From oral rehydration therapy to vaccination, antiretroviral drugs and insecticide-treated bednets, the fruits of biomedical innovation are saving millions of children’s lives each year.

These successful interventions generally derive from a straightforward strategy of innovation: Target a single disease or risk factor by inventing a device or drug to fight it. To successfully mobilize science to help more children survive and thrive – especially those who do not have access to current ‘solutions’ – we need to change how we innovate by moving out of the laboratory and working more fully with resources, people and challenges in the real world.

The need for integrated innovation design

Converging evidence from neuroscience, epigenetics and developmental biology all suggest that the very risks that threaten children’s lives – poverty, malnutrition, social exclusion, infectious diseases and violence – also jeopardize the physical and mental development, and hence the future, of the children who survive.  These threats operate persistently and through multiple channels, limiting the impact of interventions targeted at a single risk factor.  For example, researchers are finding that vaccination efficacy is blunted in developing countries by diverse inhibiting factors including micronutrient malnutrition and enteric pathogens. Moreover, the effects of multiple risk factors tend to accumulate in early childhood and then cause damage throughout the life cycle. We need to develop solutions that acknowledge the multiple sources of both adversity and protection in a child’s environment.  We call any such solution – which may involve combining and adapting existing tools or designing new ones –‘holistic’.

Meanwhile, the question of how to deliver solutions to places where the need is greatest and assure acceptance and uptake there remains sorely underexplored, with grave consequences for vulnerable children.  We need innovation to create ‘delivery’ strategies that respond to local needs by bringing to bear the knowledge and capacity of all relevant actors, from village mothers to multinational businesses.

The overall concept of combining social, scientific and business innovations for better, sustainable results at scale is known as ‘integrated innovation’. Our efforts to apply this concept in practice have suggested ways to design collaborations that help innovators generate the holistic solutions and complete delivery strategies needed to ensure that children survive and thrive.

Sketching a proposed design

Our proposed design has two main components, which we call ‘co-creation’ and ‘fast-cycle learning’:

1) Co-creation: Generating solutions by jointly mobilizing both scientists and each child’s community (including his or her family).

2) Fast-cycle learning: Building the capacity of the system as a whole – scientists, communities, families, and public and private sector partners – to learn and adapt quickly based on what’s happening on the ground.

This design is neither top-down nor bottom-up, but rather horizontal: a way to combine, on an equal footing, local and traditional knowledge and capacities (see Community in Figure 1 below) with emerging contributions from the fields of biology and medicine (see Scientist in Figure 1).

Figure 1: Co-creation


The co-creation model sketched in Figure 1 is also the key building block in our approach to fast-cycle learning. Once scientists and community members have designed an intervention in partnership, they can continue to learn from each other as results come in and can then adjust the programme design accordingly.  Participants in this approach report that they find working together this way – in their words, ‘on the same level’ and ‘without hierarchy’ – to be both motivating and productive.

The examples below highlight our experiences with innovations serving young children that illustrate the core co-creation model as well as ways of building on it for greater impact and reach.

Example 1: Scientist-Community Partnership: Community empowerment lab, Shivgarh, Uttar Pradesh

A physician-led research team engaged with local community leaders and members to improve newborn survival.  Rather than starting with a package of proposed changes, they began by agreeing on the shared survival objective and then worked together to align existing cultural values, traditions and beliefs – e.g., ‘evil spirits’ and the ‘evil eye’ – with scientists’ understanding of causes of infection.  On this basis, they co-created a package of preventive behaviour changes – e.g., clean delivery, skin-to-skin care and early breastfeeding – that fit the resources and worldviews of local birth attendants and mothers in a unified, easily understood way.  In its design, the intervention was both holistic (i.e., it addressed multiple causes of infant deaths) and responsive to local delivery challenges (it was designed with and for local providers). Uptake of the behaviour changes was strong. For example, 85 per cent of newborns in the treatment communities received skin-to-skin care in the first day of life, compared to 10 per cent in nearby control communities. The results were impressive: After 16 months of work in the treatment communities, newborn mortality was 41.0 per 1,000 live births, compared to 84.2 in control communities.

Toward impact at scale

An intervention co-created as described above will be, by design, sensitive to the local context.  So to work out how to scale up such an intervention and extend it to other places, the scientific team can form additional community partnerships and jointly adapt both the intervention content and the delivery model to multiple new settings (see Figure 2):

Figure 2: Scaling Up


Example 2: Researcher as Developer with Partners in Diverse Settings: Jamaica and worldwide

A research team at the University of the West Indies piloted a community-based package of low-cost toys and picture books that local home visitors could use with mothers and their young children to stimulate learning and play. The team then adapted the materials and ‘curriculum’ across languages and cultures, with successful clinical trials in Bangladesh, Colombia and Jamaica.  Especially encouraging is evidence from these trials that this low-cost stimulation intervention can protect severely malnourished children against the disruptions in cognitive development that malnutrition generally causes.  This intervention depends on a key scarce resource: the time of community health workers.  Therefore, innovators in Bangladesh and Colombia are now running trials on alternative ways of using these workers’ time, in order to develop sustainable delivery strategies for scaling-up to reach all the disadvantaged communities in their countries. The Jamaica team is drawing on these and other ongoing local experiments to create a web-based platform to facilitate adaptation and uptake in other parts of the world, with the ultimate goal of benefiting vulnerable children everywhere.

Figure 2 describes a design for adapting and scaling up a single intervention.  But as we have noted, children in needy communities generally face multiple threats to healthier development.  Under those circumstances, a well-tailored combination of interventions may be needed, as illustrated in Figure 3 below.  Here the community acts as a ‘solutions integrator’, working with a variety of teams of scientists to pull together a custom solution. This innovation structure parallels the way developers and solutions integrators often work in the information technology business, for example.

