[MUSIC PLAYING] We are the Association for Child and Adolescent Mental Health, or ACAMH for short. And this is ACAMH Learn. Welcome to Mind the Kids, a podcast series dedicated to exploring the latest advancements in child and adolescent mental health research and practise. I'm Clara, an academic clinical fellow in child psychiatry, and I'm passionate about understanding and addressing the diverse mental health challenges faced by young people globally. In this series, I will be joined by renowned researchers and clinicians from around the world to discuss their cutting edge research, innovative interventions and best practises in child and adolescent mental health care. And today, I have the pleasure of receiving Professor Tamasin Ford, Professor of Child and Adolescent Psychiatry at the University of Cambridge. Tamasin leads the Child and Adolescence Resilience and Mental Health Research group, the ChARM group, and her work in psychiatry epidemiology has influenced policy and practise worldwide. So no one better than her really to be discussing the topic is the prevalence of mental health conditions in children and young people increasing or not. So thank you so much, Tamsin, for being here today. And over to you. Thank you for inviting me to come and speak about this really important issue around ADHD, autism, the provision of services, and the increasing controversy with assertions of overdiagnosis over medicalization. It's such an important issue. I wanted to start by going into the heart of the matter. So in the UK, we're very fortunate to have national data, and you have been involved and led since 1999, the British National Service and the British National Service, or MHCYP for those in the audience who don't know. A representative of longitudinal study that has been following the mental health of children and young people at a population level in England, and the expert on them is Tamsin. So I wanted to ask you, what are the results of those surveys telling us? We have had three surveys for children and young people, one in 1999 that recruited about 10,500 children and young people aged between 5 and 15. And this had a follow-up of the original sample. Three years later, a fresh sample were resurveyed again. Same method, same team. This was slightly smaller. We actually over recruited massively for the first one, which doesn't often happen in research, but seeing as we had the interest, we decided to get as much data as we could. For the second one, we recruited to target and it was just shy of 8,000 children across England, Wales, and Scotland, as the first survey had been, and they were aged 5 to 16 to cover compulsory school age. The third survey was in 2017 and actually was only England, and was particularly exciting because it went down to the age of 2 and it went up to the age of 19. Now, where we can compare across the three surveys, which is obviously the age group between 5 and 15, there's no evidence of an increase in the population prevalence of ADHD. Autism was not included in the first survey. There were three questions about were there whether speech problems, were their problems making and keeping friends, and were their repeated interests or movements. Just three questions. Partly because autism at that point was conceptualised as a rare and severe condition. And with a prevalence that was estimated to be 1 in 10,000. And as we were only supposed to be recruiting, 8,500, therefore, we didn't expect to find many cases. In fact, with those three questions, we've probably got around about 30 young people that even with those questions, yes, no, and a bit of qualitative information, we were able to say, actually, I think there are cases of autism there. So for the next survey, there was a module developed to ask in much more depth. So we could be much more certain. And it's been included ever since. So you can't read three data points in order to say something is definitely increasing or decreasing. Because you can always draw a straight line between two points. And whereas for anxiety and depression, the emotional disorders, there is a very clear increase that also interesting was picked up amongst 16 to 24-year-olds in the adult mental health survey in 2014. And again, in their most recent survey, which reported this year and last year, there just wasn't any evidence that ADHD is increasing. And when you go out to the general population, regardless of who's been to see a doctor or a specialist or who's been told they've got ADHD, when you go and do an assessment and you interview the parent and you interview the child, and if they allow you, you get some data from school. At that level, there hasn't been much of a change. However, we did ask about medication use in 2004 and again in 2017, and the proportion of young people who have this research population diagnosis of ADHD according to the research assessment, the proportion taking medication had doubled, which fits with what we have we know from prescriptions. And we know that referrals for ADHD have massively increased. We know from NHS England data of what clinics are doing. But interestingly, when you look at this administrative prevalence, so prescriptions and actual appointments with clinics in the NHS, the proportion at a population level doesn't equate to the research or population prevalence. In other words, we may be treating many more children with ADHD. The number of clinical diagnoses in electronic health records has massively increased for both ADHD and autism, and prescriptions have massively increased. Actually, we are still seeing fewer children with ADHD and autism than we would expect from what