Over the last six months, we’ve talked to hundreds of leading researchers across 30 countries about the opportunities AI can bring to health data – from helping to detect and diagnose disease more effectively, to speeding up drug discovery, and delivering better care. We’ve also heard about the barriers that hinder innovation in the field.
An environment where data innovation can flourish
The AI technology itself is not the limiting factor. A significant amount of funding is directed to AI projects near clinical application.
However, for data innovation to happen, other challenges earlier on in the process need to be considered, such as:
- unrepresentative datasets – black and minority ethnic groups are more likely to opt-out of the use of their health data for research. As a result, these groups will become invisible and the algorithms developed using these datasets will simply not work for everyone.
- a lack of public trust – data protection rules are often seen as complex. This makes data-holders become risk-averse, so they will often restrict access to the data for research.
- health data can be messy and complex – it can take many months to prepare; this often deters data scientists from working with health data, and even drives them to leave academia for highly paid jobs in the tech sector.
- a lack of data science capacity worldwide – some countries in Africa do not have data scientists to address local health priorities.
What we want to achieve
The data, the tools, the governance, the skills and the public trust are crucial for unblocking data innovation – but these ‘enablers’ often get under-valued and neglected.
This is where our new programme on data for science and health can have most impact. More specifically, we aim to:
- engage society to build people’s trust, understanding and participation in health data innovation – we want to ensure the public are an important part of the solution, for example by building on the successful work of Understanding Patient Data and supporting more effective community engagement with diverse groups.
- equip and motivate data scientists to innovate with health data for public good – we want to tackle the difficulties of working with health data, by funding the development of open source tools.
Our work will go beyond the UK. We want to make sure that low and middle-income countries are equipped to participate – and we’re planning to start by looking at hubs in Southern and East Africa, and in India.
Wellcome is ideally placed to make a difference. We can bring a cross-discipline approach to the challenge – including ethics, advocacy, behavioural change, education and public engagement – in addition to our expertise in funding research and innovation with health data.
But we cannot do this on our own. It will be essential to work in partnership – for example with Health Data Research UK and the Open Data Institute in the UK, and internationally with organisations such as the European Bioinformatics Institute (EMBL-EBI), the Massachusetts Institute of Technology, the Gates Foundation, the Tata Trust and the African Academy of Science.