We are the Association for Child and Adolescent Mental Health, or ACAMH for short. And this is ACAMH Learn. Welcome to Mind the Kids podcast from the Association for Child and Adolescent Mental Health, known as ACAMH. What happens to a young person's mental health and behaviour when violence isn't the exception in their daily life, but the backdrop to it? Hi. I'm Mark Tebbs. I'm your host for today. I spent my whole career working in mental health, from frontline service delivery to director of mental health commissioning. I'm currently a chief executive of charity, and I'm a trustee of a mental health organisation and also a career coach. I'm really pleased to be hosting these podcasts. It gives me the opportunity to speak to some of the leading clinicians and academics in the forefront of children and young people's mental health. So if you're a clinician, if you're an educator, a parent, or just curious about mental health, I hope these discussions leave you with some fresh insights and practical tools to help. Don't forget to subscribe where you get your podcasts from. And also, on your YouTube channel, just search ACAMH, and you'll find us across all socials. Today, I'm speaking to Susan Swingler, a researcher at the University of Oxford who has completed her PhD looking at how young people respond to chronic stress in their environment, including violence exposure, and how that shapes their mental health and behavioural development. Susan's work draws on the remarkable Thula Sana cohort study following young people in Khayelitsha, a large community in the outskirts of Cape Town from birth through to late adolescence, spanning nearly two decades of data collection. We'll see what everyday life looks like for the adolescents growing up, what the study found about the relationship between that exposure and then conduct problems, like aggression and rule-breaking, and whether that relationship works both ways. We'll explore some intriguing possible differences between how boys and girls respond and what all that means for designing effective, contextually-grounded violence prevention. We'll challenge some common assumptions about behaviour risk and resilience. So, Susan, welcome to the podcast. Really lovely to be speaking to you today. Great to be speaking to you. And thanks for having me. Great. We'll start in the normal way. So it'd be great if you could introduce yourself, maybe say a little bit about your research interests, where you're based. And if you worked with anybody on the paper, then it's kind of a good opportunity to give them a little bit of a name check. Great. So I'm Susan Swingler. I'm a researcher at the University of Oxford, and I've very recently finished a PhD in social intervention and policy evaluation. My research focus has been on how young people respond to chronic stress in their environments, including violence exposure as a chronic stressor, and how this impacts their mental health and behavioural development. I'm interested in specifically looking at low and middle-income country contexts because this is where the research has not always been well represented in the wider literature and where we know the majority of the world's young people live. So this research in the South African context with the Thula Sana cohort was very much a collaborative effort. Jane Barlow, who's also at the University of Oxford, and Mark Tomlinson at the University of Stellenbosch, they guided the work throughout from every stage. Qing Han from Peking University and Alexis Cullen, who's at King's College and Karolinska Institute, they contributed their expertise in research design and our analytic approach. And then, really, the three people who brought their deep knowledge of the South African context and through their work with the Thula Sana cohort itself were Stefani du Toit, Sarah Skeen, and Mark Tomlinson at Stellenbosch University, who were really instrumental in sustaining the cohort over this nearly two decades. So that longterm commitment, combined with the efforts of the research teams on the ground in Khayelitsha and at Stellenbosch and from the families, who kept participating, they really made this research possible. Yeah, amazing. So let's dive in to the study, explain a little bit more about the research. So it'd be great if you could just set some of the scenes or set some of the context. What did everyday adolescence look like for the people in the study? Great. So it's important to talk a little bit about Khayelitsha, where the study took place. It's a large isiXhosa-speaking community, which is on the outskirts of Cape Town in South Africa. And it means "new home" in isiXhosa. And it was established by the apartheid government in the 1980s as part of a policy of racial segregation and forced removals. So that history has very much shaped how young people are able to experience and navigate their space. So many families and young people will face high levels of unemployment, overcrowded housing, and uneven access to basic services. And this is also alongside a significant burden of violent crime and homicide rates among young people that are much larger than the national and, certainly, the international average. So for adolescents growing up there, they have to navigate their space against this backdrop of increased risk to violence. And this may be present across various spaces in their lives. So they might witness or experience violence themselves in their neighbourhoods at school, including corporal punishment or bullying, youth violence, sometimes in their homes, of course, and then just navigating daily public space, really. So I think it's also important to understand that this means that they have very limited control over whether they experience violence or not, which may be different to other relatively less adverse contexts, where it's not easier, but there might be better strategies for avoiding