Find out about the design principles and learnings behind our recommendations to Wellcome around ‘health data challenge prizes’ – which aim to spark ethical, equitable data innovation in key health areas
In March we wrote about the work the ODI has been doing with Wellcome – scoping a series of health data challenge prizes across the UK, sub-Saharan Africa and India. The challenge prizes are intended to spark health data innovation that is ethical, equitable, and engages with the communities affected and influenced by it. A challenge prize is a financial reward to support innovative solutions to an otherwise unfixed problem.
We kicked off by identifying four key topics to investigate in more depth. We then identified potential partners and participants, and examined issues that could be addressed through data innovation and the data landscape. The topics are:
- More effective mental health treatment for young people
- Reducing snakebite mortality
- Improving antibiotic prescription and use
- Public health risks in urban environments
We’ve now completed that work and have shared our recommendations back to Wellcome’s Data for Science and Health team.
Data Challenge Prizes for Health: a Playbook
We have also published Data Challenge Prizes for Health: a Playbook to support others interested in data challenge prizes as a way of creating ethical, equitable innovation in the health sector. Based on our experience of exploring challenge topics, the playbook provides a step-by-step guide on how to identify a suitable health problem and design a challenge prize that will achieve sustainable impact. We will iterate on the playbook through other projects, such as Data infrastructure for common challenges, part of our four-year R&D programme.
Here’s a brief summary of the recommendations we made to Wellcome, the design principles we used to get to them, and some of the learnings we’ve had along the way. Covid-19 of course is a major factor for each of these topics, but as it is not yet possible to reliably predict what the impact of the pandemic will be, the recommendations do not address Covid-19 challenges specifically.
Scoping health data challenge prizes: four topics
Data science for more effective mental health treatment for young people
Key question: How can we support the development of more effective interventions for addressing anxiety and depression in young people and adolescents?
Depression and anxiety, two of the most commonly diagnosed mental health disorders, are often seen as a catch-all definition for someone’s mental ill-health experience. But just as people are different, the way that they experience depression or anxiety can hugely vary too – one size does not fit all when it comes to interventions. Most people who experience mental health problems fully recover, or are able to live with and manage them – especially if they can get the right support at a young age.
But 70% of children and adolescents who experience mental health problems do not receive appropriate support and the generalised approach to diagnosis can affect patient trust in treatment. To be able to support young people to prevent, effectively manage or overcome their own mental health challenges, mental health practitioners need a better understanding of two things: the way that these disorders manifest through symptoms, and mechanisms of a mental health intervention that can help.
Both of these areas have been explored through traditional research methods, but there is still more to learn about anxiety and depression. A health data challenge prize could help mental health experts make more effective use of data to develop a better understanding of the patterns of anxiety and depression symptoms; and a better understanding of what makes an intervention effective.
Better public engagement could help people understand and trust that health data will be used to create positive outcomes, and could also increase trust in healthcare and research organisations more broadly.
People taking part in the challenge prize could use health data and work closely with young people to understand their needs. Teams could then build data-driven tools that help young people better manage their own mental health, or that help practitioners to make better decisions about the interventions that they prescribe and apply.
Ultimately, the tools created from this challenge could help current and future generations of young people to better manage their own mental health, or to receive the right professional support when needed.
Data science for reducing snakebite mortality
How can we use data science to reduce deaths from snakebite in India?
Farming rice in the South Indian state of Tamil Nadu poses many challenges. Poverty is already endemic, and the tropical region is severely impacted by climate change, including increasing temperatures, declining rainfall, cyclones, floods and drought. For all the threats to life and livelihood, there is perhaps one least understood by authorities but no less deadly. Anywhere between 3,000 and 10,000 people in Tamil Nadu will die from a snakebite this year. Deaths from snakebites globally are on par with cervical and prostate cancer, killing between 81,000 and 138,000 people every year.
