Dr. Jane Gilmore: Welcome to Mind the Kids. I'm Dr. Jane Gilmore, honorary consultant clinical psychologist and child developmental program director at UCL.
Professor Umar Toseeb: I’m Umar Toseeb Professor of psychology with a focus on child and adolescent mental health, and special educational needs.
Dr. Jane Gilmore: In each episode, we select a topic from the mental health literature and in conversation with invited authors, sift through the data, dilemmas and debates to leave you with our takeaways for academics and practitioners. Today, we'll be talking about the relationship between parenting style and infant characteristics. This episode is called Parent plus Child Factors. Does it add up?
Professor Umar Toseeb: So, parenting is a tough one for me because I think that as someone who is not a parent, I feel like I can come at it from a different perspective. And it's very easy to get into this parent blaming trap where you say, well, children's outcomes are dictated by parenting and the parents are solely responsible. But it also feels a bit disingenuous to then not put some emphasis on parenting because in parenting is clearly important.
Dr. Jane Gilmore: I like the way, Umar, you're raising the idea of addressing the parenting issue because it's really, parenting has become a verb only very recently. You we've been doing it a long time, but we've now started to talk about parenting as a verb. And certainly in my clinical experience, there's now much more advice available to parents about best practice in the general population. So in previous times, the clinical population were quite well, sort of address in terms of content, now the general population are interested. So we've got parents actively looking for the science behind parenting advice. So the internet has democratized the availability of scientific data as well as serving up a whole lot of the bad stuff too. So in my experience, parents are sourcing peer reviewed papers or books aimed at professionals. I think that's really interesting. I like the way you're taking that thoughtful approach and, you know, of course, there's lots of very good professionals in this area who don't have children, but quite rightly, I think, there's an increased value in lived experience and mental health. So maybe there's an equivalence in parenting.
Professor Umar Toseeb: I think one of the things that I mean, maybe this is insensitive, I feel like it's very, let's not blame parents, which is great, let's not, but then equally, when there's a positive outcome, then it's easier to ascribe that to good parenting. And that might also be kind of what we're going to be talking about today, whether some children are more susceptible to... thriving in positive environments and then other children might be more susceptible to being disadvantaged in negative environments. But it's all interesting things in terms of where we place the role of parenting within the wider social context of development and children's development.
Dr. Jane Gilmore: And I know you're very excited about the paper because we're discussing big data and you love big data. Can you just say very briefly why you are such a fan of these sorts of data?
Professor Umar Toseeb: I think it gives us the big picture at the population level. So we know we always have this, oh, the sample size isn't big enough. And actually that isn't really an issue for lots of research. For some research, having a sample size of 10, 15, 20 people is actually a good thing. Like, you know, don't need massive sample sizes for lots of lots of areas of research. But if we're trying to understand what phenomena look like at the population level, and what that looks like between populations across different populations, then having these big data sets is really important. And also, you know, in areas like child development, children's mental health, there isn't one single cause of any sort of outcome for children. There are lots of different causes and having such big data sets allows you to control for those different effects include various different predictors of children's outcomes within the same statistical model. So then you can really pin down. Well, how much of this effect is being explained by parenting or whatever other factor you're interested in. So that's why. And also I just like numbers. I could spend all day in front of a data set and that will be the best day of the week.
Dr. Jane Gilmore: Well, I've got a lot of questions for Robert and one of them, which perhaps we won't waste valuable time on, but given there was 30,000 participants in the study, I want to know how big the initial spreadsheet was. Anyway, we'll tackle much more interesting questions with Robert in a moment.
Professor Umar Toseeb: Okay, let's introduce Robert. So let's do it. Let's bring him in. Today we're joined by Dr. Robert Eves who's a postdoctoral researcher at the Department of Psychology in Bielefeld University in Germany. Robert is the lead author of the paper, Interactions Between Infant Characteristics and Parenting Factors Rarely Replicate Across Cohorts and Developmental Domains, which is published in the JCPP. Welcome, Robert.
Dr. Robert Eves: Thank you very much for having me.
Professor Umar Toseeb: And actually, I'm going to go back there and be like, can I call you Rob?
Dr. Robert Eves: Of you can.
Professor Umar Toseeb: Yes, let's do it. So what was your motivation for writing this paper?