violence or seeking protection from violence. When I describe Khayelitsha [INAUDIBLE], I also don't want to fall in thinking about only deprivation or adversity. It's also a really strong, vibrant community that is resilient, that has strong family, community networks, local grassroots organisations who are creative, and where young people have really high and real aspirations for their lives. So I'll talk about how important it is to frame Khayelitsha not only in terms of deprivation and adversity, but also thinking about Khayelitsha and knowing that Khayelitsha is a resilient, vibrant community that has really strong family and community networks, local grassroots organisations that want to tackle violence and promote resilience among young people, and then also young people who have real aspiration. And like adolescents anywhere are going to school, they're forming relationships, they're playing sports, and they're thinking about their futures, things that adolescents do everywhere. But at the same time, they're having to negotiate unsafe spaces, which really puts them at risk of experiencing violence alongside material hardship. And they may respond in various different ways. Sure. And what would the education system typically or the education experience typically look like? Well, yeah, so the history of Khayelitsha and that kind of structural disadvantage has meant that the quality of education is not where it should be. And, certainly, when adolescents go to school, they may not have schools that have robust violence prevention and response policies. And there's also a history of adults perhaps responding to violence and to conduct problems with violence itself through corporal punishment. That is something that they might experience at school. Yeah. So I think you've really set the scene well, in terms of the environment. And we'll go into more detail on the study in a moment. But could you just unpack what the key headline findings were, particularly this kind of relationship between violence exposure and adolescents' behaviour? OK. So the central finding is that adolescents who were exposed to greater violence between the ages of 12 and 14 were more likely to go on to have conduct problems, like greater aggression and rule-breaking behaviour at the same time-- so also at 12 and 14 years old-- but also four years later, when they were between the ages of 16 and 19. And this association remained consistent when we controlled for broader socioeconomic hardship and deprivation. So we could really say that that violence exposure was a robust predictor of risk factor. And it was so across different analytical approaches and informant sources that we used. Another key finding was we also tested in the other direction. So do conduct problems-- theoretically, we might expect that conduct problems put young people in situations where they might experience more greater violence. But we didn't actually find strong evidence for that pathway or that direction. But it's a relatively small sample, so we can't rule out what we saw was a small and nonsignificant effect there. So the main message for listeners is that violence exposure is a risk factor, and it raises conduct problems at the population level, and that this risk might persist across adolescence as a developmental period. And it really makes adolescence an important window for prevention and response, because we already know that conduct problems typically emerge or resolve or can escalate within this period and that it's quite a malleable, sensitive developmental period where intervention can take place. Yeah. And when we talk about conduct problems, what are we talking about? What would be some of the behaviours that we would associate with conduct problems? How should the listeners think about those kind of behaviours in the context of this kind of environment that you've described? Great. So conduct problems, we would know as a research or clinical term. We might often use it in the context of conduct disorder. But in this study, we didn't think about it in terms of that psychopathology, but more as a number of behaviours that adolescents could be expressing. And these included frequent aggression, fighting. This could be verbal or physical aggression and also rule-breaking, deviant behaviours, like lying, stealing, or damaging property. And the measures that we use combined and looked at both aggression and rule-breaking using the same scales with self-reported measures, so adolescents' own understanding or awareness, but then also caregiver reported to look at the different informants or what they can see is happening. So I think the term itself can make us think that these are processes that are individual and really ascribed to personality, or they're difficult individual behaviours that exist entirely within this young person. But what this research is trying to get at is understanding how these behaviours emerge in contexts. So aggression as we think might actually be a learned response in a context where violence may be modelled. It might also be a way externalising distressing anger or frustration or a strategy that also might be protective in unsafe situations. And, of course, persistent hardship, limited opportunities can make young people express their behaviours in a way that's external. At the same time, we want to think about contexts, but also not to normalise the behaviour or make an excuse for it or think about the effects on other people as unimportant, but really, thinking about, how can effective response look beyond these individual behaviours to think about, what's happening around a young person that is maybe driving this behaviour? What pressures are they facing? And what support or protection might be unavailable and might be driving the marginalisation of this young person displaying these behaviours? So saying that in these environments, where