Dr. Umar Toseeb Hello, welcome to the Papers Podcast series for the Association for Child and Adolescent Mental Health, or ACAMH for short. I’m Umar Toseeb, Professor of Psychology. My research focuses on special educational needs and mental health in childhood and adolescence. In this series, we speak to authors of papers published in one of ACAMH’s three journals. These are the Journal of Child Psychology and Psychiatry, commonly known as JCPP, the Child and Adolescent Mental Health, known as CAMH, and JCPP Advances. All listeners to this and indeed, any of ACAMH’s podcasts, are eligible for a free CPD certificate. Do please visit acamhlearn.org for details of this, together with information on how you can access hundreds of hours of free talks, lectures, interviews, all of which you can also get free CBD certificates for. The web address is acamhlearn.org. That’s a-c-a-m-h-l-e-a-r-n.org. If you’re a fan of our Papers Podcast series, please subscribe on your preferred streaming platform, let us know how we did, with a rating or review, and do share with your friends and colleagues. Today, I’m talking to Dr. Tong Chen, a Postdoctoral Scholar at the Department of Psychology at the Pennsylvania State University in the US. Tong is the Lead Author of the Paper, “Developmental Trajectories of Child and Adolescent Emotional Problems Associations with Early Adult Alcohol Use Behaviors,” published in the JCPP. This paper will be the focus of today’s podcast. So, thank you so much for joining me, Tong. Can you just start with an introduction about who you are and what you do? Dr. Tong Chen I’m Tong Chen and I’m currently at the Pennsylvania State University, at the Department of Psychology. I also obtained my PhD there last summer. I mainly work with internalising problems or emotional problems, such as depression and anxiety, as well as alcohol use behaviours. And I focus on the developmental periods of adolescence and early adulthood for those behaviours, but I also look at how does, say, other factors earlier in life, for example, starting from prenatal or childhood, how do they relate to those behaviours later in life? I am interested in looking at genetic and environmental influences on those behaviours. Dr. Umar Toseeb Thank you, and let’s turn to the paper. Can you give us a brief overview? Dr. Tong Chen In this paper, we looked at whether emotional problems during childhood and adolescence are associated with the alcohol use behaviours in early adulthood which we measured at age 22. And in this paper, we really want to focus on the development of emotional problems across childhood to adolescence, then look at how does – the change of the emotional problems and the timing of the emotional problems might impact on later alcohol use in adulthood. And I want – just want to mention that when we’re looking at the emotional problems, we separated the effect of conduct problems on alcohol use, because that is a robust predictor for more alcohol use. And finally, we used a twin design to look at the genetic and environmental contributions to the associations of these behaviours. Dr. Umar Toseeb Okay, and let’s unpack that a bit now. Why is it important, from a theoretical research point of view, to investigate the relationship between developmental trajectories of emotional problems in childhood and adolescence and then, subsequent alcohol use behaviours in adulthood? Dr. Tong Chen So, I think the first thing is why is it important to look at emotional problems? Because there has been a lot of literature about how does, for example, conduct problems predict alcohol use? And we’ve mentioned a little bit before that they have been robustly linked with more alcohol use, but for emotional problems, which is another very prevalent behavioural problems, the association with alcohol use is actually very mixed. Like, some studies find they may be linked with more alcohol use, but some find they may be linked with less alcohol use or not related with alcohol use. And I want to mention that I’m specifically talking about the longitudinal associations, like when we’re looking at emotional problems in childhood or adolescence and we are trying to predict alcohol use in adulthood, the association is just very complicated. So, that’s what made me interested in looking at emotional problems specifically and I saw there is a gap in the literature which – for me to – that looking at the developmental trajectories of emotional problems. Because I think one possibility for the mixed finding could be that emotional problems at different timepoints could have a different impact on alcohol use and we know that emotional problems do show changes across childhood and adolescence and different people may follow different trajectories. So, that’s why we explicitly model the developmental trajectories of emotional problems and then, we measured them separately, childhood and adolescence, to look at the association with early adult alcohol use behaviours. And we did find some interesting patterns in regard to developmental change and timing. So, I think it’s important to take a deeper look of all the different aspects of emotional problems and the different characteristics along the developmental trajectory, to have a more nuanced understanding of their association with alcohol use. Dr. Umar Toseeb So, you’ve talked a bit about this, but why would emotional problems in childhood be associated with alcohol