Protective Factors For Children - video transcript
Speaker: Matt Walsh
Thank you, Paula. Really happy to be here today to talk about what we've been working on for the last two years. Also very fortunate that we were even able to have funding from the Growing Up in New Zealand data to do secondary data analysis, so we haven't collected any of this data but since it's available and there's funds, luckily from here in MSD, to look at the data. I'm very happy to bring that up. Let's see how we switch -- I can stay on this slide too, okay, there we go.
So, just some background on me, so I'm -- you can probably hear I'm from America, I've been here two years, and prior to coming here I worked for the Department of Children and Families, which is a little like OT in Wisconsin. So, when I'm coming at all of these researchers I'm in academia now but I kind of know that none of this works without interacting with the programme and policy analysts that are at OT.
So, the goal here is to kind of say what we have and then kind of have you guys have lightbulb moments, hopefully, to then have connections that we can actually bring this somewhere further than in academia.
So the goals for today, first talk about ACEs, Adverse Childhood Experiences and one of many ways to measure bad things happening to children. Then to talk about, there's been four rounds of this secondary data analysis funding from Growing Up in New Zealand and we were lucky to get some money from the first round and now the current round that's going in the third, so talk a little bit about what we did in the first round and how that informed what we're doing now in the third round and then talk about the current work.
So, just to start off with, how many people here are familiar with ACEs and the ACE literature that's out there? So, good, that's great because it's been about 20 years ago, actually even longer, that they first had a study in California where a researcher tried to figure out why some of these morbidly obese women that they had worked with, so many of them had been reporting sexual abuse as they were children.
And it came through now 20, 25 years later in a huge wealth of literature that's kind of talking about Adverse Childhood Experience. And what they really are, just intensive and frequently occurring sources of stress. So, things that are happening over and over again that a child's witnessing or experiencing, that then changes their biological processes for how they are interacting with the world. So, fight and flight responses, everything seems to be correlated with these frequently occurring stresses that are happening.
Then originally they had decided they were all in these areas here of abuse, neglect and household dysfunction. So, abuse and neglect, that's very key to what's going on at Oranga Tamariki but household dysfunction, things like having a parent that's under mental illness or having a jailed parent, experiencing a divorce, these are all different Adverse Childhood Experiences and they're common. So, throughout the world a lot of the literature is from Europe and from North America but more than two-thirds of adults in general have at least one of these adversities when they're growing up.
The literature really says the more you have the worse off it is. It's a very clear dose-response that you see. They also normally clump them together, they don't try to say, "Okay, this one's worse than the other." We're going to try do that a little bit, even though I don't usually like to do that. But normally it doesn't matter which one it is, they all seems to have the same way of changing the biological response in the child.
I'd say there's probably 10,000 scientific reports that seem to be going through Adverse Childhood Experiences and in any area you can think of there are correlations between having these adversities in childhood and then having these bad outcomes; that could be cancer, death due to cardiovascular disease, experiencing risky behaviour yourself like smoking, even educational outcome and occupational outcomes. There's pretty much -- you name the outcome and somebody probably has a study looking at ACEs in that study.
The goal here was to say since there weren't as much research going on with ACEs here in New Zealand, the first was to say, can we use the Growing Up in New Zealand study to look at -- can we measure ACEs in these children? So the kids right now in Growing Up in New Zealand are about ten, so they were enrolled in about 2009. The good thing about this -- it's almost 7,000, it's all in these three pointed-out DHBs, so Auckland, Manukau and Waikato.
So the original, about 7,000 women were enrolled and the good thing for our study is that once they enrolled the participants, they had really good retention rates through the first five years. They're still collecting data now but that's good because you will see all of our data comes from what's happening when they're 4½ years old.
So that first round, there were two reports that came out, they're both searchable. If you kind of just look for Walsh and Adverse Childhood Experiences or MSD you'll find the reports. There's a whole bunch of limitations that are going to be included in there and methods that I’m going to be skipping but they are located there, so if I'm skipping things that you want to know, I'll point you to those but I can also talk about them as well.
What we found is, yes, you can measure adversities in children through the same standard ways that they had done for the Dunedin study that was before it. By the time the children are 4½, more than half of them have experienced at least one. So, that's the zero, the blue bars, it's the per cent that 47% that had zero, 30% had one and the other, about 25%, had more than two.
So that was not surprising for people that are looking at it but if you are like telling people that aren't used to the Adverse Childhood Experience literature that's usually something they're like, "That's not good". So the first step that we wanted to do is, can we predict? At the Centre for Social Data Analytics where I work we do a lot of predictive analytics using administrative data, either through the IDI or through other sources and other partners.
