Paul Tarini, a Senior Program Officer at the Robert Wood Johnson Foundation (RWJF), paused and then asked me, “have you heard the one about the woman helping her friend look for his keys under the streetlight? After a few minutes, she asks him if he’s sure that’s where he lost them, to which he responds ‘no, but this is where the light is best’”.

Paul was describing a variant of the Streetlight Effect—a colloquial name for an observational bias arising from our propensity to look for answers where it’s easiest to do so.

This analogy provides a helpful way of thinking about the predicament that, in spite of an ever-expanding global field, many health innovations go undiscovered or remain at a suboptimal scale, failing to meet their full potential. This is at a time when many countries are grappling with complex and deeply entrenched health inequalities, increasing healthcare costs and declining budgets. New solutions are needed now more than ever before.

Over the past few months, the Innovation Mapping and Health Lab teams at Nesta—in close collaboration with RWJF—have been working on this question. We’re creating something provisionally called a ‘health innovation scanner’—an online, interactive, free and openly accessible database, search and mapping tool that is capable of illuminating new swaths of the health innovation landscape for funders, policymakers, researchers, investors, innovators and everyone in between to see.

The direction the light will shine is guided by people active in the health innovation landscape who have opened up to us about what kinds of questions a tool like this might help them to answer. The source of light is a combination of new data sources, powerful analytics, and compelling visualisations.

Below we’ve described some of the ‘dark spots’ that have emerged thus far that we hope to illuminate, as well as some early explorations of the data landscape. We’re also encouraging people active in this space to reach out to us and share their views. In a world where we’re often confronted with too much or too little data—or data that simply don’t help us to answer the questions we need answered—we want to ensure that we’re building a tool that is relevant, useful and timely.

What exactly is a ‘health innovation’?

Before delving deeper into the details, we must make a quick detour to answer an important and inevitable question—what exactly is a ‘health innovation’? A pass through the literature quickly reveals that health (or healthcare) innovation covers an expansive range of topics, institutions, actors and theories. Some authors have even argued that the overuse and consequent dilution of the term has caused it to lose meaning.

We decided to come up with a definition to guide our work in this project that reflects the shared interests of Nesta and RWJF. The work of Nesta’s Health Lab centres around ‘people- and data-powered health’, a concept that aims to make the most of diverse resources from throughout the system alongside the knowledge and commitment of health professionals. RWJF’s work is centered around a long-term, equity-oriented framework to build a ‘Culture of Health’. Drawing from fundamental aspects of both, our definition is:

New approaches that enable individuals and communities to take more control of their health, [and/or] generate and make use of new sources of data and/or digital technologies, [and/or] improve how health and care is made available and delivered. These innovations should lead to fairer and more equitable health outcomes.

The definition makes a deliberate choice to focus on health innovations that have typically been difficult to measure and have perhaps received less attention in the past, but we feel are vital for (and will fundamentally shape) the future of health and care systems.

The direction of the light: What questions might the scanner address?

Interviews with staff from within Nesta and at RWJF have been essential in framing our understanding of the health innovation landscape and have provided ample food for thought on what kinds of questions the scanner might address under ideal circumstances:

• Digital innovation: Health tech startups are developing artificial intelligence (AI) systems that can rapidly identify and address health issues and ensure more timely and appropriate care. Blockchain has migrated from the financial sector to healthcare and is poised to transform medical data management, including drastically altering data sharing across healthcare providers and personal data use in health research. How can we pick up on these and other emerging trends in digital health innovation? For example, can we see how AI or blockchain are being used in health? Or, more ambitiously, can we identify what the next important trend similar to AI or blockchain will be?
• Social innovation: Social innovators, public sector organisations and civic society groups are experimenting with social models of health—from peer support to neighbourhood networks and other ways of building social fabric and enabling communities to thrive. How can we identify projects, programmes and policies that sit at the intersection of health and its social determinants, or at the nexus of other relevant sectors (e.g. social care, housing)?
• Reverse innovation: Health innovations developed in lower-resource countries can be both unique and powerful by virtue of the constraints within which they were created. However, promising health innovations can fail to move beyond the countries where they were developed, minimising their impact. How can we pick up on promising innovations outside of our own borders that we might ultimately be able to adapt to our own context? Can we develop a deeper understanding of the health innovation landscape in countries where English is not the mother tongue?
• Innovation and societal needs: How does the current funding of health innovations relate to societal need, for example disease burden or particular challenges faced by health systems in different contexts? And what needs to change?

Are these ‘dark spots’ for you as well? What else would you want to learn about the health innovation landscape if you could leverage new data and analytics?

The light source: How can we use new data and analytics to explore the health innovation landscape globally?

There are many new and potentially relevant data sources that could yield helpful insights into these questions, opening up a wide range of exciting pathways to explore. For example, we could explore trends in open digital health innovation in GitHub—a collaborative, open source software development platform. Our initial exploration of 3.6 million GitHub projects in the EU revealed 2,000 focused on health, including 40 that use machine learning or AI.

Another possibility would be to use CrunchBase to analyse trends emerging in the private sector, drawing from rich data on nearly 500,000 predominantly startup and tech companies in 199 countries.

Once we’ve identified promising data sources, our next step is to carry out a series of ‘pilots’—short, focused mini-projects that allow us to quickly determine whether a particular data source can provide us with useful outputs. This approach provides us with the information needed to abandon ‘dead ends’ or to build on promising leads.

Our first pilot aims to set the groundwork for the project by developing a rich and detailed picture of the current landscape of health innovation funding activity globally. To do this, we’ve started scraping and extracting data on research and project funding in the US, UK and globally, drawing from data sources such as those compiled or provided by Innovate UK, Gateway to Research, Grand Challenges, and USAID. From this ‘seed’ of known health innovation activity, we can then use the Google Search API and academic databases to identify relevant connections and collaborators, expanding the initial list and uncovering patterns and trends that were previously hidden.

In the coming months, we’ll be developing additional pilots in that focus on AI in health, peer support, as well as exploring the relationship between societal needs and innovation activity.

This project is both innovative and ambitious. It builds on previous work carried out by Nesta to map innovation but it also involves working with new data sources and experimental data science methods. We will therefore seek to balance ambition in the scope and depth of the scanner’s coverage and robustness in completeness and richness of the insight generated.

We look forward to sharing more developments from this exciting project as we move forward.

---

Sincere thanks to all who participated in an interview, those who have provided input to the early stages of the project, and to those contributed to the development of this blog post, including Juan-Mateos Garcia, Minji Song, Richard Gregory and Isabella Blatter. Thanks also to Nesta’s Annie Finnis and Halima Khan for some of the background material that was used to develop the blog post.

Interested in getting involved?

We’re putting together an external stakeholder group that will provide targeted, timely feedback at key junctures in the project. If you want to participate or have any questions/comments, reach out to me at [email protected].