Last summer, my Laidlaw research project asked how equal, or unequal, access to life support is across the UK. ECMO is a form of life support that takes over the work of the heart and lungs when they fail, and the question looked simple on the surface: where were patients referred from, when, and how much did access vary depending on where a person happened to live? This work was possible because the data, although often poorly curated, is present.
For my Leadership in Action placement, I wanted to experience what happens where the infrastructure to support those analyses is disparate, underdeveloped, or simply absent. Africa is home to about 18% of the world's people and carries roughly 25% of its disease burden, yet contributes under 3% of the world's clinical trials. The reasons are many but one is self-reinforcing: without published local evidence it is hard to prove the need for trials, so Africa is repeatedly passed over for research funding and trial support.
During my placement I worked with the Africa Clinical Research Network (ACRN), a pan-African research network running active studies in six countries on the continent, with partner organisations across fourteen. I spent the placement supporting the development of their new project, REMAP, the Real-World Evidence Mapping and Analytics Platform for Africa. REMAP is an effort to build the capability and infrastructure to turn routinely collected electronic health record data into real-world evidence, and through it to support research, attract funding and improve care. The premise is fairly simple, but as one might imagine the execution is not. The data already exists inside hospitals; the work is making it findable, comparable and usable, and doing so in a secure way compliant with different institutional and national governance, regulatory and data protection frameworks.
As a young organisation, ACRN works at a very fast pace; it was clear that being a self-starter and taking the initiative was crucial both to making the most of the placement and to being useful to ACRN in the short time I had. After four years in environments with real institutional inertia, being somewhere a decision taken on Monday could be in front of external stakeholders the next day was a genuinely refreshing change.
The REMAP team in Harare. Myself, Jackie Badze (Project Manager) and Tawanda Masiiwa (Medical Officer), outside ACRN's headquarters in Vainona.
A brief overview of REMAP
I’m going to try to avoid getting too technical, but I think the following is important context for the rest of this post.
As stated previously, Africa carries roughly a quarter of the world's disease burden but hosts under 3% of its clinical trials. The data that could change this already exists inside hospitals as healthcare records, but it is fragmented, non-standardised, and underutilised despite the data already being collected.
REMAP (Real-World Evidence Mapping and Analytics Platform for Africa) is federated real-world-evidence infrastructure.
So what does that actually mean?
REMAP harmonises routinely collected health data to a common language. Instead of moving the data from the institution to one central database, the hospital holds a copy of their data in this universal language, and the analysis is sent to this harmonised data. The results from the analysis are then shared, instead of sharing the raw data. This is what is meant by federated.
As hospitals hold a copy of their data in this universal language, it means the same analysis can be sent to and run across multiple sites. Furthermore, the hospital retains sovereignty over their data. Together, this lets sponsors, researchers and regulators assess trial feasibility, compare treatments, monitor drug safety and surface gaps in care, bringing research, funding and better care to African sites while building the local capacity to sustain them. It also generates local evidence for patients whose care is currently guided by research extrapolated from other populations and other health systems.
Week one: building a fair way to choose pilot sites
Parirenyatwa Hospital (Harare)
I had arrived expecting to spend the six weeks on technical problems: data models, harmonisation, federated queries. My first proper task was none of those. It was building a fair, evidence-led, repeatable and scorable way to judge whether a partner site was feasible for REMAP’s pilot. Leading that with the platform's medical officer, on a continent whose health systems are fundamentally different from the NHS I trained in, meant I had to get familiar with the realities of African healthcare very quickly.
What I learned there set the tone for everything after. First, it showed me the importance of developing mutual understanding, trust and relationships with potential pilot sites for REMAP. Talking to primary investigators, clinicians and a variety of teams at the sites gave me an understanding of how their institutions function and how ACRN could support their development best through REMAP, as well as an understanding their concerns as a partner. Second, it showed me that the instinct to choose partners on their digital maturity alone was a trap: the sites with the least infrastructure were often the most eager to take part and were also the sites that would benefit the most from REMAP. A good site assessment needed to balance technical feasibility with the site’s appetite for REMAP. Trying to create an objective way to assess and score these aspects against defensible, weighted, reproducible criteria was a challenge but ultimately allowed us to gain understanding and build relationships with end-users. It was potentially the most crucial aspect of my placement as it formed the foundation for successful, mutually beneficial partnerships going forward.
Week two: understanding what we were testing during REMAP’s pilot
In the second week, whilst site conversations continued, I spent part of my time defining how we would know whether the pilot had succeeded. I wrote a set of testable hypotheses and assumptions the pilot would stand or fall on, each paired with the evidence that would confirm or disconfirm it and how we could measure and track these KPIs as the pilot progresses.
The discipline of writing falsifiable hypotheses gave me a greater understanding of what we were trying to achieve. However, it also showed me that naming your assumptions up front is uncomfortable precisely because it makes them contestable and that doing so is essential to robust decision making.
Week three: the people REMAP is for
Outside the NaPAZ Conference, Harare.
