out of those sorts of interests.
[02:05] Jo Carlowe: We’re going to look at the JCPP paper. But before we go into details, can you clarify what constitutes late diagnosis when talking about autism in children?
Professor Will Mandy: I mean, there’s no accepted definition and I think different people will have different views. The literature that we reviewed for this paper has tended to view late diagnosis as something that happens after children have been in their first couple of years of primary education, maybe from six or seven years onwards however. I think it is important to contextualize that with the fact that many, many people in the UK, and I think in other parts of the world, get diagnosed well after that.
We know that from actually previous work that we’ve done. If you take autistic kids who are diagnosed by the age of 14, about a third of them don’t get diagnosed until they get to secondary school. And we also have this large wave of people now who have been diagnosed in adulthood.
So whilst I think it’s a very fair definition to say late diagnosis is after the age of about seven years, I think it’s worth remembering that that’s actually setting quite a high bar compared to what’s actually happening out there in practice at the moment.
[03:11] Jo Carlowe: Well, let’s turn to the paper ‘Mental health and social difficulties of late-diagnosed autistic children, across childhood and adolescence’, which was recently published in the JCPP. What did you look at in this study? Can you give us a summary?
[03:25] Professor Will Mandy: So we were interested in emotional, behavioural, and social difficulties in a group of 430 autistic children. When I talk about emotional behaviour and social difficulties, in some aspects I’m talking about mental health, things like anxiety and depression, what people sometimes call internalizing problems. I’m also talking about behaviour and conduct difficulties, sometimes known as externalizing problems. So difficulties with regulating your behaviour or doing things that might infringe on the rights of other people. And in social difficulties, talking about things like peer problems, struggles getting along with other people, perhaps difficulties establishing friendships, perhaps becoming a victim of bullying. Those sorts of things.
So in the realm, our view was if we studied emotional behaviour and social difficulties, you’re getting quite a good overall picture of how somebody is getting on, how their life is going, how they’re doing in these important different domains. And crucially, we were interested in the trajectories of these difficulties, where people started at the beginning of our period, which was between about the ages of 3 and 14 years, but also the sort of change and continuity in those difficulties over time.
And we were interested in whether these trajectories of these emotional, behavioural, and social difficulties, whether they differed depending on whether somebody received timely diagnosis by the age of seven years or whether they received what we call a late diagnosis, i.e. one between the age of 8 and 14 years.
And we wanted to do this first to get a sense of the levels of need of late diagnosis kids prior to their diagnosis. We wanted to see OK, if somebody is diagnosed at 9 or 10 or 11, how were they getting on when they were 5 and 7 before they’d been picked up by services and before they’d been identified as autistic? So that was one reason why we wanted to do that. But we also thought that by studying the trajectories of these emotional, behavioural, and social difficulties in relation to age of diagnosis, it might actually give us some insight into what drives the age at which you were diagnosed. What factors might influence whether you get picked up earlier or whether you actually end up waiting until late primary school or even secondary school to have your autism recognized.
[05:33] Jo Carlowe: Can you say something about the methodology that you used?
[05:37] Professor Will Mandy: We used a sample from a study called the Millennium Cohort Study, which is a wonderful UK resource for researchers. So what this is it’s a birth cohort. This was when they attempted to recruit a large sample, about 19,000 kids, that was representative of the general UK population. And they did that at the turn of the century, hence this name Millennium Cohort Study. And then they followed the sample up at regular intervals over time right up until the current day in order to generate data on all sorts of elements of these young people’s lives, in order to address all sorts of questions of clinical, social, political, education importance. So that was the study that we were able to draw on. And of relevance to us and to our conversation today, there’s a couple of really important things they measured in that study. So first of all, they’ve been asking parents at regular intervals whether their child has been diagnosed with autism spectrum disorder or some variant of that term. So we have a good sense of who in the sample has received an autism diagnosis, and we also have a sense of the timing of that diagnosis because they ask each wave of data collection. And then the other thing that they did was that they measured the emotional, behavioural, and social difficulties of these young people at multiple time points between the age of 3 and 14 years. And they did that using a measure called the strengths and difficulties questionnaire, which is very widely used parent report measure of these sorts of characteristics I’ve been talking about. And then using those data, we could actually model the trajectories of these difficulties over time. So we were able to look what’s their sort of start point, in terms of where they’re at in terms of these different difficulties. But also, we can look at the rates of and the nature of change in these difficulties over time using a technique called latent growth curve modelling. And I must say that I’m just one author on this paper, and in particular Emily Midouhas, a colleague of mine at UCL and also Eirini Flouri also at UCL should take the credit for that sophisticated statistical modelling that they did there.
