Documenting childhood inequalities and the case for early intervention

In episode 11 of the DIAL podcast, Professor Gabriella Conti from University College London discusses two pieces of research part-funded through DIAL’s Growing up Unequal? The Origins, Dynamics and Lifecycle Consequences of Childhood Inequalities project.  The first investigates socio-emotional inequalities in children born in the UK in the 1970s and the Millennium and the second investigates the long term health benefits of the UK Government’s high profile Sure Start programme. 


Christine Garrington  0:00  

Welcome to DIAL a podcast where we tune in to evidence on inequality over the life course. In series three we’re discussing emerging findings from DIAL research. For this episode, we’re talking to Professor Gabriella Conti from University College London, about two pieces of research. The first compares the behaviour of children born in the 1970s with those born in the millennium, the other looks at the long-term health benefits of the UK government’s Sure Start programme.

Gabriella Conti  0:26  

So we know that early human capital is a key determinant of lifecycle outcomes. And by now we also know that if there are early life inequalities that can perpetuate and amplify a person’s life cycle. And so it’s really important to document existence of early inequalities if they’re present, thinking about what we can do about them. So there was one starting point we ha. Another one was that we know that inequality has been increasing throughout the developed world and in particular in the US and the UK in the recent years. And there is likely less evidence for inequalities in child development. In particular, we were interested in the dimension of child development whose importance is being increasingly recognised which is child social emotional development. And so this is why then we started looking into data and see what we could do to document the evolution of inequalities in the social emotional development of children as early as we could.

Christine Garrington  1:26  

So how did you go about the research tell us what you actually, what you actually did?

Gabriella Conti  1:30  

You know, the documenting evolution inequalities for a long period of time requires that you have data which can be compared across time. And in the UK we’re fortunate enough to have British Cohort Studies, which essentially follow the life cycle of cohorts since birth. And in particular, we use the data from two of these cohorts. The British Cohort Study, which has followed the cohort born in one week in 1970. And the Millennium Cohort Study which has followed a cohort born throughout 2000. So these were like, a thirty years apart. And we had pretty big sample sizes more than nine thousand for the 70 Cohort, more than five thousand for the 2000, so called Millennium Cohort. So we were able to extract from these two data sets, relatively similar questions on child behaviour as to the mothers in the two cohorts. At age five so relatively early in life, and mothers were asked questions about whether for example, the child is restless? The child is solitary? Child is screaming or fidgety? And we construct the comparable skill so for two important dimensional social emotional skills, which have been used widely in an interdisciplinary literature on child development, namely, externalising and internalising.

Christine Garrington  2:51  

Can you explain in simple terms, a little bit about what we mean by those types of things?

Gabriella Conti  2:55 

Such emotional skills internalising refer to the child ability to focus their drive and determination and externalising relates to interpersonal skills. So a child with better externalising skills is likely less restless, hyperactive, less antisocial and a child with better internalising skills in less solitary, neurotic and worried. And one important thing that it’s good to notice at this point is that while you know the same questions have been asked to the mothers across the two cohorts, and we were super careful and just used the questions which were worded in a similar way of course it could be that what is perceived as a hyperactive child has changed in thirty years. And so we use reasoned methodological advances to take this into account to make sure that we’re effectively comparing the same constructs that showing the development of children and this methodology can be used in other settings. And we’ll show that it’s really important to use as we show in our application.

Christine Garrington  4:01  

Just talk us through then the key things that sort of emerged once you’d gone through all of this, once you’d look so closely at that data, once you’d put those methods in place and once you’d used those really robust approaches.

Gabriella Conti  4:12  

Yeah, so actually, we find that inequalities in social emotional development – this externalising internalising behaviour of these very young children of five years did increase in this 30 years between 1970 and 2000. So this kid we’re measured in 1975 and 2005. And it’s important to notice that we’re not able to say whether one cohort was better or worse than another because we weren’t able to compare the levels. Because thanks to our methodology, we were able to compare the difference so the inequality between these two cohorts but no whether one was better or worse than the other. But we saw very clearly that the inequality in the cohort born more recently, so the one born in 2000 was greater than one born in the seventies. So in particular, we see increasingly this early gap between the children with the highest and the lowest social emotional skills. So for example, no matter which measure we used we saw that for example, if you looked at the difference between the 19th and 13th percentile, this has widened substantially in 30 years. And this increase was particularly pronounced for boys. For example, for boys, the gap is increased by 19% for externalising skills and even more 30% for internalising skills. So inequality in the social emotional skills so very young children are five years of age was much lower among the children born in the 70s than among those born in the 2000.