The first thing is, can we predict based on what we would call the administrative data, data that's already known about a child when they're born, can we predict which ones were most likely to have adverse experiences? So, we use a subset of the full sample randomly selected, 3,500 children and we looked at everything we knew about them when they were born. We are able to segregate or put them all in a line, from the lowest risk of having ACEs to the highest risk of having ACEs, so about putting people in little groups of 700 kids per group.
What we were able to find is that, after we did this risk, we were then able to follow them up for 4½ years and see who actually had ACEs. You will see that we are able to predict who is most likely to have these ACEs, so in this case the highest risk group, more than 55% had at least two or more ACEs by the time they're 4½ years old.
Of course that means 45% didn't have ACEs and so the goal of the research isn't really to look at who is going to have ACEs and find the risk factors for ACEs, it's really to say, who did we predict to have ACEs and who didn't have ACEs and what is the resilient factors for those children that didn't have ACEs? What can we learn from them that might give us more information for how to protect children from Adverse Childhood Experiences? So, as I said, 55% had two or more ACEs, that's this bottom row but 21% of these children that were at the highest risk didn't have any.
So, the first part of our research was to say, okay, Growing Up in New Zealand collected thousands of variables about pretty much anything you can think of and it's probably in there. What can we find that isn't already something we use for the predictive model, something that we know already about when they're born? Do they have a lot of litter in their neighbourhood? Do they interact well with their partner? What can we find out that might help us understand which children are at the highest risk but actually have no adversities?
So lots of dots here, each one of these dots is a factor that was found to be associated with doing well, so being at a high risk but not actually having any. The first thing that came out to me was there's a lot of dots at the parent/partner. From the child welfare background that I have, we had a lot of policy levers in Wisconsin to cut a deal with family finances, "Okay, we can give you this support that you need".
& We also had a lot of -- the social workers are sent in to kind of watch the interaction between the parent and the child, that's like the key that we were looking at in Wisconsin. And a little bit of parent health and wellness, you can find programmes and policies that kind of help the mum get the resources they need.
But we almost never, at least in Wisconsin, looked at this parent/partner, sometimes that was because there's a thought that the partner is the one is causing the problem, separate from the partner. But in this case it seemed like -- so the dots by themselves are good because that's something we find but also the further you're over -- I can't do like your left and your right, I get a little confused. So the higher the number, these are standardised, so that means they're having the larger effect. Not only are there more dots but they're also stronger.
Not only that but if you were to think about the number of factors we looked at, like I said, Growing Up in New Zealand was collecting a whole bunch of different things and it's based on the researchers doing that data, since it's secondary data analysis. They really were focusing on these community and neighbourhood factors. In fact, there's about 750 factors we were looking at, of which a third of them were under community and neighbourhood, whereas only about 60 to 65 factors were found in parent/partner. So, not only do we have the most number found in parent/partner but we also had the fewest things that could have come through.
So, to me the first things that came through is, okay, we need to maybe think about how we engage this parent/partner relationship more than we are doing. Maybe they're doing more here in OT but at least in Wisconsin this is an area that we haven't been focusing on as much. As we are waiting for other things to come, that was the main goal of that first part but there was some time to look at other things and Growing Up in New Zealand really collected a lot of really amazing school readiness, it might not be the best term, 'school readiness', because all children are ready for school when they turn five.
But school readiness exams that they looked at, can you write your name? So, the top one is if I clap once, can you clap twice, this inhibitory control? They also looked at here's a list of upper case and lower case letters, can you identify these letters? How many can you identify in 60 seconds? A gift-wrapping test, which is a little like the marshmallow test, if you've heard of that, which has now been a little bit debunked. Effective knowledge is, can you tell if you look at a face if it's happy or sad, that's the effective knowledge? Number and name writing, can you write numbers on a piece of paper?
And somebody would evaluate, yes, they seem to be able to write numbers. And then counting up from 1 to 10 and 10 to 1, those are self-explanatory. These are standard tests that they are administering to all the children that are 4½ years old. If you go back to that original 10,000 articles that are on outcomes, almost none of them are, do you already see negative outcomes by the time the children are 4½? This was kind of an opportunity to look at that in Growing Up in New Zealand.
These are small. The goal that you should see is that every one of these things is kind of declining. So as you go up here, these are the number bases that the child experiences, zero to four or more and every single one of these, if you go up to how many letters can you name, you see that by the time you get to three or four more you have more than half the children can identify more than one letter in 60 seconds on a sheetful of letters, whereas for the children that had zero ACEs, three-quarters of them could.
All of this here kind of showed that a very dose-response kind of fit with the literature already, does response and negative effects of these ACEs that are happening. That finding there was what kind of drove the whole second round of what we're doing and what we're now working with OT with, with Paula and Eyal, to kind of say, okay, let's look at the school readiness, is this the same in teen mothers? Do we see the same trends for that subgroup of population?