But one of the great things about latent growth curve modelling is it yes, it gives you an average trajectory for a group of people, but it also manages to chart differences in trajectories between people. And then what you can do is you can look at other factors that might predict those difficulties. And of course, in this study the one we were really interested in was timing of diagnosis. Were there different trajectories of these difficulties, and in these sophisticated models that Emily and Eirini did, that were associated with timing of diagnosis. And we were able to do that statistically whilst controlling for a bunch of other things so that we could get as close as possible to really trying to establish some sort of relationship between timing of diagnosis on the one hand and change in continuity in emotional, behavioural, and social difficulties on the other.
[08:33] Jo Carlowe: Well, let’s go to the findings. Can you give us an overview?
[08:37] Professor Will Mandy: So I suppose I’d like to emphasize three things. Firstly, generally earlier in the study period, when they were younger the late diagnosis had lower levels of emotional, behavioural, and social difficulties compared to the earlier diagnosed kids. In those aged 3, 5, and 7, they tended to score lower on our measures.
The second thing is though, which is kind of intriguing, is that subsequently they tended to show a faster rate of growth in these difficulties compared to the earlier diagnosis. So that actually, by the end of our study period, by the age of 14, the late diagnosed kids tended to score either equally or even higher on these measures of difficulties compared to the earlier diagnosis. So there was a catch up effect if you like, in terms of these difficulties.
And I suppose the third point I want to make, just to kind of nuance what I said before, I did say that the late diagnosed kids had lower levels of emotional behaviour and social difficulties than the earlier diagnosis, it went prior to their diagnosis if you like, when they were younger. However, it is important to note that there were still pretty high levels of difficulties in that late diagnosed autistic group. But I suppose what I wouldn’t want is people think, oh so this was a group of young people who were all just doing absolutely fine. There is some evidence that quite a substantial proportion of them were struggling, perhaps with their mood, perhaps with their behaviour, perhaps with their social interaction even in that period prior to their diagnosis. They just didn’t have levels that were quite equivalent on average to the earlier diagnosed.
[10:07] Jo Carlowe: Will, can you say more about why some children get missed?
[10:11] Professor Will Mandy: It’s an interesting question. I think like anything, it must be multiply determined. There’s all sorts of risk and protective factors that can sort of nudge your probabilities of being identified or not one way or the other. So the literature has identified a bunch of things that are pretty consistently associated with later diagnosis. One is being a girl, another is having higher IQ, another is having less obvious and overt autistic characteristics. Both of those make sense, don’t they? And that it’s just in a sense it’s a less obvious and a less perhaps, obviously impairing set of characteristics that mean that people go longer. There’s also a social dimension to this. So there’s some evidence that people who are from less well-off families or who are from racial minorities identity are more likely to fly under the diagnostic radar for longer. I think there’s an important dimension there.
And then I suppose what our study adds is that– what I would think is pretty plausible idea based on the data that we’ve shown in this paper, is that there’s also something about your pattern of functioning and well-being also probably influences the timing of your diagnosis. And so what we tended to find was the points at which the later diagnosed young people tended to reach the age where they were diagnosed was also the point at which their emotional, behavioural, and social difficulties started to equal and exceed those of the earlier diagnosed children.
And so one plausible explanation of that finding is that there’s something about a young person is presenting perhaps with anxiety, they’re really struggling to get along with peers, perhaps they’re acting out and they are not able to behave at school for example. And that kind of puts you on the grownups radar. They start to think, OK, well what’s going on here? And that can then I think, lead to in some cases to a referral to an autism service and assessment and so on.
[12:11] Jo Carlowe: I’m wondering about the role of diagnostic overshadowing. I believe your paper came to some interesting conclusions that contradict previous findings.
[12:21] Professor Will Mandy: So maybe I’ll give an example from another area of research I do, which is on eating disorders and autism because there’s quite a well-established overlap between autism and particularly anorexia nervosa, restrictive eating problems. And what we find in that case is that so for women who are autistic and have anorexia, the average age of their eating disorder diagnosis is about 18 years, and the average age of their autism diagnosis is about 27 or 28 years.
And I think what often happens, and this is where the diagnostic overshadowing becomes in is that they are obviously they’re in distress, they’re having severe difficulties of their mental health in terms of this eating disorder. That’s what attracts the attention and people see that. They attribute people’s difficulties to an eating disorder but it overshadows, it almost disguises the sort of underlying autism. That’s also very important to understand in that person’s life and in fact to understand their eating disorder. And there’s been evidence for example, quite a lot of evidence actually, that ADHD can diagnostically overshadow autism, particularly in the States is evidence that actually if you have an ADHD diagnosis, that can be associated with a later autism diagnosis, almost as if you go to the clinic you get assessed and say oh, yes this young person’s got ADHD and we can attribute all of their sort of struggles they might be having at the moment to ADHD and the autism gets missed.