Christine Garrington  5:50  

I will ask you what you make of all of that in a moment, but you also took a number of sort of different factors into consideration, didn’t you including the mother’s level of education, what did you see there? That was important?

Gabriella Conti  6:00  

Yeah, indeed, because as we know there have been several changes in the composition or the population composition or the workforce but also changes in women’s experience, especially in the labour market. And so we wanted to look not only what happens in these gaps across the groups but also zooming in on particular groups. And so in particular, we found that even when we compare children or mothers with different characteristics, we found increases inequalities and in particular, while having more educated mothers or mothers with healthier behaviour was an important determinant to the skills in both cohorts. We found that the benefit of having the mother we higher levels of education or a mother in employment was significantly larger for both boys and girls in the most recent cohort than in the cohort born in 1970. So in other words, the difference between the children are more or less educated mother was greater among those born in 2000, as compared to those born in the seventies. And we found this for children’s mothers with respect to mothers education, with respect to mothers employment, but those inequalities increase between children and mothers who smoked and not during pregnancy.

Christine Garrington  7:19  

Okay, so some really interesting findings. I’m interested to know whether you were surprised by what you found? And if so, how you explain what you found?

Gabriella Conti  7:26  

I might sound a bit cynical but I am going to say I wasn’t actually surprised about the findings per se, because we do know that there have been many increases inequality across different dimensions. What I found really surprising was the extent of the increase and the fact we could see that so early across such an important measure of development. So we did spend a lot of time in trying to tease out the various determinants of this increase in inequality and as I mentioned before, there being significant societal changes in these thirty years so for example, it’s also documented in the paper across the two cohorts. The average age of women have children has increased by approximately three years from 26 to 29. The proportion of women in employment has increased in our data we see from 42% to 62%. And especially the proportion of unmarried mothers is increased dramatically from 5% in the 70 cohort to 36% among mothers in the 2000 cohort. So mothers are having children on one hand at an older age and when they’re more engaged in the labour market, which is good for social emotional skills. But on the other hand, mothers are also more likely to be unmarried at birth, which is more stressful, and so it’s less good for social emotional development. And we have used the methodology to decompose these factors. And we found the changes in these factors explain about half of the cross cohort increasing inequality when it comes to externalising skills. 

Christine Garrington  9:02  

Okay, so let’s consider them the, the implications of these widening inequalities that you’re hinting at here and talking about here, especially in the you know, recent context of the COVID-19 pandemic. What would you say overall Gabriella that we learn from this work?

Gabriella Conti  9:17  

So first of all, let me remark that we’ve seen during the pandemic because there’s been increases in inequalities in learning experiences at all level by the children in the home environment. So the children would have been better able to have learning for example, at home or will have had the parents will be better able to cope. They certainly been affected less severely than children in a more disadvantaged situation. And so now it’s starting to be documented that in particular, children’s social emotional skills have been affected by the pandemic. And on the other hand, as have documented those in other work, support especially for the more disavantaged is also diminished during the pandemic. So for example, done work on the Universal Health Visiting therapist showing that they’re in addition to the cuts of course for many years, they’ve been an unequal in widespread deployment across local authorities. And so the pandemic there has been a double hit on the one hand families that have been unequally affected depending on individual circumstances. On the other hand, also, you know, public services so which are supposed to help families and to help preventing inequalities have been also unequally affected. And so it’s really crucial that the government takes stock of this and hopefully, in the upcoming annual review, it provides more support for the earliest and prevent at least the widening of these inequalities.

Christine Garrington  10:47  

Yeah, on this whole idea of the importance of early intervention being key to reducing inequalities. We’re really fortunate to be able to talk to you today about another piece of research that you’ve been doing, and looking at one of the best known really policy innovation interventions in this era, particularly here in England, which is called Sure Start. Now, I think many people are familiar with it, but for those who are not I wonder if you can just talk us through what Sure Start is?

