This piece is part of Apolitical’s spotlight series on scaling social impact, in partnership with the Bernard Van Leer Foundation. It also appears in our government innovation newsfeed.

The early childhood intervention “Reach Up” was first conceived in Kingston, Jamaica, nearly 50 years ago. Now it’s been trialled in 10 countries, rolled out as a national policy in Peru and seems to be on the cusp of government adoption in two more: Bangladesh and, at long last, its country of origin.

Reach Up started from the observation that children born into poverty develop very poorly. They’re disadvantaged in many ways: not only are food, shelter and healthcare scarce and uncertain, but their parents are often so focused on providing these things that they don’t play with their children. That might sound like a relatively minor omission, but it’s crucial. Because playing with a child is instrumental for the development of their brain.

Reach Up was one of the first programs to really drive this home.

In its initial incarnation, it involved sending doctors and nurses to visit mothers every week in their homes, bringing toys and books and showing the mothers how to play with their child to improve their development. That way, rather than learning just once a week when the nurses visited, the children would learn from their mothers all the time. The visits went on for two years.

The resulting studies found that children whose mothers received instruction from the doctors and nurses made significant developmental gains. But the really remarkable thing was how those gains lasted.

In 2014, a study by the Nobel-prize winning economist James Heckman followed up on one of the earliest cohorts of children to go through the program. Twenty-two years earlier, those children had been through the Reach Up program, as delivered by trained visitors, but not nurses or doctors. He found that they had long-term improvements in cognition, did better in school, and were less likely to be involved in violent crime. They also earned more money: a full 25% more than equivalent peers without the treatment, and the same amount as their more advantaged peers.

In short, Reach Up seemed to compensate for early developmental delays, narrow the achievement gap and reduce inequality later in life.

“That trial shows the potential of the intervention,” said Professor Susan Walker, who now leads the Reach Up research team in Jamaica. “But we have to be a little cautious about extrapolating to what could happen at scale up. Because scale up of necessity will change things.”

Scaling and changing

The challenge for the Reach Up team was clear: they had to take a small, tightly-controlled and researcher-led program and turn it into something that governments could deliver effectively and cheaply at the national level.

The first thing to do was to make the program more cost-effective.

The researchers started using homemade, recyclable toys. They devised a structured curriculum that could be delivered by community health workers rather than highly trained professionals. And to ensure that quality was maintained, they introduced occasional supervision for the home visits.

Then, through a series of trials, they tried various changes to the frequency of visits and the duration of the program. “We found a kind of dose response: the more intense the visiting, the bigger the benefit to the child,” said Professor Sally Grantham-McGregor, one of the original designers of the program. This meant that weekly visits were ideal — but perhaps not absolutely necessary.

Nowadays in Jamaica, a year-long Reach Up program with weekly visits costs around $245 per child — still far from cheap. Calculating the benefits of early childhood interventions is tricky, but Reach Up's are conservatively estimated at around $928 per child. That gives it a benefit-cost ratio of around 3.8.

Having reduced the cost, the next challenge was to reduce the reliance on the core team of researchers in Jamaica. To this end, the team codified the program in a series of digital manuals.

At first these were freely available online, but those designing the revamped program came to decide that more control was needed to ensure fidelity to the original. “The manuals are available at no cost, but early on we decided to insist that the initial planning and adaptation, and the initial training of supervisors, was done involving someone who was experienced in the program,” said Walker.

One of those manuals is dedicated to adaptation, and Reach Up has now been adapted to ten countries and counting, from Bangladesh and China to Brazil and Colombia.

“When we take Reach Up to other countries, they want to make it their own,” said Grantham-McGregor. “And you want them to do that, but at the same time you want fidelity to the original. Where do you draw the line?”

Some things, like the toys and materials, are always adapted. “You have to think hard about the child’s environment and reflect it back to them in the materials,” said Grantham-McGregor. “We run a survey to collect all the local games and songs and toys, and we include those. For instance, finger and toe games. In England they have 'this little piggy went to market…' And there’s an equivalent in almost every country I’ve been in.”

“If you look at the books available in the shops, they’re useless,” added Grantham-McGregor. “Things like birthday parties with white middle-class kids and talking bears. Ours are very simple, and they seem to work.”

Beyond just adapting the materials, the Reach Up team look for opportunities for family interaction in a country's culture. “For instance, in Peru sometimes they have a story time at night, when the family sit around,” said Grantham-McGregor. “The mothers also carry the kids around on their backs for long periods. And in Jamaica they plait the children’s hair.”