Now our findings actually are really rather incompatible with a picture of diagnostic overshadowing actually. Because what our findings seem to be saying is that the more mental health difficulties or associated difficulties you have, the more likely you are to receive an autism diagnosis. And it’s that process I talked about before, I think that that’s probably going on, which is just that somebody is feeling anxious and depressed in an obvious way. They really are having trouble getting along with peers, people become concerned about them and start to think what might be going on, how could we understand this, how can we explain this and that can, in a sense promote that person’s chances of receiving autism diagnosis.
So in a way I mean, I think that’s potentially a bit of a good news story and the implication is that these young people when they are being assessed clinically are getting a good holistic assessment. It doesn’t just stop it at their anxiety, but it also considers the possibility that might be a neurodevelopmental condition there.
[14:41] Jo Carlowe: Let’s turn to something you mentioned earlier. You mentioned that girls were overrepresented in the late diagnosis group. Can you elaborate? How do you explain that?
[14:49] Professor Will Mandy: There’s been a lot of interest in this recently. So I suppose the first thing to say is there is a really widespread and widely replicated finding that autistic girls and women are more likely than autistic boys and men to fly under the diagnostic radar. So for example when autistic girls are diagnosed, it tends to be later on average than the boys. And we also know that autistic girls are more likely to be missed because we see disproportionately high number of women in adult autism assessment services than men with the implication perhaps, that there’s a kind of– they particularly likely to be in this kind of lost generation of autistic adults.
So people have been trying to understand this to work out why this is happening largely so they can sort of try and put it right and address this gender imbalance or inequality here. And the idea that people have really pursued is that there’s basically consistent differences in how autism presents in girls compared to boys and women compared to men. And this more female typical presentation might not fit our current conceptualization of autism, which after all are largely based on male cases.
So for example, there’s been quite a lot of work done on levels of social interest and social motivation, with the finding that again, on average, I’m not talking about every girl and every boy here, but on average autistic girls are more likely than autistic boys to be quite socially oriented and socially motivated. And I think that can delay diagnosis. There’s stuff done around the focused interests or the fascinations that many autistic people develop with findings that when these occur in boys, they often look more striking and unusual than they do in girls. And I think that can also make it more likely that a boy is sent for an autism assessment than a girl.
And then there’s the sort of other area, which is to do with something called social camouflaging. So there’s this idea that fundamentally, autistic people find themselves living in a world that wasn’t designed well for them, that was designed for and by the non-autistic majority. And so very regularly find themselves in situations that are fundamentally are unaccommodating to them. And so what do you do when you’re in that situation? One solution that many autistic people find is that they develop strategies to kind of adapt their behaviour so that they can pass as non-autistic. Sometimes people call this things like adaptive morphing or mask.
What we’re finding is that whilst boys and girls, men and women both camouflage, girls are more likely to do it than our boys. And so that might well also be delaying diagnosis and resulting in people being picked up less. And I think it’s really interesting to note at this point, we don’t have the designs, the research designs, that can really help us tease apart to what extent might a sex gender based differences we see on the autism spectrum be driven by biology and to what extent are they driven by culture, to what extent are they sex differences, to what extent are they gender differences.
But it does seem very, very likely that gender if you like, the kind of societal norms and constructions around how boys and girls, the expectations and how they should behave in the norms, they should conform to, that seems to be really important here as well. And certainly speaking to autistic women who were missed in childhood and only diagnosed as adults will often talk about how because they were girls, they were kind of almost expected to camouflage and mask more, whereas they felt that boys would get away, almost, with just being off doing something else and not being social and so on.
We know, in fact, actually, there are some stereotypes where people fundamentally, they could see the same behaviour in a boy or a girl. And they’re slightly more likely to attribute that behaviour potentially to autism in the boy rather than the girl because they just harbour stereotypes that autism is more of a male condition. But there’s a whole range of factors that drive that gender difference in age of diagnosis.
[18:48] Jo Carlowe: It sounds like a lot of work needs to be done on changing the classifications really or updating them.
Professor Will Mandy: Yes, and that’s an interesting point you make there. Or I think changing the way those classifications are applied just to encompass more diverse ways in which autism can be expressed and the ways in which it’s influenced by sex and gender.