Dr. Robert Eves: So I think you briefly spoke about some of the ideas that are also important to me in the introduction. Generally when we're interested in developmental psychology, we quite often use somewhat small sample sizes, but which most importantly are not nationally representative. So we don't have a good distribution relative to the actual population of the UK or Germany or whichever country that we're interested in. And this is probably quite problematic, because it results in us not having that much confidence in our findings and whether they are true for everyone in the population. So I'd kind of read some of the past papers in this field and I kept seeing like, you know, a sample size that I thought could be improved upon. And because I had quite a lot of knowledge of big data of these nationally representative samples, I thought to myself, well, I already have the data. I already know how to test these kinds of things. Why don't I come along and try and show whether these effects are really happening at a population level.
Professor Umar Toseeb: Yeah, excellent. And I suppose my understanding of what this paper looks at is you've got parenting. So how parent interacts with their child, the child characteristics. the temperament of the child, for example, and how those two things work together to influence the child's outcomes. You've described various theoretical frameworks or theoretical explanations for why those two things the ways in which those two things might interact to predict child's children's outcomes. And I think this is the real challenge, I think, on this podcast, which is how do you explain those four theoretical frameworks in language that's accessible because, and I've said this before, I have been trying to explain some of this stuff for about four or five, six, seven years, how many other years it's been I've been teaching. And every year I stand in front of the students with the same set of slides and I'm like, how can I explain this in a way that's accessible? So I'm going to lean on you and rely on you to try and answer that for me.
Dr. Robert Eves: Okay, should we try and break it down? So let's start with a simpler, most basic idea of the world. Okay. So let's say that birth weight, so how much the baby weighs when they are born is important for long-term developmental outcomes, right? So it might predict how many mental health problems they may have in the future. It might predict something like their IQ or their cognitive performance. And what we might expect is that as the birth weight increases, we might also expect the cognitive performance to increase at the same time. As you said, there are many things that influence things like cognitive performance. We might also think something like stimulating parenting is important. So how often you read to your child, how often you take them to a museum, for example. And if we looked at that measure, how much stimulating parenting a child receives, we might expect a similar pattern. So as you increase the stimulating parenting, you start to increase the cognitive performance of the child. The most basic idea here is that these effects are additive or independent of one another. So this means that as the birth weight increases, the IQ increases as the stimulating parenting increases, the IQ increases, but they don't influence one another. Okay, so that's like the independent or additive effects kind of idea. What this specific paper is really, really interested in are interaction effects and specifically different types of interaction effects. So feel free to interrupt me when I go on a bit of a tangent, but we'll try and describe these three different interaction effects in a way that's ⁓ relatively easy to understand. So, the first one is that the effect of parenting may differ depending on your birth weight. Okay. So for example, we might have an idea of vulnerability. And so let's say this, this was in statistics language, we would call this a diathesis stress model, or a gene environment interaction kind of model. And this is the idea that children who are born with a lower birth weight may be particularly sensitive to environments where there is not enough stimulating parenting, where there is limited stimulating parenting. And so if you take a child who is low birth weight, and in an environment without stimulating parenting, we might expect them to perform disproportionately lower relative to the rest of the population. So this would be that those two factors in combination have a kind of synergy and kind of result in disproportionately poorer outcomes.
Dr. Jane Gilmore: So it's greater than the sum of the parts, essentially.
Dr. Robert Eves: Exactly. And this is a kind of slightly negative kind of concept or idea. So that's the traditional diathesis stress kind of model. There's then the flip side of that, sometimes called vantage sensitivity. And this is that some people have an infant characteristic, which makes them more sensitive to good environments. Okay. So this would mean that in a bad environment, a child with a normal birth weight and a low birth weight, we might not expect them to differ at all. But as we make the environment have more and more stimulating parenting, we might expect the difference between normal birth weight and low birth weight children. To start to increase and increase and increase. But what we would be expecting is that the normal birth weight children start to really benefit from the environment where lower birth weight children may be less sensitive to the good effect of the environment. And so they're not improving at the same rate. And so this would mean that we would expect the biggest difference at the highest quality environment with those normal birth weight participants performing really, really well, and their lower birth weight participants not performing as well. And then we had, yeah.
Dr. Jane Gilmore: So this is sort of, sorry Rob, this is like the Matthew effect. So the rich get richer, right?