use in adulthood differently to emotional problems in adolescence being associated with alcohol use in adulthood? Dr. Tong Chen The first, I guess, straightforward answer to that is because adolescents are just closer to early adulthood. So, I think there’s also a reason to expect that the emotional problems during childhood may also influence alcohol use in adulthood and may occur through different mechanisms as compared to emotional problems during adolescence. One thing that is interesting and important in the alcohol use and substance use literature is that although we know that they mostly starts to happen in adolescence and then, they peak in, for example, early adulthood, it just does not happens all of a sudden in those developmental periods. There’s a lot of literature suggesting that a lot of things happening in childhood might impact later alcohol use or substance use. For example, the family processes, the parenting behaviours, the peer group characteristics, the neighbourhood you live in, all these things happening in childhood could have a long-term impact. And one way that could happen is through what we call the cascading effects, so these earlier developmental process, they then might lead to later developmental process and then, they ultimately contribute to adulthood outcomes. So, it’s like snowballing, so one thing leads to another. Dr. Umar Toseeb Could it also be that the profile of emotional problems in childhood is different to the profile of emotional problems in adolescence? And what I mean by that is maybe emotional problems in childhood are caused by different things compared to emotional problems in adolescence and that might mean that those emotional problems look different, which then might mean they have different impacts upon subsequent outcomes. Dr. Tong Chen Yeah, I think that’s a really good point. So, for example, there are studies looking at the genetic and environmental contributions to emotional problems across childhood and adolescence. And Researchers did find that there are new genetic and environmental influences that contribute to, for example, emotional problems in adolescence, that were not contributing to childhood emotional problems. And also, some of the influences on childhood problems, then they do not seem to be influencing later emotional problems. So, I think that’s a great point. Like, they could be influenced by different things and those different things could have a different impact on alcohol use. Dr. Umar Toseeb Let’s move onto the design. You used the Twins Early Development Study, TEDS. Can you tell us a bit about that dataset? Dr. Tong Chen It is a longitudinal twin dataset based in the UK. So, it recruited the twins born in England and Wales from 1994 to 1996 and they just followed them since then. I think the study is still ongoing and the twins completed their 26-years-old assessment with a lot of measures on mental health, cognitive abilities, family processes. So, it’s a very rich longitudinal twin dataset. Dr. Umar Toseeb And one of the great things about the dataset is that it’s twins. Can you tell us a bit about the twin design and what its benefits are? Dr. Tong Chen The twin design includes identical twins and fraternal twins and because identical twins share 100% of their genes and fraternal twins share, on average, 50% of their genes, just like any other siblings, due to these differences in the genetic similarity, we could then infer on genetic influences on certain behaviours. For example, if we find identical twins are more correlated on emotional problems than fraternal twins, we can infer there are genetic influences on emotional problems. And because we’re able to examine that contribution of genetic influences, we then are also able to examine the contribution of environmental influences. And there are some environment that the twins share and there are some environment that the twins don’t share, so we’re able to separate those environmental influences, as well. Dr. Umar Toseeb And those will become more important when we go on to talk about the results. Hopefully, we’ll unpack that a bit more. But before we get there, can you just tell us about the measures that you use to measure emotional difficulties, conduct problems and also, the alcohol use measures in adulthood? Dr. Tong Chen For emotional problems and conduct problems which we controlled for, we used the SDQ measure, the Strength and Difficulties Questionnaire. So, a very widely used measure for child and adolescent behaviours. We used the parent report on SDQ from childhood and then we used the self-report on SDQ in adolescence. And then, for alcohol use, we used the AUDIT scale, which is a very brief, but very powerful, screening questionnaire that you can use for alcohol use problems, but we used it to measure two aspects of alcohol use. So, one aspect is alcohol consumption that just captures how much you drink, how often you drink, things like that. And the other part of the AUDIT scale, it measures alcohol-related problems. So, it’s measuring whether drinking is causing any problems to your life, whether it’s impairing your functioning, like in work, in school, like, these kind of things. Dr. Umar Toseeb Excellent, thank you. So, moving onto the analysis, I think you’ve done two very exciting pieces of analysis. You’ve done the piecewise latent growth curve models and also the twin modelling. Let’s unpack those two