• Analysis
  • November 30, 2018
  • 13 minutes
  • 1

Early intervention gives children huge benefits. But why do they often fade?

Answering this question is vital if early childhood policy is to deliver on its potential

“There is tremendous enthusiasm for early childhood now,” Sally Grantham-McGregor says, “and rightly so. It’s a very important stage.”

Grantham-McGregor can speak confidently: she’s worked in the field since long before that recognition arrived. Now emeritus professor of international child health at University College London, she led the team that pioneered a transformative program in Jamaica in the 1970s that improved the way parents played with their young children.

The promise of that Jamaican intervention, and other early childhood programs like improved preschool and childcare, is that they can have lifelong benefits — helping to overcome disadvantages that set in from the first few months of life.

But in recent years several new studies, in countries from Colombia to Malawi, have suggested that the picture is not so straightforward.

Improving childcare or parenting skills can deliver an immediate benefit. But it may well fade within a few years — and nobody is sure whether that’s because of difficulties scaling up, poorly targeted interventions or any of several other possible explanations. What researchers do agree, though, is that answering these questions is vital if early childhood policy is to deliver on its potential.

Total eclipse of the boon

The fact that children’s early years are crucial to their development is now almost indisputable. Grantham-McGregor described a particularly striking finding from Bangladesh, where cognitive gaps between children from the top and bottom wealth quintiles began to be significant from seven months of age.

By five years, the disparities were enormous. In the United States, longitudinal studies have found these kinds of gaps in childhood translating into social outcomes like incarceration rates and employment prospects in adulthood.

Early intervention is meant to stop those disadvantages accumulating. Some of the most impressive results are from Grantham-McGregor’s original Jamaican trial, now known as Reach Up, which used home visits to teach parents to play with their children in ways that improve healthy development. Adults who’d been in the program 22 years earlier earned more, had better educational outcomes and were less violent than their peers.

That success inspired versions of Reach Up in other countries. But their success has been more limited. A study conducted from 2012 to 2015 in Colombia, implementing a modified version of the program in 96 towns, found that the initial benefits disappeared two years after the program ended.

A similar “fade-out” result emerged from a World Bank trial in Malawi. It aimed to improve the general quality of childcare centres and, additionally use them to teach parents better approaches to playing with their children.

A year and a half after the intervention, development outcomes were better, although that improvement was “on the smallish side,” said Berk Özler, the Bank’s lead investigator on the study.

By the 36-month follow-up, though, “the fade-out was completely universal,” Özler said.

These studies echo even more puzzling findings from the United States. Famous programs, including the Perry Preschool and Abecedarian projects and the government-run Head Start, delivered immediate benefits which disappeared after a few years. But later assessments found them resurfacing: adults who’d been in the programs had better employment and other social outcomes.

That phenomenon, sometimes called benefits “eclipse”, only adds to the vexing question of why fade-out occurs — and what can be done about it.

An ongoing puzzle

There is no shortage of possible explanations for the problem of fading gains. Grantham-McGregor believes that one core issue may simply be the size of the initial effects: smaller benefits aren’t able to be sustained over time.

“I just thought that that benefit wasn’t enough to be sustained”

That hypothesis is not yet supported by research. But it’s a hunch which led Grantham-McGregor to correctly predict that the Colombian evaluation, which had smaller benefits than the original Reach Up program, would likely find a fade-out. “I just thought that that benefit wasn’t enough to be sustained,” she said.

Why were the gains smaller in the first place? Grantham-McGregor pointed to a raft of cost and time-saving changes from the original program that have been common in follow-ups, driven by the demands of working at regional or national scale.

But the challenge may also go beyond scaling implementation, to issues of initial program design. Özler said that the Malawian project he evaluated had been generally well-implemented. But parents were trained in particular ways of interacting with their three- or four-year-olds which were no longer adequate by the time those children were six.