Dr. Robert Eves: Yeah, so yeah, that would be an interesting kind of comparison. I've never thought about it in those words, but I think that would certainly, certainly fit. And then finally, we have what we would call differential susceptibility. So this is a another nice long statistical term. But what it really means is for better or for worse. And so it means that you are sensitive to good environments, but you're also sensitive to bad environments. So let's say that being having a low birth weight, makes you just generally more sensitive to the environment. And what this means is that when you're in an environment with limited stimulating parenting, you may perform disproportionately worse. You might have the worst performance in regards to the general population. But because you're sensitive to the environment, if you're put in a really good environment, it means that you can take full advantage of it and you can actually be the best performing in the environment, when the environment is right. So this is suggesting that certain people can be ⁓ more sensitive to their environment. And our three different interaction models suggest three different things. So diathesis stress is that you're sensitive only to the bad, vantage sensitivity, only sensitive to the good, and differential susceptibility, you're sensitive to everything.
Professor Umar Toseeb: That's very clear. think one of the, if I didn't know about, well, if I hadn't read your paper and I hadn't read other similar papers, I would guess that differential susceptibility. So this idea that there are just people who are sensitive to positive environments and also the same people are probably sensitive to negative environments. And this idea that there are some people who are just more sensitive to the environments and there are other people who are less sensitive to the environments. That would be the one that I would say, intuitively makes more sense to me but I'm as we'll find out when we go through the findings of the paper that doesn't seem to be what the data is showing.
Dr. Robert Eves: Have you ever heard of the idea of dandelions versus orchids? So this is another way that people sometimes talk about it. Orchids are beautiful flowers, but you need to put them in the right environment for them to grow fully. Where dandelions, you can put them anywhere, right? They're pretty much the same if they're growing on concrete or if they're growing in a botanical garden. And so essentially, we might have children who are sensitive to the environment, which we could ⁓ an orchid child and they need to be put in the right conditions in order to flourish.
Dr. Jane Gilmore: Wonderful. How would you and obviously there's 30,000 participants, so you haven't individually given feedback, but how would you describe your findings to the parents who took part? Because, of course, some of those parents are vulnerable and some of those parents have had a background that might be difficult. So how would you describe those findings in ⁓ a fairly brief and accessible way? And again, Rob, this is a difficult question.
Dr. Robert Eves: Yeah, so one thing I would want to emphasize is that we're using observational data. We don't want to say that just because something is associated with better outcomes that we want to say it causes it. So I would just want to caveat or preface it with this kind of thing. But what I think we can quite confidently state is that the number of books you read to your child, how often you take them to museums, and how often you respond sensitively to their emotional needs, that's all strongly associated or at least associated with later development, both behavior, of internalizing problems or externalizing problems, and also cognitive development, things like fluid or crystallized IQ. ⁓ And this appeared largely true for all types of infants, okay? So it doesn't matter if your child is low birth weight or normal birth weight, it doesn't potentially matter if your child has a difficult infant temperament, so difficult to soothe, or if they had a sort of slightly more easy temperament or a normal temperament, what it seemed to be was that parenting factors were just generally associated with later development, and it didn't depend on what the infant characteristic was.
Dr. Jane Gilmore: It's a really good clear message and it's underlying and underlining, I should say, some of the data that has been somewhat obscured, I think, by the different models and so on. And I think it's, I think you described that brilliantly and it's about heading towards what we know is helpful. Here's what we can do. We know this has an impact in later development. That's fantastic. Thank you Rob.
Professor Umar Toseeb: In the paper, I think there's a part where you describe that some of these ⁓ person environment interactions might be different depending on the outcome that you're interested in. for example, it might be that in one situation where you're interested in a cognitive outcome like language or memory or whatever that might be, you have a particular interaction model or interaction effect that explains the effects more. Do you just want to explain that a bit more to us?
Dr. Robert Eves: Yeah, so let's go back to our idea of dandelions versus orchids. And we sort of talk about the orchid being beautiful, right, but this is just one outcome, which is like our visual look at it. And we might think that something like cognitive performance is important, and we want to know about it. But there are lots of other domains that we might be interested in as well. So the mental health of the of the child as well. And if our idea is that certain children may be more sensitive to their environments than others, what we want to test is, is that true if we're interested only in cognition? Is it true if we're only interested in mental health, or is it a general factor? And so when we're talking about a general factor, it means that they'll be sensitive to their environment generally, and it will influence lots of parts of their lives long term. And if something's more domain general, as a psychologist, I would generally find it a little bit more useful for the real world. Okay, so if it was only predictive of IQ, I'm not saying it's not helpful, but it would be really interesting if your something about you predicts how sensitive you are to the environment for lots of different outcomes.