violence is normalised, the reaction can be adaptive, I guess, but not to deny the impact of that, but maybe as an opportunity of intervention and change? Yes, exactly. So moving away from the idea that normalisation for young people-- it doesn't mean that they're unaffected by violence or that violence is regarded as acceptable, but it might mean that repeated exposure to violence can make it feel like something that's expected or something that isn't out of the ordinary or shocking. And that's something that has to be negotiated to or adapted to in everyday life. So yeah, thinking more critically about that normalisation language or desensitisation language and thinking that even a small statistical effect in the study, it matters enormously for when exposure is widespread. Like you say, we need to think about interventions that prevent violence while also trying to respond to violence in the behavioural manifestations that it might have while these young people are adapting to their environments. Yeah. And I'd like us to talk a little bit about your methodology as well. You mentioned this longitudinal study and the commitment from researchers over several years. So what did it take to sustain that kind of approach in this environment over such a long period of time? Yeah. So it's really quite a massive effort that I, as a researcher, came on after the data was collected. So, really, it's good to take time and reflect on what effort this took. So the Thula Sana cohort began between 1999 and 2003, over 20 years ago, where 449 women were recruited during pregnancy into this Thula Sana cohort. And Thula Sana means "hush, baby" in isiXhosa. So from pregnancy through various phases of all time points in infancy, they were engaged in intervention and data collection. And then there was quite a gap. And then they were recontacted between the ages of 12 and 14 for an assessment, follow-Up assessment, and then again at in late adolescence between 16 and 19. So, yes, it's a low-resource, highly-mobile community. So a major challenge came at that early adolescent, mid-adolescent period, 12 to 14, which was, over a decade later. And researchers had to come in and look at the records and find that the addresses were outdated. Families that lived in informal communities, their street names may have changed during that time, or they actually might have moved. So the team at Stellenbosch and at the research centre in Khayelitsha had to work with local community networks to trace participants. And some and many of them were still in Khayelitsha, but other participants were eventually located across three different provinces in South Africa, which included 25 different towns. So those who were in Khayelitsha and who were able to travel came to the research centre in Khayelitsha. But for those who wouldn't, the researchers actually made the effort to go and conduct home visits. And all of the measures were adapted and to isiXhosa. And the data collectors or field workers were trained and were isiXhosa-speaking research workers. So that meant that by the time I came across the data and worked with the research teams, around 70% of the original cohort were retained in the study, which is a really remarkable achievement, in my opinion, in this context of socioeconomic adversity. And so I think it goes beyond good organisational skills and really is a community effort requiring a lot of local knowledge and trust and time from the families and young people who participated and shared these difficult experiences that they were having over many years. So, yes, many thanks to them and all the research teams involved. Yeah, quite an achievement. Was there anything in that follow-up rate that you think impacted on the results? Was there any sort of sampling biases detected in addressing those challenges and the longterm follow-up? So we did do analyses of what could be driving missingness in the studies. And there were a few things, so baseline characteristics, like socioeconomic status, maternal age, and also baseline levels-- at the early adolescent period, baseline levels of conduct problems, which could be driving missingness. So we really wanted to address that. In our analytic methods, we used approaches to dealing with missing data, like multiple imputation, using all the data key variables that we did know about those children who were missing. And then we also have to say, this is an observational study. And there are issues then with generalizability to the cohort as well as the community at large, so just thinking about that, in terms of a caveat to our findings. Yeah, brilliant. Thank you. And you've mentioned the association between violence exposure and conduct problems. And then you briefly mention that the reverse pathway association wasn't found there. Could you just explain that a little bit more, just so we fully understand that bit? Yes, of course. This is a really interesting question that we had about this reciprocal relationship between violence and conduct. Violence can drive conduct problems. And then conduct problems might mean that adolescents find themselves in situations where they experience more violence. They might receive more punitive responses. They might be kind of marginalised, in terms of more pro-social peer groups. Or family members, say, might distance themselves. So they might not have access to protection, and they might experience more violence or be among peers who are more violent. And there's evidence from relatively less adverse or higher-income country contexts that clearly says there is this reciprocal relationship, to the extent that sometimes conduct problems is more a driver of violence exposure than the other way around. But in our study, we found that it was not quite the case where violence exposure is driving conduct problems. But then that reverse pathway, the effect size was a lot smaller, and it wasn't