Reaching out to Latin America

Reach Up has a strong evidence base: in fifteen trials it has always had a significant, positive impact. But most of these trials have been at a relatively small scale. Latin America is the exception.

In Colombia, the government wanted to trial an even cheaper version of Reach Up. It cut costs by piggybacking the program on the infrastructure of an existing cash transfer scheme. People receiving those cash transfers were recruited and trained to deliver the program as home visitors, and the frequency of supervision for the home visits was slashed. Inevitably, the quality dropped.

“In Jamaica we used to supervise every week, but in Colombia they were doing it every nine weeks,” said Grantham-McGregor. “And for the visitors, the only requirement was for them to be literate. It was a very low hurdle to climb.”

In the end, they did see a significant, positive impact, but a much smaller one than in Jamaica. And, two years later, the impact seems to have faded back to nothing.

The cost-cutting in Colombia went too far and sacrificed the program’s effectiveness. Nonetheless, it did show the program could be delivered at scale. And that’s what caught the eye of Peru’s government, which had promised to deliver early childhood services to its rural population. Cuna Más, the development program it started in 2013, was the result.

Cuna Más scaled at breakneck speed. By December 2016, the latest figures available, it was reaching 85,000 families — approximately 32% of the target rural population. It has to hit coverage targets every year.

The first evaluation took place in 2015. It showed positive and significant impact — but even smaller than in Colombia. “The impact in the evaluation was relatively modest, but it showed a lot of promise,” said Dr Marta Rubio-Codina, who helped analyse the trial. “But implemented at this scale, by a government agency, you lose control of many elements.”

One issue was the sheer speed of the roll out. There was no time to stop and consolidate. There were government targets to hit every year. This meant that the program “kept developing as it was rolled out,” said Rubio-Codina. “So the Cuna Más program that I saw is not necessarily what is now happening in the field, and I cannot assure you that the same programme is being used in every area.”

Two other issues have eaten into the quality of delivery: the sheer size of the organisation required to support it, and the changing priorities of government.

The program involved a central team in Lima, regional offices, supervisors who train the visitors, and the visitors themselves. The more layers of management, the harder it became to ensure quality on the ground. “That’s been a big change since Jamaica, when it was Sally, her colleagues, and a handful of home visitors,” said Rubio-Codina.

Then, the same thing happened as in Colombia, only in reverse: other programs started piggybacking on the infrastructure of Cuna Más. A new government came in that saw chronic malnutrition and anaemia as priorities; for them, the home-visiting strategy of Cuna Más was an opportunity to deliver medicine and advice. Now, more and more time in the visits is given over to those things, at the expense of the Reach Up curriculum.

Several lessons emerge from Colombia and Peru. The first is to scale slowly and in steps. The second is that you can only cut costs so far before you undermine a program. And the third is that you never get something for free by just using existing services and their infrastructures: there is always a cost in quality.

Back in Bangladesh

It’s clear that governments that need Reach Up feel unable to afford it. The model remains too expensive, but to cut further costs without compromising impact it is necessary to identify what really makes the program work, and what therefore must be preserved.

In Bangladesh, a different model of Reach Up has emerged that may shed some light.

It operates out of clinics, thus eliminating the need for home visits. That also means that the program can be delivered by health workers with degrees, which reduces the need for supervision. It also brings mothers in two at a time, rather than doing one-on-one visits, and cuts the frequency of visits to once a fortnight.

The results aren’t published yet, but the impact looks to be on a level with the early studies in Jamaica — and at a much larger scale.

Compare this delivery model to the original one in Jamaica and they are strikingly different. It’s true that not everywhere would be able to use this model — it relies on existing clinics — but it highlights just how many elements of a program can be flexible.

This begs the question: what are the inflexible, essential elements of Reach Up?

“I can tell you what I consider to be the essential elements, but I don’t have the evidence,” said Grantham-McGregor. “Working with mother and child is critical. The secret is to get the mother onside and to boost her confidence. To make her realise that she can change a child’s development.”

“I think the materials are important, but other people will tell you not. The children look forward to the visit because we bring something different every week. It’s like a toy library. The children greet you at the door, pulling at your bag: ‘What’ve you got?’" Grantham-McGregor added. "But I think supervision is the key.”

In Bangladesh, supervision was not so necessary. But everywhere the program is delivered by non-experts, it is vital. It’s the only way to ensure that the chain of learning, from visitor to mother, and from mother to child, is preserved. And that’s the essence of the intervention.

Reach Up isn't limited to the Global South. Here’s how it is being adapted to the US.

(Picture Credit: Christina Xu/Flickr)

Tom Graham