we found in the national surveys. In other words, we are not overdiagnosing ADHD. We are underdiagnosed it, but it is certainly true that we are recognising it more. This is a really important point and a really interesting one as well, because I think because of the overlap and the interface that ADHD has on educational attainment, for example, something we have much more awareness today. And as you said yourself, some people have claimed that ADHD has been overdiagnosed, but actually we have data to prove that that's not the case. It's more that people are talking about it more. How do you see this increase in mental health awareness? Do you think that this also influences people's perceptions on what gets diagnosed? I don't know for certain, but I'd be very disappointed if it didn't. But I think it's really important that everybody understand, understands the fact that ADHD and autism diagnoses require a careful assessment that takes into account multiple perspectives. For example, if we take ADHD to meet diagnostic criteria, you have to have impairment and it has to be across at least two, but usually every aspect of life. So a situation where a school might be very worried about a child and the parents aren't would raise questions. And the other way around, which is more typical that the parents are worried and perhaps the schools are not seeing it. But the schools might or parents might not be seeing it because they don't know what to look for. And I guess understanding that there are other things beside ADHD, for example, that can cause problems with focusing your attention. If you're very anxious or very depressed, for example. If you have an eating disorder that's severe enough, you'd be malnourished. Your concentration, your memory not going to be good. And that's why assessment needs to take into consideration, not only does this child have ADHD or does this child have autism, but does this child have anything else that might explain these problems, and that's not an either/or. Actually, if you have ADHD or autism, you are at an increased risk sadly, because life is difficult. It's more difficult for you than it is for children without these problems. You then have an increased risk of developing other conditions such as anxiety, depression, behaviour problems later on if you are not adequately supported, which is why picking these conditions up quickly and instituting support rapidly is so, so important. We do have very effective treatment for ADHD with a really strong evidence base. The problem is that families in a recent survey reported waiting. 40% of them waited more than a year from the point of referral to the point of receiving a diagnosis. That's a whole academic year, or three terms of missed education. Not to mention the fact that the child almost certainly had problems right from the beginning of going to school. But there are certain points in a educational career where ADHD is more likely to become prominent and children are more likely to be referred, and those relate to when the academic demands suddenly step up. So there's some on entry to school. There's some at the beginning of key stage one, when the expectation is very much that you're able to sit, listen, organise yourself, follow instructions, which a child with ADHD would find harder. And also the step up from primary to secondary school, where you go from being in one room with one teacher who knows you well to having to relate to a whole swathe of teachers having to find yourself a way around a big school, and to have the right books and the right equipment with you for the right lesson, which is a huge increase in organisational demand. You mentioned something very interesting about the importance of those assessments being done, like using more than one source of information and by professionals who have training. And also interestingly, you mentioned, we have survey data and research data showing that on one hand, we know that there isn't an overdiagnosis at population level because at population level, the prevalence of ADHD is not increasing. But we have an increase in the number of diagnoses in electronic health records and an increase in prescriptions, which and clinically as well, I've been seeing an increasing number of patients that come with private assessments and assessments that are done elsewhere, which not necessarily is a problem. But I think, as you said, it's really important that when people seek out those assessments because the waiting list is really long and sometimes you don't want to wait, which is completely understandable. But it's done by people who have the appropriate training because as you said, having ADHD, it's not just having an impairment in one area of your life, but more than one. Yeah. I think this is really important. Interestingly, when I did my PhD, which was all about which children with what types of mental health conditions got to which services, I did ask about private health care and got so few responses that I couldn't report it. And in fact, most of the private activity was around educational psychology assessments and legal support for tribunals appealing for special educational needs. But again, there were tiny, tiny numbers and so few that you risked identification of those individuals if you reported them. So I didn't talk about it in any of my papers. Flash forward to 2017 and the picture is very different. It is now a really quite large chunk of the access to services around mental health is via private practitioners. Now that in and of itself is not a bad thing, but it does mean that people are functioning outside the integration of primary care