[19:10] Jo Carlowe: Will, I want to turn to something else you mentioned earlier. So your study shows that the late diagnosed group have milder problems in early childhood, presumably making them harder to identify. Given this, what more do you believe can be done to find, assess and correctly diagnose children?
Professor Will Mandy: It raises some interesting dilemmas I think, because what we don’t want to be doing is going out there and hunting out kids who are doing absolutely fine in their lives and enjoying their lives, and sort of slapping diagnostic labels on them for the sake of it. I don’t think that serves anybody. However, on the other hand, what we do know is whilst many of these later diagnosed kids may have had on average lower levels of mental health difficulties, they did have some difficulties. So many of them could have benefited from the diagnosis. And also we know that for many of them, there will be this growth in these difficulties over time. And so that does raise the question. We don’t know the answer to this yet but does raise the question of whether had they been diagnosed earlier and then given appropriate support. Maybe they wouldn’t have experienced this growth in mental health difficulties. I’m not saying our study proves that but I’m saying it raises that as a question.
So I do think that there is this group of people who could benefit from a more timely diagnosis, especially things like before they make that transition to secondary school, when there’s this real escalation and the complexity of the kind of educational and social environment. And for many kids who are autistic and going into that transition without the diagnosis and the support and understanding it can bring, that can go wrong and really damage their well-being and their life chances. I think that one of the ways we can really help matters is actually through training people.
For example, one of the interesting things about trying to improve diagnosing autism is that autism is often diagnosed by specialists. It’s considered to be a diagnosis that should be made by specialists who know a lot about autism. But actually what we need to do to make sure that children who need an assessment are getting to their specialist clinics is to better educate generalists. And so I really think there’s a role, for example, training GPs, family doctors in the UK who are really kind of an access point to most forms of health care in the UK about identifying children who might be presenting with, let’s say anxiety or social difficulties and who might actually benefit from an autism assessment. And also teachers I think are another key person in this because they spend a lot of time with the kids, they see them in all sorts of contexts and can be very, very important in helping identify those who might actually benefit from autism. So I think there’s some quick wins almost actually, that could be gained from bit of training and just generally raising awareness and making sure that people’s understanding of autism is up to date and reflects if you like, the more subtle presentation.
[22:14] Jo Carlowe: Will, you raised the possibility in your paper that there are subgroups of autistic young people who have distinct trajectories in terms of their emotional, behavioural, and social difficulties. I’m not sure if you’re going to be able to answer this. But are these trajectories inevitable for certain subgroup or would early intervention change the trajectory by mitigating against later difficulties assuming these children are diagnosed earlier?
Professor Will Mandy: Yeah. I think it’s just an incredibly important question as you alluded to. And your question it’s not one that I know the answer to yet, certainly based on research that’s been done so far. And I think that could be a really interesting next step actually, is to really model a bit more what these subgroups are, test whether they really are groups for example, who we might expect. There might be one group for example, who is characterized by a decline in levels of anxiety and another who’s characterized by rises in it. And to understand the differences between the people in those groups so we can begin to understand a bit more about what mechanisms might be driving it. But also it would be very interesting to study, naturalistically looking at these groups and trying to get a sense of what interventions have they access? What in their environment, what’s changed or adapted to help them? And we might begin to generate ideas from that sort of work about the extent to which trajectories can be impacted by environmental change and psychological intervention and so on. And could help us generate ideas about good practice that in turn could give rise to experimental work like randomized controlled trials that could really nail down whether there is a causal effect of certain interventions of environmental modifications on mental health in autistic young people.
[23:58] Jo Carlowe: Let’s turn to CAMH professionals. You mentioned earlier the need for more training for generalists. I’m wondering what the implications are of your findings in this paper for CAMH professionals.
Professor Will Mandy: I have a view that we are too specialized in our autism assessments. This is just my view and some people won’t agree with me. I preface my comments with that. But almost through historical accident in my view, there are some conditions where you can walk into your GP and say a few things and you can end up with a diagnosis on your record that maybe you’ve got depression or so on with a very widespread professionals feel confident in making a diagnosis. And then there are other conditions like autism where there’s a general sense that you need a very intensive assessment from people who are very, very specialist in that area. And OK, there’s a lot to be said for that. It’s not always easy making an autism diagnosis but there are some problems with that.
Firstly, certainly in the UK, we’re seeing very, very long waiting lists for autism assessments at the moment. And I think that partly reflects the fact that this specialist model that we have just can’t keep up with demand in the population now. I think another issue with that specialist model is that what we now know is that undiagnosed autistic people present with difficulties across a range of services– eating disorder services, mental health services, and I also suspect, actually a range of health care services as well. And so there is an argument that if we actually spread the expertise a bit more widely about how to do autism assessments, that could lead to a better care model.