Professor Umar Toseeb: Very helpful. So let's go back to this idea of big data. A single data set with like five, six, seven, 10,000 kids for this kind of question would be helpful and would be useful. But what you did was you took four large sets of data from different countries. Why is that important and necessary?
Dr. Robert Eves: So my background generally is doing ⁓ meta-analyses where you're interested in whether an effect is found in lots of different datasets, and then you're interested in where does it not work or where is it slightly different. So my background was more in how birth weight predicts cognitive performance generally. And what we did was we were interested in whether certain countries which have certain medical systems or education systems might be doing something different, which then might help children in a different way. So I had this kind of idea in my mind that, you know, parenting might influence developmental outcomes differently in different countries. So we had data from the United Kingdom, Ireland, Australia, and the United States. In some ways, these are all quite similar countries, they're all Western countries that speak English as their first language, or largely as their first language. But there are also cultural differences. So this would be adding extra evidence that what we're finding is really robust. And one thing that I would then want to emphasize is that if we're going to take data or results from one country and use it as evidence for how to change our education system or our neonatal system or whatever in another country, we need to first show that these effects are robust across countries. So that's the general aim. So that we give extra confidence because if we'd only use data from the UK, some people might quite rightly say, well, the UK is a little bit different from other countries that I know. How do I have confidence that that effect is also true where I live?
Dr. Jane Gilmore: And I really liked the way when I read your paper that you look very carefully about exploring consistency across the data sets so that you could assume there was parity, the data set that you are pulling from was irrelevant. I looked very carefully at some of these ⁓ infant characteristic questionnaires just to see, because it was a slight, there were slight differences in the position of, presenting the ideas of fussiness or characteristics of the infant to the parents. But you explored that really carefully. And I think you were confident that there were equivalents across those datasets in terms of the different measures used so that you could be confident that the whole dataset could stand for a population factor.
Dr. Robert Eves: Yeah, so this is a really good point, right? So this is potentially a slight limitation of using secondary data. So you can't change how questions were asked after they've already been asked 10 years ago. So I have to use the data that is already available and all of the different studies might use slightly different scales or questionnaires about how they ask things. So as you said, all the four different studies that we used had slightly different measures of how they measured temperament or how they determined whether a child was fussy or difficult relative to slightly more easy or kind of normal infant temperament, normal in a quote marks.
Dr. Jane Gilmore: Typical maybe, typical.
Dr. Robert Eves: Typical, typical is a much better choice of word. Thank you. And so one thing I did was I looked at these individual items that were being asked and I thought to myself, yes, I think there's quite a lot of overlap. They are generally being asked in very, very similar ways. I am happy to combine this data. So I feel like I'm comparing apples with pretty similar apples, right? I'm not saying they are exactly Granny Smith's or whatever, but I am pretty confident that the general fruit is the same in all of these different questionnaires.
Professor Umar Toseeb: And let's talk about your findings and the implications of your findings. So my understanding of what you're finding is that for the most part, it's the additive explanation that so that these two things operate independently in predicting the child's outcomes that are significant, that's coming across as important across various different data sets and different models. And then to some extent for some outcomes, it's the diathesis stress model and advantage point model. And that for me intuitively, I'm struggling to get my head around it because it's suggesting that parenting operates independently of the child's characteristics. But actually, if you asked parents, they'd say, well, actually, I'm guessing you'd say, well, actually, how I interact with my child depends on their characteristics too. So how do you put those two things together and make sense of them?
Professor Umar Toseeb: So I would say part of the rationale for the paper was that there were these past publications talking about moderation or interaction effects saying that certain children are more sensitive to their environment than others. And that's the key thing that we wanted to test, right? We wanted to test for moderation. The idea that you're kind of discussing or talking about is a slightly more closer to what I would think of as mediation, right? So the infant characteristic, influences the parenting behaviour, which might then influence the developmental outcome. And I think that's a very interesting research question, which you could certainly investigate. But it wasn't the key hypothesis for this study. So what we were instead was interested in these moderation effects. I did have a quick look at some of the correlation tables from the paper again last night. And I did see that there is a small correlation between for example, the infant temperament, which is rated by the mother, and the sensitive parenting that the mother also reports. So this could be an indication that, you know, when you have a certain type of child, that might then influence the parenting a couple of years later. If we wanted to really investigate that, I'm not sure that these datasets would be the best way to do that. I would instead want to do something like a sibling study or a twin study, where you could then look at how for example, twins can differ on their birth weight. So you could have a study where you look at whether the mum acts differently to her lower birth weight twin because she thinks that they are more vulnerable potentially as an idea. So that would be a really interesting future step, this kind of test of mediation or how the child influences the parenting. And that's certainly backed up by like some of the ⁓ most famous developmental psychologists of our time, right? So like Michael Rutter had theories that a certain type of child elicits certain types of parenting behaviors. And that would just need to be tested in a new study with different statistical models relative to this study.