separately. So, firstly, can you tell us about piecewise latent growth curve models, what do they do and what do they try and explain? Dr. Tong Chen In general, the latent growth curve models models the developmental trajectory of any behaviour. So, it estimate an intercept, which is where you start. For example, if the measure starts at age four in our case, it captures your levels of emotional problems at age four. And then, it estimate the slope, which is capturing how fast it change, like, from, for example, from age four to age nine, in our case. And in the traditional latent growth curve model, you often estimate, for example, another slope, which is the quadratic slope. That’s, kind of, if people don’t follow a simple linear change pattern, they might increase scores and then decrease. And you just use that quadratic slope to modify the linear shape of that trajectory. But the reason we chose to use the piecewise model is because in many cases, the estimation of the quadratic slope is very difficult to interpret, especially if you want to use that to predict something else, because it’s telling you that the trajectory is not linear. It’s a modification to the linear shape, but the quadratic term itself is very difficult to map to, like, a practical interpretation. So, we chose the piecewise model so that we can allow a similar non-linear shape, where we can have, for example, a linear trajectory in childhood and then we can have another linear trajectory in adolescence. So, looking together, childhood across adolescence, the trajectory is not linear, so it’s, kind of, doing the same thing as the quadratic slope. But then, because we separated into childhood and adolescence, that each have a linear shape, and we could have easier interpretation on how does the trajectory look and how does that impact later alcohol use? So, for example, we have an intercept and a slope for childhood emotional problems and then we have intercept and slope for adolescent emotional problems and then, we can know how do all these trajectories predict alcohol use? Like, where you start in childhood, where you start in adolescence and how you change in childhood, how you change in adolescence, how do they impact alcohol use? So, I think it just makes sense to me, as a Developmental Psychologist, to look at it in this way that allow me to look at developmental change and timing effects in a more concise way. Dr. Umar Toseeb Thank you, that’s very helpful, because I think in the past, I’ve only ever used latent growth curve models and then, if I was interested in things at multiple timepoints, I would always just include the quadratic function. But I can see why this piecewise approach is different and needed in some situations, so I will look into it next time I need to do that. Thank you. And then the next part of the analysis was taking advantage of the twin design. My understanding of the added value of twin design is that it allows you to understand whether the relationship between two variables continues to exist after you’ve controlled for the shared genetic and environmental effects. So, in this case, is it, like, are emotional problems and alcohol use correlated because they are related to each other directly or because they are both influenced by third and fourth unmeasured factors, which in this case, might be genetic and environmental influences? Is that correct, is that the added value of a twin design? Dr. Tong Chen Yeah, I think, and especially to this paper specifically, that’s exactly what we did. Because using twin design, we could differentiate the genetic and – versus environmental contributions, to, for example, emotional problems. And then, because we have this added information, we can test whether the same set of genetic and environmental factors that influence emotional problems are also influencing alcohol use. And I think this is often under-looked possibility in studies without genetic information, I think because there’s really no information to actually test this possibility, as opposed to the two behaviours are directly related to each other. So, we were able to compare the two models with different assumptions. The first model assumes that emotional problems is directly linked with alcohol use and the other model assumes that there is genetic and environmental factors influencing both emotional problems and alcohol use. And we cannot, like, say 100% that they’re directly influencing each other because we don’t have a straight causal design, or, like, randomised control design. But we are more confident that their associations is not confounded by sheer genetic or environmental factors if we do see the model is preferring that they’re directly linked to each other. Dr. Umar Toseeb Yeah, excellent, and let’s turn to the results. What are the key findings you’d like to highlight? Dr. Tong Chen First of all, we controlled for the effect of conduct problems on alcohol use. So, all the findings I’m talking about in terms of emotional problems is after we controlled for that part. First of all, we found that higher levels of emotional problems at age nine is associated with less alcohol consumption at age 22 and that’s only in males. The other finding is that the increase of emotional problems during childhood is associated with less alcohol-related problems at age 22. So, these are, kind of, our two main findings. Dr. Umar Toseeb And so, when I was reading this, I