& Then also a lot of these factors, a lot of questions like, what can you change? What would you change from a policy standpoint here at MSD or at OT? How would you change this to kind of help the children that are experiencing ACEs? So the first thing we did was we segregated the same analysis we saw before where we said it was over half of the children had experienced at least one ACE. This is now based on the age of the mother at the birth of the child. If a mother was less than 18 at the birth, about 60% had experienced zero or one and then here for the 18 and 19 group it's a little bit higher, so 60% -- oh, little lower I guess, 55% had zero or one, so that's a little worse for the children. The larger of these three areas, that's the two, three or more ACEs.
You can see that there is a definite association with mothers that are having babies as teenagers, their children are experiencing more adversities. Again, not a surprising finding for anyone who is looking at this and it's not because of the teen mum, it's for a whole bunch of other reasons that are going on. The question then is, who do we approach? If we want to minimise Adverse Childhood Experiences, we just involve all teen mums or should we use this sort of predictive approach that we were talking about earlier where you look at all the factors you know and try to decide?
If that's your question preventing Adverse Childhood Experiences, what you have here are these curves of -- if you remember before I kind of made a big line of people and I split them into five groups. I should have probably split them into ten because this example now has them all split into ten groups. I have this whole line from the lowest risk for the predictive model all the way to the highest risk for the predictive model and I'm looking at the top 10%.
& So this is this group, so this would be that area that I'd highlighted before and if I was doing the five groups it would be nine and ten together. What you'll know teen mums, mothers that have the babies when they're still teenagers, they do have slightly higher risk of having Adverse Childhood Experiences because the line is slightly above it but you would still want to not just select all teen mums because, as you can see, there's a lot of teen mums here that are at much lower risk than choosing the whole population based on the predictive model.
Policies that are just saying, okay, we're just going to roll all teen mums, might not be the best use of resources. If you knew the two things, you knew whether or not they're a teen mum and you knew their predictive score, if you only had enough money to -- if you first select teen mums that are in ten, that would be the area that you'd want to focus on most. But then the next one you'd want to focus on all people are at ten before you would start taking teen mums that are lower.
& That's like the first thing to me because there are a lot of programmes and policies everywhere that just kind of make it easy and say, okay, if you're a teen mum you're going to be enrolled. That said, that was to prevent Adverse Childhood Experiences. If you want to get school readiness better for children you'll see that on every one of these curves, so teen mums are grey again and the non-teen mums and the higher you are the better you are at passing that school readiness test.
At every area, no matter where you are in the curve, teen mums' children are doing worse. In fact, if you look at this here, at zero ACEs, a teen mum at zero ACEs only 52% are passing this letter test, that's how many letters they can name in 60 seconds. So if you kind of go across that, that's like the equivalent of a non-teen mum with three or more ACEs. Again, it's always kind of complicated, like do you want to deal with preventing ACEs or do you want to deal with actually having the children be school ready? I guess it might depend if you're at OT or MOE or who knows what your priorities might be. But every one of these and you'll see these are just two that I'm showing to kind of split it up but every one of the six school-readiness tests in every single instance for the same level of ACEs the teen mum is always doing worse than the non-teen mum, mothers that have children in their teens; I'm not sure how to phrase that without tongue-twisting.
Again, it kind of shows that instead of dealing with just bucketing all teens maybe in this case you would, if you really wanted to be having children ready, you would be like, well, it doesn't matter. Teen mums need the services because all their kids, on average, are doing worse on the school-readiness test. That was the first thing we started connecting with Paula and Eyal here at the evidence centre. Then they really kind of pushed back and said, "It's good you're doing this total ACEs thing but I want to look at individual ACEs and which ACEs are important or not". I pushed back as much as I could but then I thought, okay, it makes sense; they seem to want this.
I went through the process and this is a little hard to tell, the number is the number of the school-readiness tests, so there's up to seven tests that we are looking at. This total is kind of what I was showing before, so six out of the seven school-readiness tests were statistically associated with having ACEs. So the total number of the ACEs, the more ACEs you have the less likely you would be ready for school, if we're using this terminology.
Splitting it up you'll see that it really seems to be driven by the physical abuse indicator. Now, that could just be because for children that's the one that they're feeling more and more often and that's what's causing the different response, biological response. But it could also be, who knows? There's hundreds of reasons; I could just sit there and make something up. But it's interesting and I think it's interesting for OT because that's one of the main areas that you guys are focusing on is physical abuse and mental abuse.
Unfortunately, there's no neglect currently being collected, so we couldn't see that one. But these are the ACEs again and you'll see, we did look at the teen sample because that was the goal for this. The numbers get real small in the teen sample, so you can't see it too much. But, again, you will see that the physical abuse seems to be the one that seems to be most associated with the school-readiness exams. So, that was the teen portion of it.