with GPs and secondary care with CAMHS, and may not have the training, supervision, et cetera that they need to do it properly. And that's where misdiagnosis could come in. And I think certainly anecdotally, I do hear comments, and it's the reason that much of this work then gets duplicated, because NHS services get landed with a request to prescribe for an assessment that they didn't do or their colleagues didn't do, and they're worried about the quality. So actually, we end up with duplication of effort. The private sector isn't streamlining and it's not necessarily better. I don't think you necessarily need to be hugely experienced in children's mental health. I'm not suggesting that all diagnoses have to be made by psychiatrists, for instance. I think it is perfectly possible to train a health practitioner in a way of systematically assessing the history, the developmental history, what the difficulties now and how they are impairing, and to consider a list of what else might be causing it. You don't necessarily need highly, highly trained individuals, but you do need to do it thoroughly. And you can back that up with tests like the QB test, which is not diagnostic in and of itself, but it can provide evidence. I think a big part of the overdiagnosis debate and it's that people think, oh, we have this increase at a population level of certainly depression and anxiety in children and young people. And we have services that are struggling to keep up with this demand. And yeah, and I read this really nice essay arguing that basically the problem was not if too many people were being diagnosed, but whether we were willing to put in the effort and the funds to invest in science and solutions as we've been talking about, solutions that are flexible and that accommodate people who, for example, are on the waiting list. And speaking of which. So I know that you and your team, you have been working on interventions for young people who are on wait lists, for example, and other interventions as well. But yeah, I wondered if you could tell us a little bit more about these interventions. And if you think these interventions targeting people who might have more mild difficulties or who are on the waiting list still like who haven't yet reached services, they could be rolled out to the NHS at a wider scale. I do think we need to get smart about this. If we are only managing a fraction of those who need help, we're not in a position where we're going to double the number of child psychiatrists or clinical psychologists or personnel working within child mental health services. I think for ADHD, there is really interesting work going on that is looking at how we can support families whilst they're waiting. So there is work going on led by Abby Russell in Exeter, at the medical school there, where she has co-developed with teachers and with parents, and to a lesser extent, children with ADHD. A toolkit for teachers, so that if they're worried about children who look like they might have ADHD or indeed have a diagnosis of ADHD in a primary school, there is a systematic way of narrowing down and thinking, what are the main one or two problems that are tripping this child up in the classroom that I want to solve? Let's try these strategies. Let's do it in a systematic way and review how it's going and see whether it has an impact. Now, that's yet to have randomised control trial evidence that it's effective, but it is certainly feasible and acceptable and people are really keen to access it, which is a hint that it's certainly looks like there is a niche. So hopefully, there will be a trial conducted of that shortly. There are some non-pharmacological ways of helping children with ADHD that we could think about seeing if we can develop into interventions that can easily be accessed either via an app or via perhaps with or without a little bit of supported help. I think the small but growing literature in this area suggests that just handing people an app or a book is not really felt like you're really taking them seriously. So there's a kind of perception problem there, and it may be that a little light touch support and guidance makes a big difference to whether or not people engage. And certainly, once you get to the self-help stuff that's just out there, it's really interesting how even if it's a really popular app and it's downloaded thousands and thousands of times, only a tiny proportion of people then actually open it and log on. And maybe that's particularly so in families of children with ADHD, because of the nature of the condition and the stresses that face these families, that it's particularly difficult to organise the time. So I think we really need to get smart as to how we can support people better and wiser, and that it shouldn't necessarily make wait for diagnosis. We need to tier the response or stage the response as we do for things like blood pressure and obesity. You first tell someone with very mildly elevated blood pressure, you need to try and lose some weight. You need to try and exercise. You need to watch what you eat. There's some lifestyle changes. And in fact, similarly for children with mild ADHD, there's some scaffolding that parents and teachers could do, which might just make the difference between the child coping or not coping and the family feeling able to deal with the situation or not and likewise the school. When we come to medication, which for the moderate and severe ADHD has a very, very strong evidence base that clearly needs a higher level of assessment and a higher level of monitoring, blood pressure and pulse rate and weight, et cetera because of potential