The World Health Organization recently classified snakebite as a ‘neglected tropical disease’, with the goal of 50% reduction in deaths and disabilities by 2030. However, data on the extent of the problem is lacking. Healthcare systems, strapped for funding, struggle to supply and allocate needed medicines and expertise in the right location and quantity. Non-governmental organisations (NGOs) dedicated to prevention, education and healing – with equally limited resources – need better information to tackle the problem effectively.
A health data challenge prize in Tamil Nadu would galvanise the research and data science community to create impactful solutions from existing but disparate data sources. The challenge would draw on the research insights across healthcare, herpetology, climate, ecology and urbanisation to improve the epidemiology of snakebite.
These tools could help the healthcare system better plan a response according to the areas most at risk of snakebite, ensuring that the anti-venom and medical expertise needed in rural areas are available. They could help NGOs plan their prevention and education better. Importantly, the models created in Tamil Nadu could then be adapted to other regional contexts across southern Asia and sub-Saharan Africa, meaning the snakebite health data challenge could provide a substantial benefit for communities at highest risk.
Data science for improving antibiotic prescription and use
How can we use data science to optimise the use of antibiotics to minimise the local emergence and persistence of antimicrobial resistance in Uganda, Malawi and Kenya?
Antimicrobial resistance is one of the major public health concerns of this century. Antibiotics have been a staple of modern medicine to fight illnesses, but antibiotic treatment has become less reliable as antimicrobial resistance has evolved in many bacterial pathogens, causing the need for costly new treatments to be developed, while antimicrobial resistant organisms continue to take thousands of lives per year. If no action is taken the United Nations warn that drug-resistant diseases could cause 10 million deaths each year by 2050.
Antimicrobial resistance occurs naturally over time. However, the misuse and overuse of antimicrobials speeds up this process. Therefore, understanding how antimicrobials are being prescribed and how patients are accessing them is key to understanding and stopping the spread of resistance.
Antimicrobial resistance is a broader term than antibiotic resistance, and includes resistance to drugs to treat infections caused by other microbes including parasites (for example malaria), viruses (for example HIV) and fungi (for example Candida).
In the UK, access to antibiotics is regulated: people need to visit a doctor and get a prescription before accessing antibiotics. However, that is not the case everywhere. In some low- and middle-income countries, many antibiotics are available without prescription over the counter, in unlicensed drug stores and at markets. People buy the antibiotics they can afford, take a lot of different ones at the same time, and not necessarily for the right amount of time. This can accelerate AMR and is difficult to track as there is so little data on the prescription and consumption of antibiotics through these informal dispensaries. These issues have been observed in Uganda, Malawi and Kenya.
A health data challenge prize could play a key role in motivating people with expertise in data science and public health to work together. Using behavioural and epidemiological data from a series of current and upcoming initiatives, the teams could map the flow of antimicrobials through community pharmacies and into the population. These insights could then be used to increase understanding of the scale and impacts of resistance. Currently, much of the data about AMR comes from western sources.
These insights could be used to create tools that support antibiotic use by giving the right information to decision makers for better interventions; to healthcare providers to encourage an appropriate use of antibiotics; and to NGOs to target campaigns to incentivise better prescription and use behaviours. If policymakers and healthcare providers can start working from a shared understanding of the scale and nature of antibiotic consumption, this has the potential to support real progress in the fight against AMR which is likely to become even more vital as Covid-19 continues to spread.
Data science for public health risks in urban environments
How can we better understand and resolve public health risks in rapidly growing urban environments?
Over 55% of the world’s population live in urban areas, a proportion that is expected to increase to 68% by 2050. By 2030, projections indicate that two billion of the global urban population will live in slums, mostly in Africa and Asia. However, the health of people who live in slums is a topic that has received little attention. A person's risk of disease is affected by both personal factors, such as diet and genetics as well as environmental factors such as sanitation and pollution, aka ‘neighbourhood effects’.
Health authorities generally track information about individuals to inform decisions about public health and health services, but because of the high concentration of people in slum areas, environmental factors affect a far greater number of people at any one time. Information about neighbourhood effects is therefore incredibly valuable for identifying highly impactful public interventions, but there is very little data about these communities.