The other portion I'm going to say, okay, what can we do with children that have already experienced ACEs? The first time we went through we went through all these factors that were in community and parent/partner. Now we wanted to look at factors, the governmental programmes that are going on. What are the factors, preference, quality, access, kind of the health services' research terminology that's used often?
We ended up choosing healthcare, early care and education, social services and splitting them into those four buckets. If you look at this you'll see that most of the factors that Growing Up in New Zealand were collecting were on health service utilisation, so that's the biggest bucket there; 135 out of the 372 but there are other buckets that we're looking at. Doing the same thing that we did before where we kind of like see which one of these 372 factors are associated with each of the tests, kind of seeing how important they are for the tests we did. We start getting a whole bunch of -- we can spit out graphs like this to the end of time.
Hard to follow exactly, it wasn't as clear to me and when I looked at the parent/partner one that was like, okay, well that's obvious, that's parent/partner, not like that but it's something that we're now looking at. Here it's a lot harder to say, I mean you have more dots in the healthcare utilisation but, like I just said, there were a lot more factors that we were looking at for healthcare utilisation. So this is where we really need people to kind of be excited about their own programmes, to look at what the results say for them, try to decide what does this mean for what we're doing and how could we change what we're doing, based on these results?
Another thing that we are kind of working on, so a follow up to what we just showed there is to kind of say, okay, these were interactions, so these are children that have ACEs have a different effect for the programme that you're looking at. One example that I'm going to go through, just in the physical abuse and the total ACEs there is almost 100 of these. Some of them make sense, some of them don't. They're all proxies, so don't ever think, okay, this is going to your GP, you shouldn't go to your GP when you're 24 months because it's not the GP that's doing it, it's all this other stuff that might be associated with going to a GP.
So here we have a red line, those are people that have the physical abuse indicator, so there's an indication that they're having this Adverse Childhood Experience of a physical abuse. The blue line are children that don't. You can see that if they don't report, so there's a question, who did you do your well child check with when you were 21 to 24 months old? One of the options is the GP. The people that don't say, yes, they went to their GP, these are these people here and you can see that there's not that much difference in their scores for school readiness for whether or not they went to the GP when they were 21 to 24 months old. But if they report going to the GP and they had Adverse Childhood Experiences, for some reason you'll have to get the people that are involved with GP programme and policing to kind of realise -- I mean we can come up with the fun ones ourselves but they're all just kind of -- for some reason people that go to their GP that have been physically abused, that's a real big indication that they're not going to be ready for school for this test.
Counting up is a score of, can you count up to ten? If they go one, two, three, four, five, six, seven, they stop right there they got 7. If they start at two and they go to ten, they get 8. So, the score over here is how many numbers in a row that they got. Most children that don't go to the GP seem to be pretty much right around 8½ and maybe they skip the seven, like my four-year-old seems to skip a lot. But for some reason if you have been physically abused or you have an indication of being physically abused, again, the definitions for physical abuse are not quite the same as OT; you will have to go back to those reports to kind of see the differences. But for some reason they have a worse score.
The question is, what do you do about that or what could you do about that? Again, trying to write this in another different way here is, if you use the GP during your 24-month well-child exam, it may not indicate whether or not you can count to ten but if you have a physical abuse indication as well, that is an indication. If I'm a doctor and I get people coming in, if I don't know anything about them I would just think, they're probably going to be about 8.5 to 9. But if there's some screenings to all that might give the same indication for a physical abuse, I would automatically know, not because of their coming to the GP, but I would automatically know these children might need more services.
If there was a way to kind of then think, okay, can we get the people in the room that are involved with setting policy and programmes for what surveys need to go out for the GP or the 24-month well-child exam, can we add this component into them? If we can, what services can we offer in addition to those kids that have indications for physical abuse?
Like I said, there's 100 of these, so really kind of requires working through them individually with the programme and policy analysts themselves because I can play with numbers all day and it doesn't really get anywhere. The goal here is really kind of say, what do we have to work with, Paula and Eyal to kind of package that to a point that somebody can read that and understand if they want to be involved with that? There's a lot of limitations for the research that we did and the methods that we used. The biggest one is there is no a priori hypotheses, we just threw -- there was 100 things that came out. There is probably 2,000 tests that are going on. There's going to be spurious findings, to know that this is the first step to kind of start the conversation to do an intervention and to do a test that might go through.
The goals for what we want to do, one is increase awareness of ACEs. There were still -- glad I saw a lot of hands but it's also OT, so you guys are kind of involved with this area a lot more. The first part is really to set up a strength base, so try to say, okay, what are these protective factors? Then we also want to develop strategies for the mother/partner and then for the work that we're doing right now is to really give more information to people that are doing programmes for teens and then also going through these examples and finding partners that want to go through it.
So, I really want to thank the Growing Up in New Zealand staff who gave us access to all the data and really thank Paula and Eyal and everybody else that I work with and thank you for listening.