was trying to use my understanding of latent growth models to try and understand what you’d written. So, is what I’ve read and understood correct, so, kind of, hopefully, it’ll help some of our listeners? So, you said in the paper, “The childhood linear slope of emotional problems was negatively associated with the AUDIT-Total,” which was the alcohol use. So, is that the finding that you’re referring to, which is increases in emotional problems during childhood are associated with fewer alcohol-related problems? Dr. Tong Chen Yeah, hmmm, that’s right. Dr. Umar Toseeb Okay, so yeah, the faster you increase in emotional problems, the less alcohol you drink at the age of 22? Dr. Tong Chen The less alcohol-related problems you have. Dr. Umar Toseeb Oh, related problems… Dr. Tong Chen Yeah. Dr. Umar Toseeb …you have at the age of 22. Okay, and what about in adolescence, what were the findings in adolescence? Dr. Tong Chen To summarise, we did not see much of a change of emotional problems during adolescence to be associated with either alcohol consumption or alcohol-related problems. But the finding I mentioned about emotional problems at age nine, that’s, kind of, our start point for the trajectory of emotional problems during adolescence. That has something to do with, like, the entire waves we have. So, we, kind of, made the cutoff at age nine. So, age nine is technically not adolescence yet, but it’s at a transition to adolescence where a lot of physiological, biological, or perhaps social changes that are already starting to happen as children transition into adolescence. So, I think the fact that we’ve found that only the emotional problems at age nine and not a change thereafter is associated with alcohol use, is probably telling us something about the role of emotional problems on alcohol use is probably established before most adolescents start to drink. Because the alcohol initiation usually peaks at mid-adolescence. So, my understanding of that is these children with high levels of emotional problems at the transition to adolescence, they may engage in less social activities with peers, or they may hang out with peers who share the same tendencies and you just keep selecting peers like that, or you keep hanging out with the same group, and you maintain that tendency throughout adolescence. So, then when people starts to drink, like, for example, in mid-adolescence, you just are still involved in less social activities like that, you have less access to alcohol and you start drinking later. And that’s, like, when the cascading fact we described earlier, all these step-by-step and then, leads to less alcohol consumption at age 22. I don’t think that means, like, emotional problems in adolescence has no impact on alcohol use, but it means that the process might start before, like, adolescents start to drink. Dr. Umar Toseeb And then, when you fitted the twin models, what was the added findings when you applied the twin models to your data? Dr. Tong Chen We did find that emotional problems and alcohol use behaviours, they are directly linked to each other. They are not explained by sheer genetic or environmental influences that are unmeasured. But one interesting finding is that we found that some associations between the conduct problems that we controlled for and alcohol use, they’re not directly linked to each other. Some of the associations were explained entirely by shared genes. So, when we adjust for those effect, we actually see that emotional problems in childhood, the increase of that is no longer associated with alcohol-related problems, but it did affect our finding on the alcohol consumption finding that we just talk about. Like, at age nine, emotional problems are associated with less alcohol consumption. So, it’s, kind of, complicated, but it’s also interesting that one of our findings changed because we adjust the way that we model conduct problems. So, I think that means that even if conduct problems is not something of our main interest, that we want to think carefully about how we control for that effect. Because before, we’re assuming that conduct problems is directly influencing alcohol use, but since the twin findings study, as some of the associations are confounded by shared genes, that means they are not directly influencing alcohol use. So, it’s like we shouldn’t really control for the effect of conduct problem in this case, because it’s not directly influencing alcohol use. And that’s why it impact our findings on emotional problems. Dr. Umar Toseeb So, if I’ve understood this right, I think when we were talking about it earlier on, I think you said that potentially, the causal effect of emotional problems on subsequent alcohol use was via conduct problems, potentially. And then I think what – I think that’s what you said, and then, what we’re saying here is the twin models show that the reason why conduct problems is associated with alcohol use is partly because they have shared genetic effects. So, the causal pathway between emotional problems and alcohol use might not be via conduct problems, is that right? Dr. Tong Chen Yeah, yeah, because we see that conduct problems effect on alcohol use may not actually be there, which is due to shared genes, so it’s different from what we has assumed, that conduct problem leads to more alcohol use. So, in this case, we were artificially controlling for the co-occurrence between emotional problems and