Let’s say that a young person presents with adolescent onset anorexia and she goes to her local service, and questions are raised about whether she’s autistic. And if in the eating disorder team where they have expertise in eating disorder and they can also do the autism assessment, A, she doesn’t have to wait two years for her referral to another service. And B, there’s more chance of integrated care where the insights about her autism assessment are really properly integrated with insights about her eating disorder. So I suppose my first point I want to make is a kind of service redevelopment point, which is, I think, there needs to be a kind of a bit of a structural change in terms of more training and supervision for general CAMHS workers, perhaps with specialist CAMHS workers, autism workers supervising them. But nevertheless, a sharing of that expertise across the system.
So another thing that I would advise CAMHS workers is to always consider the possibility of neurodevelopmental factors in pretty much any work they’re doing. Working with a young lad with conduct problems who keeps getting excluded from primary school. We know that undiagnosed autistic boys are very highly represented in that group. Working with a teenage girl who’s developed severe social anxiety. Again, we know that autism and social anxiety in adolescence are associated. So I think it needs to be factored in routinely to assessment.
And just the third thing, I think that CAMHS workers have a real role to play in skilling up the people if you like who feed into their service. So for example, I know of colleagues in one service in North London who did some work with teachers in the area telling them more about autism in girls. And what that did was it actually changed the male to female ratio in their clinic. They found they were getting more appropriate referrals of girls who if they hadn’t done that work, would have had a higher chance of being missed. So I think those are a few things that I think are relevant from this work for CAMH Service.
[27:28] Jo Carlowe: Will, are you planning would follow up research that you can share with us?
Professor Will Mandy: We are, we’re really interested in the mental health of autistic young people because we think that it’s something that has a huge impact on their being and their life chances and on their families. And also because we’ve been told repeatedly actually by the autism community that this is a very, very high priority for them. That’s what they’d like to see researchers doing. So we’re going to be doing more of this longitudinal work. A, charting the different trajectories and whether there are different trajectory groups of autistic young people who are characterized by distinct patterns of change and continuity over time in their mental health and functioning. And B, we’re trying to get at mechanisms a bit more, trying to understand what are risk factors for example, being somebody who’s anxiety greatly escalates during adolescence? Because we feel that can pave the way for the development of both preventative measures and also the development of intervention. And I think another thing I’d like to emphasize in this work is that we’re trying to be especially interested in environmental risk and protective factors. There’s been a bit of a tendency to date for researchers, and I would include myself in this, to take quite an individualized approach to trying to understand mental health.
So the basic idea would be you think OK, autistic people are more likely to feel anxious. What is it about those individual autistic people that causes them to be more likely to develop anxiety? Is it because they have an intolerance of uncertainty or is it because they have sensory processing difficulties? My view, those are reasonable questions to ask but you’re slightly missing the point as well if you’re just asking those questions. Because you could say, well, what is it about them and the environments they are expected to be in that is making them feel anxious? What is it about the interactions, if you like? And in a way we know, for example, that bullying is incredibly common unfortunately, of autistic people. And so to study the development of anxiety, depression without thinking about peer victimization seems like you’re really missing the point there.
So we want to include a much better consideration of the environment and how that influences the lives of autistic people. And crucially, and this is a tricky thing, in my view to model, the fit between the individual and the environment. That individual’s particular set of capacities and what their environment is asking and demanding and requiring from them. Because my view is it’s actually instances where there’s a misfit between those two things.
[30:08] Jo Carlowe: Finally, Will, what is your take-home message for those listening to our conversation?
Professor Will Mandy: I think the take-home message is that there are a group of young people who are autistic and have a good chance of being diagnosed once they get to secondary school, but who would likely benefit from their autism being identified earlier than that, in their primary school years. And through training of teachers and family doctors and other people, and to generally spreading awareness and making sure we have assessment methods that are sort of sensitive to a range of autistic presentations, we can have a better chance of picking those kids up early and hopefully giving them the support they need in order to flourish and to have satisfying and enjoyable lives.
[30:53] Jo Carlowe: Brilliant. Thank you so much. For more details on Professor Professor Will Mandy, please visit the ACAMH website, www.acamh.org and Twitter @acamh. ACAMH is spelled A-C-A-M-H. And don’t forget to follow us on your preferred streaming platform. Let us know if you enjoy the podcast with a rating or review and do share with friends and colleagues.