Dr. Jane Gilmore: That clarification is so valuable because I think parents do describe day to day and I know because I am one but I also talk to parents day to day about the relationship you know relationships are two way experiences and of course they can influence one another but you're pulling apart the different findings from this study is not exploring relationship issues this is exploring ⁓ the impact of various environments on a child and their outcome.
Dr. Robert Eves: Yeah, and of course things aren't perfect with how you measure things like stimulating parenting and sensitive parenting because something like how many books do you read to your child a week? It's, in my mind, it would ideally be unidirectional. And what I mean is that the parent is the environment and they are influencing the child. But when we think about this in reality, if the child has no interest in reading a book and the child has many other hobbies, then it's not really unidirectional, it's bidirectional, right? It's that the number of books read is dependent on what the child wants to do. And so it might not be that the reading of the book is causal or the reason why their cognitive performance is higher. It could be other factors going on. So yeah, absolutely. When we're looking at the environment, it's quite hard to see a factor that's just unidirectional, where it only influenced the child, but the child doesn't have any impact in the reverse direction.
Professor Umar Toseeb: And that reminds me of some work that I did a few years back where we looked at children with developmental language disorder, which is they have impairments in using expressive and oral language. And we found that children with developmental language disorder, when they were in their first two years of life, they had a poorer early home environment. And I remember when I was presenting those findings, I was like, this is not evidence to suggest that the poor early home environment is causing...their language disorder, it's also dependent on the child. And it's very difficult to create an environment that's rich in language, communication, literacy, all of those things, if your child is not responding to it, because then they'll find where you stop or you reduce that kind of input. And I think that's the point that you're making here too. The outcomes that you looked at, I think they were at the age of five. So that tells us that everything that we've been talking about today, really relates to what we can be sure about for the age of five or early childhood. Might those effects be different if you looked at kids later in primary school or as they progressed through adolescence into young adulthood?
Dr. Robert Eves: Yeah, I think you're already reading my next grant application. yeah, yeah. So generally, I think it's very interesting to with developmental psychologists, we're interested in the lifespan, and to assume that everything is solid or fixed at age five would probably be quite a big oversimplification, right? So I definitely would want to investigate these things longitudinally. One thing I would say is that one paper can't do everything, it can't test all of these models different perspectives, different outcomes and different time points, because at some point you're writing a book, not a manuscript anymore. In regards to how much do I think these things change over time? I think that potentially cognition is a relatively stable trait. So for example, if you look at the correlation between Scottish, I think there was like a really nice study, which was Scottish children who were measured at age 11 and then measured again at age 80. It's like one of the longest longitudinal studies in the world. And I think the correlation was surprisingly high, like 0.6 or something like that. Don't quote me on that. We'll have to look it up. In regards to mental health, I think there's a lot more variation, right? We talk about adolescence as the key period of time where we start to see things like anxiety and depression really coming about. So potentially in this way, in this regard, in this paper, five years is a slightly too early measure. But the thing that I was really keen to focus on was that we did the same thing in different data sets in order to test the robustness. So that's why we focused on age five years. It was because the different data sets all had similar data at that time point. So we had more confidence that we were comparing those apples with apples.
Professor Umar Toseeb: I think you make a really good point there where, you know, there's, and I find it interesting to say it sometimes where someone's like, well, why did you focus on this time point or this measure? And there's usually a good theoretical explanation, but also there's, that's the data that was available and that was consistent. So that's what we looked at. And I think it's nice that you've mentioned that because you can't always have the data that you want and you can't have the time points that you want and you can't do everything. So you have to, yeah, like you say, you have to make it manageable and realistic within the scope of the project that you're doing.