Figure 3: Toward an integrated solution


Example 3: The Community as Solutions Integrator: The New Haven MOMS Partnership, in Connecticut

A multi-neighbourhood initiative in New Haven, Conn., has developed a strategy that draws on the experience of communities with Lady Health Workers in Pakistan. New Haven’s Community Mental Health Ambassadors (CMHAs) are neighbourhood mothers, peers of those being served. They identify families’ needs, participate in service delivery, suggest new strategies and provide feedback to modify existing ones.  They are members of both the neighbourhoods where they live and work and the citywide Guide Team that integrates feedback from all sources and takes decisions for the partnership.  The Guide Team includes representatives of all participants in the partnership – local businesses, social service groups, CMHAs, government agencies and researchers.

Thus, MOMS has a structure that can a) identify needs at the family level; b) draw on a range of local resources to craft solutions; and c) bring service shortfalls and unmet needs to the fore.  But MOMS has also been able to link the needs it identifies locally with national sources of innovative strategies: In 2012, MOMS joined a national learning network on early childhood science, policy and practice.  Stimulated by the ambassadors and the unmet needs they highlighted, MOMS has identified and adapted interventions from throughout that network, ranging from video coaching to stress reduction to building social capital and employability skills.

A major unaddressed issue identified by CMHAs was stress in mothers’ lives.  The MOMS Partnership adapted an eight-week cognitive behavioural therapy intervention, structured and presented it as a stress management course delivered in a group setting, and trained the CMHAs to co-lead the groups with a clinical psychologist.  A large clinical trial is in progress, and the work is already showing promise, e.g., in the level of uptake:  72 per cent of enrollees are completing the course (attending at least six of eight), whereas comparable rates for delivery outside the MOMS context are less than half that.  MOMS has now begun work with scientists at the University of Oregon to adapt an individual video coaching programme to a group setting as a follow-up to the cognitive behavioural sessions.

The MOMS Partnership has a strategy, illustrated above, for discovering and addressing unmet needs. This emphasis on what remains undone, rather than on a programme’s successes, also drives the fast-cycle learning strategy now being piloted in Washington State.

Example 4: Fast-Cycle Learning: State of Washington, United States of America

In 2012, research and development teams from four universities formed an ‘innovation cluster’ with five community-based agencies in Washington State.  Working with a set of common metrics and a shared overall theory of change that emphasized executive functioning and related cognitive skills, the teams have co-created and field-tested multiple interventions in diverse settings. Keeping the various scientific and practitioner teams connected enabled them to mix and match strategies to meet multiple needs.  Each pairing of scientists and community agencies began by asking what was happening with families whose children were not reaching desired outcomes.  They then co-created interventions to help those children.  No single trial could suddenly solve the identified problems.  Therefore, rather than wait for results from one large, multi-year test, the teams used small, quick trials so they could keep building on initial results – including initial failures.

For example, Childhaven, a therapeutic day-care programme for highest-risk children, is working with researchers from Berkeley on experimental games for children between the ages of 4 and 5, with the aim of building executive functioning (EF). After an initial 10-week randomized microtrial, approximately half the children enrolled in the programme showed major gains in cognitive flexibility, a core EF skill.  But surprisingly – since the game design targeted EF in general – those children showed no gain in selective attention, a second executive functioning skill.  By analyzing detailed data on the different groups of children, the team produced hypotheses on why the second group of children had no benefits, and why the first benefited from only one skill.  They are now running a second fast trial, which mixes in a mindfulness intervention already being tried in Washington by a different scientist-agency team in the cluster.

Even disappointing results have been constructive for the Washington cluster because participants have fostered a shared culture of learning, a ‘community of practice’ committed to discovering solutions for all children. Such communities generally require an engaged sponsor and convenor, a role played in this case by the Washington State Department of Early Learning.

Figure 4: The Washington State innovation cluster

Figure4a Figure4b


The examples discussed here show that researchers and community members can collaborate effectively, on an equal basis, across long cultural distances.  These are early days for results, but the emerging strategies have shown striking local uptake and real promise as part of the larger effort to make holistic, resource-appropriate solutions available to the world’s most vulnerable children.  These strategies emerged not by accident, but by design.  Key principles include:

We call on scientists, communities and stakeholders in all sectors to work together so that the potential of the generation now being born is not wasted.  The innovators described above have made progress by paying sustained attention to what’s not working, and by keeping an eye out for surprises.  As more innovators engage from lab to village, we expect their work will continue to surprise us, helping children have better lives.


The authors wish to thank the researchers and communities who collaborated on the innovations described in this essay, friends who commented on the text, and colleagues who have been with us on the learning journey. Among others: Jane Abrams, Christoph Berendes, Raquel Bernal, Silvia Bunge, Russell Eisenstat, Philip Fisher, Jason Gortney, Jena Hamadani, Lorri Hope, Annmarie Hulette, Rebecca Jaques, Aarti Kumar, Vishwajeet Kumar, Liliana Lengua, Alicia Lieberman, Linda Mayes, Denis Pelli, Ximena Peña, Christine Powell, Joseph Rotman, Holly Schindler, Elizabeth Segal, Jack Shonkoff, Peter A. Singer, Megan Smith, Fahmida Tofail, Susan Walker and Tassy Warren. The innovations were supported in significant part by Grand Challenges Canada (funded by the Government of Canada), the Washington State Department of Early Learning, and the Center on the Developing Child at Harvard University.

Click here for additional background, references and information from the authors.