Gabriella Conti  11:11  

Yes, thanks so much for this question. Indeed, with colleagues at the Centre for Fiscal Studies we have been working for a few years now on Sure Start and Sure Start is really a major early education initiative in UK, which was originally area based. And for those who don’t know, it has quite a long history. It was indeed first introduced by the Labour government in 1999, so called Sure Start local programmes, and the idea was to give quality services for under fives only in disadvantaged areas then it was so popular and so well accepted by families that the government gave it the major change in 2003 and changing it to Sure Start Children’s Centre which were gradually rolled out across England and at the peak they reached more than 3500 of these centres. And what was really nice about the centres was first they provided a physical place for parents to go to bring their children to interact with other parents and it has a wide variety of services which the parents can use. Ranging from early education services, parenting support, childcare, there were health visitors providing the visits there in the Sure Start centre. There were signposts for job search assistance. So think of it really, as a one stop shop for families with children under five. If you had a child and you needed help for health, for other reasons – so you wanted a childcare place, then you could just go to Sure Start centre. And it’s also to be said that at its peak in 2010 Sure Start also received a third overall earlier spending as much as £1.8 billion a year. But then unfortunately after 2010 spending has fallen by more than two thirds, many centres have been closed, they’ve been scaled back or they’ve been renamed children’s centres. Now you don’t even hear Sure Start  or they have been integrated into family hubs.

Christine Garrington  13:10  

Yeah, so big changes over, over the piece there and your research on Sure Start has looked quite specifically at its effects on children’s health which is, you know, really important and you started by looking at hospitalizations and links with that. What did you, What did you find there?

Gabriella Conti  13:26 

So we have a data set which contains the exact address and opening date of each Sure Start both local programs and children’s centres so we were able to look at the origin of the programme and its expansion up to 2010. And then in England, we have a very good dataset called Hospital Episode Statistics where there is collected data on the universal patients using English public hospitals. So this is important because we can really have a very comprehensive look at the effects across all Sure Start centres for all England the local authority so it wasn’t like a selected or a small sample. It was quite good administrative data. And so what we did – we essentially will be focused on the expansion period of Sure Start so as I say, 1999 up to 2010. So where there were all these centres opening up. And then we’re looking at what happened essentially when you have access to Sure Start since they’re in your area. And what we find, that we find very interesting effects which changed across the life cycle of the child. So first of all, greater access to Sure Start in your area initially increases hospitalization at age one and this shouldn’t be really surprising, given that it’s common in other studies as well and my reflect the fact that the programme would help referring the parents to proper health care in the case of childhood illness, but also brings their exposure to infectious illnesses from other kids being all in the same setting. However, what is really remarkable is that these early increases are more than outweighed by longer term effects. So as we look throughout childhood adolescence, we find that there is a reduction in hospitalizations.

Christine Garrington  15:18  

And there were even longer term benefits for these children as they got older weren’t, weren’t there? Talk us through, talk us through what those benefits were.

Gabriella Conti  15:25  

So we found is that exposure to an additional centre per 1,000 children at ages 0-4 average around 7% of hospital admissions at each file. Which goes up to 8% by the end of primary school, so age 11 and 8.5% by age 15, which is the final age of this study. Now this represents approximately 2800 fewer earlier hospitalizations at age five and over 13,000 hospitalizations prevented of 11 to 15 year olds each year. So that’s quite a substantial number. It’s pretty much an 8% reduction on the pre-Sure Start hospitalizations rate and this is all completely looking essentially at the peak level of Sure Start provision where there was one centre per 1000 children available throughout England. And what is really interesting, while Sure Start was for children under fives, we’re able to detect impacts which are increasing over time and are able to detect them essentially 10 years after the children have aged out of eligibility. And if we look at the drivers of these reductions hospitalizations, we see that it’s important conditions such as external causes, so things like injuries and also mental health related admissions in addition to infectious illnesses.

Christine Garrington  17:00  

Those are just really striking findings Gabriella and I know that you also tried to get a feel for the potential cost savings that Sure Start would have generated. What were you, what were you able to say about that?