adverse effects. I don't think any of this is beyond the health service to do it, but I think some of the ways that services are organised and delivered currently are really tripping us up. So for example, we moved to a pathway model in child mental health services so that you don't have services used to work with age. So children and adolescents broadly speaking, and then started having specialist ADHD clinics or neurodevelopmental disorder clinics. And then it's become pathways which are very rigid and very focused on only one condition, which means inexperienced staff who work in those pathways. Actually, that's all they see. And if you have a hammer, everything can look like a nail. Maybe you miss some of the comorbidities, but it also means for families that they're allocated to a pathway, often on the basis of quite a brief letter. And they may sit on that pathway waiting for an assessment for a long period of weeks, months, sometimes sadly, years, and then be told, oh, no, you've not got ADHD. You have to go right back to the beginning because we think you've got anxiety or depression, and that's not great for the families. It's not great for the impact on the child, which is ultimately what's important. And I suspect it's not great for the clinicians. So I think we need a hard think about how we organise services so that it's less siloed and less rigid, and there's smoother ability for teams to manage more than one condition or get prompt intervention for another condition. And I think we need to upskill people who are outside specialist mental health services if they're doing ADHD assessments. So we don't get this duplication. We don't get the desperate family that maybe pays a considerable sum of money to get an assessment and then gets sent back to either their GP or specialist CAMHS, who says, we don't believe that. You've got to wait and we'll redo the whole thing. It must feel absolutely bewildering to families who are caught in these loops. So Tamsin, we were talking about innovative solutions such as large language models. The idea would be to direct people to personalise the sources of support. And I was asking in terms of evidence and validation of these models, do you see them as a solution to the problem we were talking about of not existing enough child psychiatrists to meet the demands that we currently have? I think AI is a potentially really exciting tool. I think one of the things it can do for us is put loads of data together that we would take too long to put together. And if that allowed children and families to access solutions that maybe they could try themselves before going anywhere more formal, I think that's a really exciting prospect. And I guess the next level up is the feedback the family gets is either, oh, it looks like you could do with some help. Here are some sources of information or these are the people you can go to see, or even here are the-- here is the referral already made. All of that's really exciting. And I think we shouldn't forget that we haven't called it AI, but actually we've had bits of this information available to us for quite a long time. So based on the first national survey and the standardised diagnostic assessment that we used, Robert Goodman and his colleague developed an algorithm that would count symptoms that were just very structured yes, no definite answers that the AI or the computer algorithm could count in those days to highlight where a child might be having difficulties and to put together the reports from parents, young people, and from schools. What the computer couldn't do in those days is assess the more qualitative information that this assessment also gathers. But it would grade the kind of difficulties into no difficulties at all in any domain. And the clinical raters working in the national survey would just reap the qualitative information to check for safeguarding issues or anything that had been overlooked. The next level up pulled out police contact. And because that might be an indication of a safeguarding problem or fears because they're quite tricky to systematise. For example, if you are really phobic of bees and wasps, you might live every summer in absolute terror and be almost housebound for fear of going out and encountering a wasp, but keeping all your windows closed so they can't get in. However, bees and wasps are only around for a very short time in this country and you might not fulfil the time criteria. So in other words, the computer counting symptoms might not think this is a real problem and that takes a real judgement call. So that was another thing that we would make sure that we read and we thought about. And then the computer would give two other types of level 1. Well, there's quite a lot of symptoms and quite a lot of impairment for the clinicians to think about. And one way the computer said, I think this child meets diagnostic criteria. And counting symptoms, it probably did more reliably than the raters, the clinicians who looked at it. However, the clinicians can judge conflicting reports. They can make links across the assessment. For example, we've talked about how difficulty concentrating can be a mark of depression, can be a marker of severe anxiety, as well as ADHD and other conditions. And a clinician could put all that together. So I don't think that clinicians are going to be replaced anytime soon, although I'm also aware of really interesting work in the computing department with Hatice Gunes, who is actually building social robots and exploring social robots save clinical time by them asking children questionnaire questions, interviewing to get that assessment. And