For example, poorly managed solid waste has health, environmental, and economic effects that multiply as waste accumulates. Uncollected solid waste increases exposure of all individuals in communities to vector-borne and zoonotic infectious diseases carried by birds, insects, and rodents. Over time, uncollected waste accumulates to block waterways, resulting in flooding, contaminated surface and groundwater, and emissions of greenhouse gases like methane.
Abuja is one of the fastest growing cities in the sub-Saharan Africa but the city is said to lack the modern management techniques to meet the requirement of a rapidly expanding city, often resulting in major floods that lead to significant deaths.
There is a thriving community of academics and data scientists using non-traditional data such as satellite imagery to develop tools and techniques to measure neighbourhood phenomena, for example the identification of slum area boundaries over time, or using machine learning to identify access ways and roads which can be used for specific public health risks such as solid waste management. However, these researchers may not have health or medical backgrounds and therefore may not package or distribute data with health decision-makers in mind.
A data challenge prize which closely engaged with city authorities to understand their specific municipal challenges, and the communities themselves to understand the effect of these challenges, could motivate and support data scientists working in the field of remote sensing (data collection from a distance, typically from aircraft or satellites) to collaborate directly with public health experts.
By combining remote sensing data such as satellite imagery, with administrative, municipal, health and community-collected data, the challenge teams could develop models, tools and applications that support cities to make better decisions regarding public services that impact public health in urban areas.
The nature of ‘neighbourhood effects’, enables cities to identify the specific interventions in slum areas that will have the biggest impact. This will therefore support the health of a huge population, and help equip the authorities for dealing with rapidly changing urban landscapes.
What we have learned
The most important and exciting learning from the project is that there appears to be an appetite among the scientific community to use a data challenge prize mechanism to build the field of data science in health. Discovering multiple, credible opportunities with a range of stakeholders and potential delivery partners provides a strong indication that there is room for this approach alongside other methods of innovation generation and scientific funding.
Below are more key learnings around how to scope and design a data challenge prize in health.
Address a significant health problem
Focus on health problems which can be addressed using data science but that also deliver tangible, visible impact to the wider health sector and are understandable to a wide range of potential participants and stakeholders.
Use existing data, do not focus on making new data
Where possible, demonstrate that there are already a huge number of opportunities for the use of data science in the health sector with existing data. This will focus the challenges prizes on innovation rather than creating data infrastructure.
Co-create with the community
Transparency is an important part of building public trust, but trust also requires participation. Therefore the patient groups and communities most affected by the challenge also need to be part of co-designing and delivering the prize.
Enable all participants
One criticism of challenge prizes can be that they require a large amount of free labour for the possibility of financial remuneration. In many cases, early-stage health companies – who are often the most agile and innovative – are 'pre-revenue' and rely heavily on grants, subsidies and investment.
Similarly academic, civil society and public participants often struggle to engage with prizes due to a lack of funding. Therefore, to enable everyone who has relevant experience and innovative ideas to take part, it is important to offer seed funding for all stages of participation.
Build capacity
Where possible, aim for the challenge prizes to build data science capacity across a peer network of experts by bringing communities together and providing training and mentorship through the prize. This is another way of ensuring sustainability and meeting the objective of increasing data science capability across the health sector.
Make it generalisable
The challenge prizes should be relevant to a range of places so that innovations created can be replicated or scaled elsewhere, or the challenge prize itself can be repeated in another location.
Be driven by the end user
To create things which actually answer the challenge topic in a practical and usable way, it is crucial to position the challenges around the needs of those who are directly affected by the issue.
Make the impact visible
Show the impact of these innovations through measurable change and human stories. Show the impact of both the innovations produced in the challenge prizes and the wider impacts of the programme of challenge prizes.
Create sustainable solutions
Sustainability is crucial to ensuring impact from the innovations generated through the challenge prizes. Consider how they will be adopted and sustained after the term of the prize.