conduct problems. So, the way I think of it is emotional problems may actually have some aspect that leads to more alcohol use, but because of the fact it’s associated with more conduct problems and we controlled for the effect of conduct problems, we’re taking out that part of the effect of emotional problems and we were, sort of, wrongly attribute to the effect of conduct. We were wrongly assume that this is occurring because of conduct problems, but in fact, it may not be. It’s be – it’s due to emotional problems. That’s my understanding, but… Dr. Umar Toseeb Excellent, and then, can you tell us a bit about the gender differences that you’ve found? Dr. Tong Chen Yeah. So, we found that emotional problems at age nine was associated with less alcohol use at age 22 and that’s only in males, not in females. I think these differences was not necessarily surprising, but certainly adds something new to the literature, because previous studies reported mixed findings regarding whether there is sex differences. I think my understanding of these sex differences is linked with our interpretation earlier about emotional problems might reduce social activities that involve drinking. There is literature suggests that compared to females, adolescent males may be more susceptible to peer pressure on risk-taking behaviours, such as alcohol use or other substance use. And in this case, because emotional problems might be linked with reduced social opportunities of drinking with peers, so that it might have a particularly salient effect on reducing alcohol use in males because they’re more prone to peer pressure on drinking if they are involved in those social activities. Dr. Umar Toseeb Given your findings, what might be the potential clinical implications? Dr. Tong Chen We’ve thought a lot about that because we, kind of, hypothesised through the literature that emotional problems may, to some extent, increase alcohol use due to that people may use alcohol to cope with emotional problems, but we did not see that in our main findings. I think first of all, I want to make it very clear that our findings should not suggest that emotional problems are necessarily protective factors for alcohol use, for the exact reason we talked about, that our findings are what we found after controlling for conduct problems. There might be children or adolescents who show both high levels of emotional problems and conduct problems and that in this case, they might show more alcohol use or more alcohol-related problems in early adulthood. So, the control of conduct problems is purely, I think, for a research purpose. We want to understand the independent effect of emotional problems so we can understand more specific mechanisms as associated with the emotional problems. But in clinical implications, you don’t, like, treat someone just for their emotional problems and ignore other things. So, I think that’s a important distinction here. And the other thing I want to mention is that in adulthood, emotional problems might still be associated with more alcohol use and that’s a robust finding in the literature. Our findings suggest that the emotional problems and alcohol use association is very age-specific and very dynamic across the developmental period. So, I think maybe that is something to pay attention to when we’re talking about, like, preventing alcohol related problems. We might think less about those indicators, like emotional problems or conduct problems, but think more about what’s driving the drinking behaviour at that specific developmental stage, and think about why are those behavioural problems associated with drinking? Through what mechanism? I think the mechanism, like, the motivation for drinking, things like that, is what we could focus on, instead of having, like, a simple, like, cutoff or indicator of risk. Dr. Umar Toseeb And where next with this? Do you have research in the pipeline that you’d like to share with us? Dr. Tong Chen One follow-up that could be interesting to look at is to measure all these drinking motivations that we talked before and to look at whether they explain the effect of emotional problems on alcohol use. And the other thing I have in the pipeline is that I further differentiate different types of emotional problems. So, I looked at whether depressive symptoms and anxiety symptoms have different impact on alcohol use behaviours. The short answer is they do, but mostly they don’t. So, yeah, that’s something that’s in the pipeline. Dr. Umar Toseeb Thank you for that fantastic wide-ranging discussion. Just to wrap up, what’s your take home message for our listeners? Dr. Tong Chen Our findings shows that emotional problems in childhood and adolescence are not necessarily associated with more alcohol use or misuse in early adulthood, especially when it does not co-occur with high levels of conduct problems. And the associations between emotional problems and alcohol use show very interesting developmental patterns. So, I think there needs to be more research that pay close attention to the change of their associations across development and how do the mechanisms explaining their associations may change with development. Dr. Umar Toseeb Thank you so much. For more details on the paper and Dr. Tong Chen, please visit the ACAMH website, www.acamh.org, and Twitter @ACAMH. ACAMH is spelt 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 your friends and colleagues.