Gabriella Conti  17:11  

Yeah, I think it’s quite important is this not only to provide evidence so that a programme is effective, but that it also provides enough bang for the buck. And with Sure Start it is particularly important given as I said before all the money, which was spent at the peak of the expansion. So what we did in this case, essentially, we had on the one hand, the money that the government had provided, and so we computed in this way a cost per child which is really more like the amount of money that the government spent on Sure Start per child, and we computed this amount to be approximately £415 per eligible child which is lower by the way than the cost of other programmes. And then what we did we looked at our estimates, and we considered first of all, we costed only the results, which for which we found significant impact. So as I said, subset of external conditions in particular injuries and also poisoning, then infections. So respiratory, parasitic and then mental health as you can imagine, you know, those have huge cost. And we considered three different types of costs. So on the one hand, the reduction in hospitalization has a direct cost saving in terms of the healthcare sector, so you will have less money essentially spent because fewer kids are going to be in the hospital. And this is more like a short-term cost, the cost at the time at which we estimate the impacts. Then there are assorted indirect costs, so if parents don’t have to spend time taking care of the sick child, they are not going to be absent for work. So it’s all savings in terms of averted the loss of productivity, and also in case the parents need, for example, to buy additional drugs. And on the other hand, we also included the long-term cost. So, for example, injuries experienced by a child can have a serious long-term consequence. Mental health conditions, especially experience in adolescence, which is the time at which you will find that Sure Start significant to reduce hospitalizations also have very costly long-term consequences. And so when we add up all this benefits together – all this averted cost, we come up with number which is around £330 million. Now of these, approximately a little bit over 10% like £3.9 million is attributed to the direct cost saving to the NHS and the rest is from the longer-term averted cost. And given that we’ve computed the cohort, this is going to be a total cost of a little bit over £1 billion then these represent approximately a third of the spending on the programme. So taken together, the savings from reduced hospitalization offset around 31% of spending on the programme. And importantly, this is only considering health benefits. Our calculations haven’t included yet any potential benefits in other domains such as for example education, social care or crime.

Christine Garrington  20:34  

Goodness me so just looking at health benefits alone some, some really important good things that would change and, and help children grow up to live better, healthier, happier lives, but also an incredibly cost-effective programme. One which really doesn’t sort of exist in the same way anymore. So what are the key takeaways for policymakers here would you say? 

Gabriella Conti  20:56  

Yeah, unfortunately it’s not been good in my opinion, that Sure Start has been dismantled so quickly. So first of all, I think a key takeaway. So one key message for policymaker and not that this is a hard one to deliver. Because there are policy cycles and usually policymakers work based on these policy cycles but good programmes might take time to deliver their benefits, but it’s crucial that when a policy is decided, it has to be looked at the evidence and also it has to consider the long term benefits not only the upfront cost. Another important lesson I think, is that what probably made Sure Start so successful. Well, there are different components. I think one key component is to have everything in one place and to make it easier for the parents to access it. So having this suite of possible interventions and providing a place for parents to congregate and have access to resources. I do believe it’s key and it’s also the case that if you provide a variety of services, they are you know, their combined effects is not only the sum of the different ones, there are interactions and there are synergies. Among them. So now, you know unfortunately, Sure Start does not exist anymore. So there is this family hubs which are being shaped up and at least my hope is given that Sure Start no longer exists all these lessons will be considered into the shaping on the family hubs. And a particular part in this regard, there is a lot of talking about proportion at universities. Now one key finding that we have is that all the benefits that we find, for Sure Start are concentrated in the poorer neighbourhoods. So we don’t find the particular benefit in the richer ones. And so this is important why? Because on one hand Sure Start did help to reduce inequalities but also because going forward an important lesson for the new services such as family hubs is that the model combining universal services with a narrow base focus of disadvantaged neighbourhoods can be a successful approach to earliest interventions.

Christine Garrington  23:10  

Inequality of socio-emotional skills: a cross-cohorts comparison and The health effects of Sure Start are research part funded by DIAL’s Growing-up Unequal? The Origins, Dynamics and Lifecycle Consequences of Childhood Inequalities (GUODLCCI). More information is available on the DIAL website. Thanks for listening to this episode of our podcast, which was presented by me, Chris Garrington and edited by Elina Kilpi-Jakonen.