the preliminary data says the children find it very acceptable. Interestingly, you get slightly different responses than you do from asking parents or asking children on paper and pen. And obviously, there are safeguarding issues, et cetera, but it's acceptable to the children who chat away and they're aware that they're talking to a robot. Again, I don't think we'll be replacing robots as clinicians anytime soon. And like all tools not used wisely, there's a capacity to do harm. So I think we need to really think about it. But I think certainly in the identification of young people who need help and directing them towards sources of support that I think holds great promise, but we need to make sure that the support is there for them when we get there. So on that note and that cheerful note, because it's really good to know that you don't think AI is going to replace completely clinicians, as I would hope to keep working in child psychiatry in six years time. But on that note, as a final question, so we spoke a lot about the increase in the prevalence of mental health conditions, namely depression and anxiety. And that we actually do not have an overdiagnosis problem, but we do have an increase in the demand on our services, and we need to make sure that they are resilient and capable to deal with that demand. So speaking of resilience of services, AI as we've been talking about, could be a potential tool in aiding us on the identification of mental health conditions. But how to make sure as a final kind of question that our services are also offering equitable care. I oftentimes think a lot about the biases within AI. So for example, we talked about ADHD and autism. And we know that the most treated and diagnosed group are boys-- school-aged boys. And we know that, we all as clinicians as well, have biases. And I suppose what I'm asking is how to make sure that we are reaching, like every young person, the most equitable way possible. Good question. And something that many clinicians hold close to their hearts. And it's a founding principle of the NHS is equal access free at the point of delivery for all who need it at the time that they need it. And I think at the moment, we're not there. And as the need to go and get a private assessment in order to get a timely diagnosis and perhaps get some treatment becomes more pressing. What we're having is potentially a two-tier service where children living in families, where parents can advocate for them, or where they have the financial resources to seek private assessment and sometimes treatment, get a service more quickly than those who don't. And of course, our prevalence studies show that the level of need is concentrated in the groups who have the less lower financial resources. So that's inequitable. This is where the systematic analysis of data to highlight children with problems, provided we use that data mindfully. And we think about where there might be biases and we keep reviewing. That's where I think it's hopeful because you're systematically looking at everyone, not just those who are seeking help and nowhere to go and know how to access it, and have the resources to take the time off work to get the child to the appointment, et cetera. So I think that's a positive, but none of our data sources, in terms of administrative records or even our assessments are without bias. So we are testing out a theory that, for example, the development and well-being assessment, the doorbell, which I just spoke about in relation to the national survey might underrecognized girls with ADHD if you have the screening barrier on. So at the moment, there are a few key questions. And if the answer to those is not yes to all of them, you don't go through the session. And it just seems from those who use it in clinical service that if you ask the entire schedule, regardless of everyone, you pick up more girls. So we have a clinic that we're working with that's specialists specialising in neurodevelopmental conditions where they use the doorbell systematically and we're collecting the entire doorbell. And then we're going to put the screening items on and see the screening algorithm on, and see whether that makes a difference to who actually gets the diagnosis at the end in terms of what the computer thinks. And I think if we are going to use these measures the same way, you would keep an eye at blood tests or scans, or any other tests you were using. You'd be alert to hang on a second. Has something gone a bit wrong here. You have to keep your critical faculties about this tool that I'm using does not seem to match my reality. And if we spot those disparities, we need to research them, evaluate them, and if we need to remediate them by changing the algorithm in some way, shape, or form. And that's one of my anxieties about AI use that as with many things, it's running ahead of the evidence behind it and may just become sort of buried in practise in a way that's very kind of not thought through, and we miss the opportunity to keep refining it because these tools and they are going to keep-- need to be repeatedly refined. Thank you so much, Tamsin, for this thoughtful and reflective answer. And thank you so much for talking with us today. And yeah, I think you gave lots of food for thought to our listeners. And if you want to know more about the work Tamsin and her colleagues are doing in the ChARM team, there will be a link to the ChARM LinkedIn web page, which you can access and see the projects that her group is working at the moment. And thank you so much for listening today, and I look forward to speaking with our next guest. Bye. [MUSIC PLAYING]