Many parents had almost entirely abandoned their new habits by the time of the 36-month follow-up in Malawi — perhaps because they could see their irrelevance to older children.

Third, Özler and Grantham-McGregor agreed that another important factor is likely what happens to children after early interventions stop.

In Malawi, most went from the program’s childcare centres into primary schools of extremely low quality, with as many as 100 children to a classroom. In some Reach Up-inspired parenting interventions, children are left for a year between the program’s conclusion and the start of preschool.

But there are plenty more candidate explanations. Grantham-McGregor listed off the duration of the intervention, its intensity, children’s age at its beginning and end and how much the program targets the most needy families.

“We don’t know,” she concluded. “This is all speculative,” Özler agreed.

Fade management

Despite this deep uncertainty, there are steps policymakers and researchers can take to avoid some of the most obvious risks that contribute to fade-out, such as poor initial implementation.

Perhaps the most important is simply to make sure programs being introduced to a new country are piloted first, even if they’ve been successful elsewhere. Cultural norms mean programs can have unexpected different impacts in different contexts. Researchers in Bangladesh found that mothers responded better to parenting advice when they received it in pairs — quite different to how they’d reacted in Jamaica and other places.

“Nobody wants to pilot, when they start somewhere new,” Grantham-McGregor said. “They want to go straight into it, and they want to go to scale straight away, and I think that’s very dangerous.”

Another key to maintaining high-quality implementation is to pay attention to management and supervision structures. At large scale, details of an intervention can easily be lost — so it’s crucial to integrate new measures into the accountability framework that workers respond to.

“If you go to the clinics, staff report up the line every few weeks how many immunisations they’ve given, how many children they’ve seen, all of that,” Grantham-McGregor explained. If tasks related to a new intervention are to be prioritised, she said, they “have got to be on that report.”

More questions than answers

The overarching reality, though, is that most studies have focused on finding benefits rather than understanding how they’re achieved or what makes them last. So researchers still have a lot to learn before they can hope to seriously tackle the risk of fade-out.

Three questions are particularly important to that task.

First, how exactly do interventions like Reach Up work? One obvious possibility is that they directly improve children’s cognition, and so the way they interact with others as they grow up.

But another is that the main impact is on parents. Studies in schools have shown that teachers’ expectations for their students are a major determinant of final outcomes. If early interventions convince parents that they can make a difference to their children’s development, and to have higher aspirations for them, that itself could fuel the gains. “It may be that it’s as simple as that,” Grantham-McGregor said.

“Nobody wants to pilot, when they start somewhere new, and I think that’s very dangerous”

Second, how do small changes in program design affect outcomes? Advice to parents on playing with their children, for example, could be delivered individually or in groups, at home or in clinics, and at the age of one or three. But there’s been very limited experimentation using randomised trials to assess which would deliver most benefit. And similar questions can be asked about other types of intervention, like preschool programs.

Finally, how do early childhood programs interact with the effects of subsequent primary and secondary education? Nobel economics laureate James Heckman, author of several of the studies finding gains lasting into adulthood, has discussed a “scaffolding” model in which each intervention needs to be followed up if children are to see sustained benefits.

“Maybe the general answer is … don’t do anything like we did if you’re not really sure that the rest of the education system is in line and connects with it,” Özler said — though he reiterated the speculative nature of that suggestion.

Those questions give early childhood researchers a lot to be working on. And they could lead to a big change in the ability of policymakers to replicate successful programs and deliver them at scale.

“It’s quite an exciting time, really, but people are jumping the gun,” Grantham-McGregor said. Researchers and policymakers need to figure out how to make childhood programs scalable and sustainable. But the need for more caution, she suggested, shouldn’t detract from the importance of further investment in early intervention.

“It’s too late once they reach primary school,” she said. “There’s a terrible need — inequity begins at this age.” — Fergus Peace

(Picture credit: Unsplash)


Leave a Reply

to leave a comment.
Master the skills you need for the public service.

Discover inspiring resources, tools and policies.