statistically significant. Important caveat is that our sample size was relatively small and not well powered, with this being an exploratory question that we had. So we can't rule out that individual behaviour is not driving exposure. But theoretically and thinking about context, this could support a theory that, in this context, individual behaviour isn't as much of a driver of exposure, because adolescents don't have that much autonomy in whether they witness or experience violence in their homes or community because of the pervasive nature of violence in Khayelitsha. It's important to both say we can't draw firm conclusions, but we stick by our analysis and say that we can't support this pathway, but also encouraging future research with large samples to ask this question. Yeah. And did that surprise you? Was that what you expected or hypothesised that you would find? We didn't make hypotheses regarding that because we thought it was exploratory. The main question was regarding violence exposure and thinking more broadly about violence prevention interventions. But then in the violence field, there are a lot of reciprocal relationships. So we certainly wanted to think about controlling for that reciprocal pathway in trying to establish violence as a prospective risk factor. So I, as a researcher, didn't make any hypotheses but wanted to think about this exploratory finding and what it could mean for future research. Yeah. And I think you looked at the possible sex differences as well. What did you find in that area? So, similarly, this was an exploratory analysis, where we did models by gender as subgroups, but we also did formal interaction analyses. And that was our primary method. We did not find significant differences between boys and girls, in terms of violence to conduct, future conduct, pathway and then the reverse pathway. But at the same time, there's that issue with sample size, that we've essentially reduced the sample size by half. So we need to be careful there with saying whether we have enough power to look at these gender differences. But we wanted to discuss what we found in these subgroup analysis to generate hypotheses that again could be tested in larger samples. I won't say that there are differences between boys and girls, but when we look at each model separately, we can come up with some further questions. So among boys, looked like their cumulative contextual socioeconomic adversity was a stronger predictor of their conduct problems at age 12 to 14. So that's that cross-sectional finding. And yeah, stronger than violence exposure-- so that makes us think, are boys responding not to the violence that they are witnessing or experiencing at that time, but are they responding to and perhaps externalising their distress to their socioeconomic circumstances? That could include anger or frustration that is expressed as aggression or, equally, rule-breaking or lying or stealing that comes from things like caregiver unemployment or food insecurity. So that's a plausible kind of hypothesis we might like to test. But at the same time, their violence exposure was a longterm predictor. So there might be a longer-term learning or a process that might be taking place for boys, where they're still at risk for conduct due to their violence exposure. For girls, violence was relatively strongly associated with conduct problems. And we didn't see that same link between cumulative contextual adversity and conduct problems in early adolescence. And also, there was a kind of suggestive pathway among girls, where their early adolescent conduct problems did appear-- or the size of the effect appeared to be-- larger, in terms of that reciprocal association. So their conduct problems may be driving their longer-term exposure to violence. Thinking about the different gender pathways that might be occurring, girls are not always expected to externalise their behaviour through aggression and rule-breaking. We more often expect that they will internalise. So for girls who are aggressive, for girls who may be stealing or lying, they might face kind of this double deviance. Not only are they behaving in this way, but they're girls who behave in this way. So they might have more harsher punitive responses in their relationships, in their peer groups, from teachers. They might face greater marginalisation and higher risk and, ultimately, gender-based violence. Again, saying this is a hypothesis that needs to be tested and it's not a firm conclusion that we can draw here-- and we might need to look at other things going on. Like, restrictive gender norms might be another variable to include in future work. So where would you like the research field to go next, in terms of exploring some of issues and findings that you got from this study? Right. So I think we need to contextualise these findings even more. So in this study, we looked at violence, contextual adversity, and conduct problems. But we might need to look at broader emotion regulation, as well as internalising symptoms and anxiety and depression, to get a fuller picture of how boys and girls are responding to violence in their environment. And in addition to our findings and the questions it asks, this could help inform gender-responsive, contextually-grounded interventions in Khayelitsha and other similar high-adversity contexts. So in my own work, I will be trying to answer these questions, call other researchers to do the same, and then also practitioners and policymakers to think about the broader ecology of young people's lives and how different interventions are needed to address risk at different levels, and perhaps differently for boys and girls. Yeah. And is the research field at a point where it can say what effective, textually-grounded violence prevention looks like? Is it at that point where it can start to develop effective interventions? So I think the evidence is