Agnes Kiragga (Head of the Data Science Program at the African Population and Health Research Center) and Kara Hanson (NIHR Director of the Global Health Research Programme and Professor of Health System Economics at the London School of Hygiene and Tropical Medicine) recently joined Anne Wozencraft, HDR UK’s Director of International and Global Partnerships, to talk about equitable research in the global context. The discussion focused particularly on equitable research within health data science, highlighting both the challenges and opportunities of moving from principle to action.
Kara, how would you explain or describe equitable research in the global context?
“Equitable research has three key dimensions. First, research questions should be shaped collaboratively and reflect the needs of the people most affected – whether locally or globally. Second, equity must be reflected in the research team and partnership itself, acknowledging that expertise is distributed across all partners and valuing the expertise wherever it exists. It also means thinking carefully about who gets visibility and recognition; from authorship on papers to opportunities to present high-profile findings. Third, equitable research requires meaningful community involvement throughout the entire process – from shaping methods and guiding access to influencing how findings are shared. Crucially, the communities involved should also benefit from the research.
“We have advanced these principles at NIHR by making equitable partnerships and inclusion explicit in funding requirements, and by ensuring funding committees give these criteria proper attention. This has been an interesting challenge, as funding committees are often heavily focused on the science and may not interrogate how research will be carried out through an equity lens. We’ve piloted giving designated committee members a specialist focus, such as equitable partnerships, capacity strengthening or community involvement. Bringing these issues explicitly into committee discussions, by making someone responsible for assessing them, has helped highlight parts of applications that might otherwise be overlooked.
We have also developed a community engagement online course, hosted by The Global Health Network, which outlines effective community engagement at each stage of a research partnership and how to budget for it. In addition, we have published a practical Global Health Research equity and inclusion toolkit, which includes guidance on resourcing inclusive research.”
Agnes, how can we best translate equitable research principles within the health data science context?
“In this particular context, we must focus on the fundamentals – such as the quality of the data, who owns it, and the specifics of data sovereignty, as well as the algorithms used in analysis. Many tools rely on algorithms not tested on the relevant data sets. It’s no longer just a case of ‘garbage in, garbage out’ – it is ‘garbage in, disaster out’. If a tool is built on the wrong algorithm, it will fail to do what it is intended to do. Technical infrastructure is equally critical: without sufficient computing capacity or infrastructure, processing power and cloud credits, projects can’t be delivered. From the outset, equitable data science projects must assess institutional infrastructure and technical capacity where projects will be delivered. Sustainability must also be considered from the outset: for example, by asking questions about who will maintain datasets, secure computing environments and specific tools over time.
Moreover, equitable global partnerships require clear agreements on data sovereignty, data sharing, and how systems will be put in place to support data and data use locally. Many institutions now operate under strong national data protection policies, which must be reflected in project design. In some partnerships, federated platforms and shared computational support have helped make this possible, underpinned by robust data governance and data sharing agreements.
Is there a project that you feel demonstrates real progress?
The Data Science Without Borders Initiative, which is running across Cameroon, Senegal and Ethiopia, demonstrates progress in this area. Two years ago, we began working with organisations at very different levels of data maturity – from an institution with decades of research experience to a public hospital with no research department. The starting point was how could hospital records be used safely for research that responds to the local needs? Since then, we have seen a transformation. Doctors and clinicians are now developing locally led research questions, based on local data, and a stronger research culture is beginning to take shape. Crucially, we have created infrastructure that will support future local projects. This progress has only been possible because of trust: trust in who is leading the work and in how the data will be used.”
Agnes, from your perspective as a leading researcher in the field, how can equitable research principles best be embedded within the health data ecosystem?
“Empowering researchers to know that when research projects are trying to solve local problems, they should be at the forefront of discussing how it should work and in a manner that will benefit the communities where the research is located. When discussing open science, we should ask how open is open. You don’t want to be in a situation as a researcher where you are being asked to share data that you haven’t analysed yet or where you no longer have any control of the data set. There’s no reason why you should accept such conditions. Another example would be if a potential project is looking at a non-priority health outcome: in such a case, it is right for the researcher to ask whether the research can be modified so that it speaks to the needs of that community.
We need to guide researchers to really take the lead and discuss such matters from the outset. It’s no use being shy about needing a server, a sandbox to innovate, or to train x number of PhD students in this project to have continuity and capacity.”
Finally, Kara and Agnes, how do you both go about monitoring whether research initiatives have truly been equitable and inclusive?
“From a funder’s perspective, we ask award holders to tell us about how their research engages with marginalised communities, how that engagement is influencing the research process, and to report on innovative practices that share power more equitably.
At NIHR, we monitor diversity across all parts of the process, including committee chairs and membership, but there are challenges in the global context. In UK domestic funding programmes, diversity monitoring has evidenced disparities in funding outcomes, and so this data provides an entry point to try and make processes more inclusive. That is much trickier in the global context because the constructs we apply in the UK (centred around the UK Equality Act) do not make sense when we are talking about global majority populations. For example, it doesn’t make sense to ask what proportion of researchers and a research team in Kenya are black, because that’s not a marker for risk of marginalisation. There are also important issues of trust, confidentiality and interpretation when collecting data on sensitive characteristics in contexts different from the UK, such as sexual orientation in countries that can place individuals at real risk. But if we don’t try new ways of measuring these issues, we won’t make progress.”
Agnes: “I agree that is difficult, but it needs to be done. As a research consortium in Africa, we now have systems in place to ensure that data used for data science projects is accessible, good quality, and most importantly, it adheres to institutional data sharing guidance and frameworks.
We are also careful to recognise the contribution of different teams across the whole pipeline of a data science project – from those involved in project design, research question formulation, data collection, data labelling and technology solutions, through to those that will develop the final products from any data science project and the IP for the product. We take care along the whole pipeline to ensure that we don’t allow anyone’s input to be forgotten.”
We are grateful to Dr Agnes Kiragga and Professor Kara Hanson for sharing their insight and experience, and for helping to spotlight what equitable research looks like in practice. Our discussion highlighted the need for organisations, funders and researchers alike to embed equity across the full research cycle. These examples of both principles and actions taking shape, in spite of the challenges outlined, illustrate what is possible.