Dr. Jane Gilmore: But better to make a strong point at five than a slightly dubious one, you know, at 10 and seven. So I think that position is right. And Umar, I think you make a good point about the practicalities sometimes driving the hypothesis, but that's so important and useful for those that are thinking about developing a project to hear. You sometimes you have to be practical and think about what you can ask and so that you can answer the question with some validity. I think we should probably let Rob go because we've spent quite a lot of time exploring these ideas, but maybe we can ask him to come back with the next grant and hear what happens next.
Dr. Robert Eves: My pleasure. Thank you very much for your time.
Professor Umar Toseeb: Thank you, Rob. I think he explained some of those gene environment or person environment theoretical perspectives or positions very well. I was very impressed with that because I was like, this makes much more sense to me. I mean, it made reasonable amounts of sense to me anyway, but I now have new ways of trying to explain it to students.
Dr. Jane Gilmore: And I think as academics, we do sometimes become a little bit tied by our language. And actually if we're forced and required to unpack what we mean by, you know, a technical term, if you like, we are required to think in a different way. One of my PhD training tasks was to write a tabloid headline and a tabloid article about my PhD. And I remember one of the facilitators at the Wellcome Trust, you know, sort of helping us do this sort of task. And one of the students said, my data is too complicated to write a tabloid to headline is a bit sniffy. And the facilitator said, if you can't explain it, then you don't understand it. And I thought that it was really useful information to me. And I think what Rob described was how well he understands something, but if you were required to put it in simple terms, it really does enrich the learning and the thinking. So it was brilliant, wasn't it?
Professor Umar Toseeb: Yeah. And I think also it was nice when I've tried to explain those person environment interactions, I've done it in with a different set of variables. ⁓ whereas Rob did it with his own set of variables that he's interested in so hearing it from a different applied perspective where you're talking about birth weight and temperament is not something I would usually do. So I'm like, this is really interesting, but I have a slight tangent. I think I just want to bring up here about the point of being able to communicate your research in a very accessible way, which I think Rob's done really well here. I was on the train yesterday and as an academic who does research on child and adolescent mental health and development and education, like one of the things that you want to do is impact policy and practice. So who is sitting on the train like four rows down? Who is it?
Dr. Jane Gilmore: I don't know, the health secretary. No, shut up.
Professor Umar Toseeb: The education secretary Bridget. And I was like, okay. I was like, is that who I think it is? And I'm like, yes it is. So obviously I Googled her and I was like, yeah, that's definitely her. And then I found, this is sad. Like then I found her Twitter and I was like, yeah, that's definitely her. She's wearing the same clothes on her Twitter photo. She's out voting or something. And I was like, okay. And then I text my boss going, is it appropriate to approach an MP on a train? Because like she is the correct person that I need to speak to about my work. And my boss was like, yeah, go for it. So after about half an hour of like psyching myself up, I was like, right, I'm going to go for it. And I went over and I talked to her about my research. But actually after that interaction, I was like, this is why you always need like an elevator pitch of this is the impact of my research. Because I was like, what do I need to say to her? What do I want to say to her? And I couldn't articulate it, but I did in the end. But what I also did was I was like, do you listen to podcasts and she was like, I was like, excellent. She's like, yes. And I was like, excellent. We've just got a new podcast. And I was like, this is what it's called. And I showed her my phone and she, she was like, okay, what's it called? So she might be listening. She might not be listening, but anyways, the point is it's really, it's a good skill to have to be able to communicate.
Dr. Jane Gilmore: I love that because you know actually and I'm also I like the fact that she listened and asked questions. I mean that's great. You know you're sort of on on on parade all the time but I think that's wonderful. I think you've got to be able to have your headline or your punchline ready before you tell the joke every time you meet the you know the Education Secretary. ⁓ Thank you Umar, that's brilliant.
Professor Umar Toseeb: Should we talk about what we think are some of the implications for us based on the conversation that we just had with Rob and the paper? So I think for me, it's interesting because these theories of gene environment interaction or person environment interaction, there's lots of theoretical frameworks like we've described. And this is the second set example where it seems that the interaction theories aren't supported. So like, well, not supported as much as the separate additive ones. And there was some work that I did previously where we looked at polygenic scores, are indices of genetic propensities and interactions with the environment. And we also didn't find an effect. And as part of the review process, the reviewer was like, this isn't that unusual. These indices of polygenic propensities tend not to predict stuff when they're in, well, they do predict stuff, but when it comes to an interaction effect, they don't seem to, they don't seem to produce the results that you'd expect. So this lack of person environment interaction or the lower salience of it in this example is more evidence that maybe we need to do this differently because I mean, I think it's, I don't know, is it intuitive that people's characteristics interact with the environment, which is the way we're investigating it seems not to be working.