looking to the INSPIRE framework, which is developed and introduced and updated by the World Health Organisation as addressing violence at multiple levels. So contextually-grounded intervention needs to recognise that. So no single programme alone is this magic silver bullet that's going to resolve violence. But we need to think about interventions that not only address individual behaviour and help adolescents to navigate their environments, but also think about interventions at the family level, at the community level, at the greater societal level. So for individuals, this might mean that interventions will be effective if they're trauma responsive in low-resource settings. Group-based cognitive therapy approaches have found evidence of effectiveness, especially in resource-constrained South African and broader sub-Saharan African contexts. At the family level, there's evidence for parenting support being effective in not only reducing violence in the home, but also preventing child conduct problems. School-based interventions, like social-emotional learning, strengthening, violence prevention policies, and reducing corporal punishment, are important. And then thinking about those findings related to contextual adversity or socioeconomic status and the disadvantage there, we need to think about poverty alleviation alongside community safety, while also acknowledging that young people who are at risk for future violence involvement may not respond to punitive or deterrence approaches, this might only serve to marginalise them even further and increase their risk for violence. So thinking about this as a system of violence prevention and response that we can act on at multiple levels and involves efforts from multiple sectors to achieve this. A great summary. So we're coming to the end of the podcast. So it's just an opportunity-- if policymakers, clinicians, teachers, youth workers took away just one message from this conversation, what would it be? Yeah. The key message I would want clinicians, teachers, youth workers to take away is that aggressive or rule-breaking behaviour can't be understood separately from violence or other risk factors that young people experience. That doesn't remove individual responsibility or responding to individual behaviours. But it does change how we think about effective response. As I mentioned, punishment or deterrence is unlikely to address the source of the problem. To prevent conduct problems or to prevent violence itself actually becomes more important. It's important to stress that young people, they need safety. They need to be supported in their relationships. And they need opportunities to develop different ways of handling conflict and difficulty in their lives. And ultimately, adolescence isn't too late. It's actually a key period. A lot of our research might be in early childhood or in childhood before the transition to adolescence. But we need to think about adolescence as the malleable period for change. They're beginning new relationships. They're thinking about their futures. And it's a key period to try and prevent violence and its impact on adolescent mental health and behaviour. And I would also encourage the listeners to go and read the paper, which is free to access on JCPP. Good stuff. And if people wanted to find out a little bit more about your work and where you're based, is there a website that people could follow up with? Yes, absolutely. So if you google the Department of Social Policy and Intervention, which is based at the University of Oxford, you'll be able to find my profile there, along with several of my colleagues. And also, be sure to check out the work of the Institute of Life Course Health Research at the Stellenbosch University. Brilliant. Susan, it's been a pleasure speaking to you on this really important topic. Great. Likewise. And thank you so much for having me. I guess what struck me first was the Thula Sana Sana data and Khayelitsha itself, not just a place of deprivation, but as a genuinely resilient, vibrant community with strong family networks and real aspirations. And then I guess there's the headline findings. Adolescents exposed to more violence between ages 12 and 14 were much more likely to show greater aggression and rule breaking, not just at that time, but also four years later. And that held up, even when you're taking into account socioeconomic hardship. What I found particularly interesting was, the reverse didn't hold nearly as strongly, that conduct problems weren't clearly driving further violence exposure, which really challenges the idea that these young people have much agency in avoiding violence in the first place. I was also struck by the gender differences. Susan described boys' conduct problems seeming more linked to broader socioeconomic adversity, girls more tightly linked to violence exposure itself, with that idea of a kind of double deviance girls might face for behaving this way. Susan was rightly cautious about calling these firm conclusions, but compelling hypothesis for future work, no doubt. I can't talk about this study without mentioning the sheer scale of effort behind it-- a cohort followed over two decades, researchers tracking families across three provinces, 25 towns, and still retaining 70% of participants. That's a remarkable testament to the trust built and the perseverance and dedication of the researchers. Finally, the point that stayed with me most is Susan's reframe around normalisation, that violence becomes expected for these young people doesn't mean they're unaffected by it. It just means it's something they've had to adapt to. And a closing message-- that adolescence isn't too late for intervention. It's actually a key window-- feels a really important and hopeful note to end on. I hope you join me next week as I talk to Asia Murray about her research with ADHD and emotional regulation. [RELAXING MUSIC]