Dr. Jane Gilmore: It’s so it's so interesting because it does. mean these models that Rob explained earlier are fairly well established in the literature and you know the interaction effects are too. But these data turned upside down some of the frameworks that have been pretty well established which I think was so interesting. It really did make me think about this paper quite a lot. It really was on my mind for a few days after I read it, which is unusual, I think, in an academic paper. So it's interesting. It's really interesting. I mean I thought about the clinical implications of this because I thought, well, on the basis of these strong data, these are population-based data, if you like, how does our messaging change to families? I think the short answer is there is no change because we know that warmth, consistency, sensitivity and predictability characteristics of best parenting practice were shown in these data to have a positive effect on infants in their later outcome. So in some ways you could argue that, you know, parenting practice hasn't moved on much since the 70s and Bowen Ryan's classic authoritative parenting, but we're much more aware of emotional literacy and understanding the nuance and communication and so on. The second thought in terms of clinical implications was the idea that, of course, there is no interactive effect here, but there is ⁓ the fact that we have both, if we are both a vulnerable baby and we have less optimum parenting, we will have a poorer outcome. So of course, we know that these additive factors often cluster in a family. So we're not talking about interaction, but the fact that, so for example, parents living in economic deprivation are more likely to have a low birth weight baby. And we know that socioeconomic deprivation predicts vulnerabilities in parenting approaches. Now, of course, there are complex environmental reasons for these patterns. But the point is that these additive vulnerabilities are likely to cluster in the same families. the models that Rob was exploring are very important. And I guess, and also, it's not a but, and also the idea that these factors are likely to cluster means that we are dealing with the factors in a clinical or day-to-day ⁓ basis in a very complicated way too.
Professor Umar Toseeb: I also thought the cross cohort comparisons were really interesting. And I think he did raise the point of the cohorts that he's comparing, the countries that he was comparing were quite similar, nonetheless in their different contexts. So they still need to be understood. And I didn't ask my usual question around what does this mean for other cultures that aren't white Northern European cultures? But actually, it applies here too. I randomly came across this paper a few days ago where they compared, so you know, we always talk about mental health deteriorating amongst young people and it's getting worse, the mental health crisis. And there was this paper where they compared mental health in adolescents, so young people in the UK, I think, compared to Brazil, I think it was. And actually they found that in Brazil, it wasn't getting worse and it was getting a bit better, I think is what they found.
Dr. Jane Gilmore: That's interesting because I was looking at it across the of across the globe and there was there were some countries that had unique characteristics but there was also quite a lot of consistent data. Now we're talking about it and I can't remember the papers that described it but cultures that were quite different who'd also got increases in mental health distress as well as those that didn't. So what to make of that I don't know maybe that's another podcast episode.
Professor Umar Toseeb: Yeah, I think it is. think we'll have to do a podcast about cross-cultural differences in mental health crisis or not in some countries. I think so. Shall we? ⁓ We have been talking quite a long time. Shall we go to our takeaways? Come on. So my takeaway is that ⁓ when we're looking at the interactions between the person and the environment, there are a broad range of possibilities and Rob described three and there's the additive model as well. And actually we should be considering all four possibilities probably at the same time so we can compare the effects and usually and I'm definitely guilty of this. I just look at one. I'm like, oh, I'll look at the diathesis stress model because that's the one I know about and that's what makes sense. But actually there are four different variations at least here and we should be looking at all of them and seeing which one is the best at explaining the outcome that we're interested in.
Dr. Jane Gilmore: Brilliant. And I think from a clinical point of view, it's about clinicians engaging with both the group effects, these population effects that we've discussing and the individual differences. So these big data findings are crucial for policy and theory development. I hope the education secretary is listening to that. And also the well-established risk and protective factors that had the predicted impact on these infants at the age of five. So it's a reminder that these vulnerabilities need to be considered you know, on a case by case basis and consider these treatment packages in that way too. Matt, our producer, has just reminded me that you can get a CPD certificate for listening to this podcast. And there are hundreds of hours of evidence-based child and adolescent health content on the ACAMH Learn website. You'll find it on acamlearn.org.
Professor Umar Toseeb: Join us again next week, we'll be speaking to Germán Carcia Grande about the intergenerational impact of mothers and fathers on children's word reading development.