Much like any research, venture or business, it is crucial to understand who your work is for and who it’s going to impact. I prepared a presentation on utilising federated data, machine learning and advanced analytics to transform research in Africa, delivered by ACRN’s CEO to the National Physicians Association of Zimbabwe’s winter conference. I also had the privilege of representing ACRN on the ground at this conference. Meeting and talking to the physicians REMAP is ultimately meant to serve was an incredible experience where I learnt not only about their thoughts and concerns regarding REMAP but about them as doctors practising and innovating on the continent. Three in four of the doctors present used AI in their clinical work every day, yet most still worked on paper or part-paper records that could not feed those tools, and almost all stated the research opportunities as the biggest benefit to a platform like REMAP. Together, this showed me that the clinicians were ahead of the systems designed to support them, and were driven by a need for locally generated evidence to guide the care of their own patients.
During this week I was also invited to consult on the design of an external, multi-country HIV-care platform trial, as well as being invited to conversations with REMAP’s funders. Arriving in a new country, with a new culture, working environment and in a healthcare system I was previously unfamiliar with had felt daunting at the start. But being asked to contribute, not just observe, reinforced that I was capable and valued.
On the ground at NaPAZ's winter conference
Week four: the organisation, and governance as code
By the fourth week my attention turned to how ACRN itself would build and sustain REMAP. I helped think through what REMAP needed from its next hires and where the real gaps sat, which taught me that an organisation's design is as much a part of the product as its technology. A team can hire excellent people to execute, but it still needs someone holding the cross-cutting story and trajectory together. I also spent time on the technical frameworks and built a proof of concept for REMAP on fully synthetic data, to move the project from a slide-deck to something demonstrating technical feasibility.
In a platform like REMAP, the governance is the technology. The legal, regulatory and data-protection commitments are not policy checkboxes. They are written into how the system executes. Alongside this I was responsible for developing external-facing material, including a comparative view of how REMAP sits against other data networks worldwide. Making a complex proposition, in both technical and governance terms, legible and defensible to a non-technical funder or partner whilst explaining REMAP’s business case was a skill in itself.
Victoria Falls from the air. Mosi-oa-Tunya, the smoke that thunders, a sacred place for the indigenous population. Home to ancestral spirits and a place of spiritual healing.
Week five: what the data can actually give you
After a weekend in Victoria Falls, I left Zimbabwe to move to Johannesburg. These final weeks turned to the data itself. I conducted a scoping review of the literature to understand how others had defined a minimum, adoptable dataset. I used this to assess the data side of REMAP: what routine African records can realistically yield, what the pilot could and could not do with them, and where the genuine analytical value would come from. The lesson was learning which variables make or break an analysis. It’s easy to understand how these variables support an analysis but what is far more challenging is designing a minimum dataset based on variables a site can reasonably curate, as well as how REMAP would support the prospective improvement of data quality, enabling local capacity building at its pilot sites.
Rosebank, Gauteng. Not such a bad view from the office.
Week six: a minimum dataset
The last week went on the largest single piece of work: a first tiered minimum dataset for REMAP, the variables, definitions and quality rules a site could realistically adopt, mapped to a common data model and built on the results of the scoping review that preceded it. Specifying a variable properly is more than just naming it. For each one you need a concrete, universal definition, a way of encoding it that translates cleanly to a shared standard, a sense of what a plausible value looks like, a rule for what to do when it is missing, and integrated data quality checks.
Six weeks taught me that the ceiling on clinical research across much of Africa is not ambition or talent, both are abundant, but the infrastructure to support them. Those foundations (in REMAP the shared standards, the governance, the definitions a site can actually adopt) need to be built with purpose and scalability in mind. That work is slower and less visible than analysis or trial outcomes but is where the real development sits. And a genuine opportunity to leapfrog forward rather than retrofit, and to turn a continent's own data into evidence for its own patients.
Public Healthcare in Zimbabwe and Johannesburg
Parirenyatwa Hospital (Harare). One of the largest teaching hospitals in Zimbabwe, with around 1,800 beds.
Alongside the desk work, I spent time in Parirenyatwa Hospital (Harare, pictured above), and the Chris Hani Baragwanath Hospital (Johannesburg), two of the largest public teaching hospitals on the continent. What struck me most was the distance between the scale of what these teams deal with on a daily basis and the resources they have to work with. Yet, at the same time, the staff were cheery, welcoming and exceptionally talented clinicians. The work they do is remarkable, especially when one considers the conditions and constraints they work within.
At the National Physicians Association of Zimbabwe conference, I was struck by the entrepreneurial drive the doctors had. They were committed to innovation and providing the highest quality of care. With public systems stretched thin, much of the drive for innovation and change came from the clinicians themselves.
This humanity was evident throughout my six weeks in Africa, especially in Zimbabwe. There was a wonderful feeling of togetherness. People look out for one another, are genuinely interested in you, your health and happiness, are caring and kind. People you meet walking down the street stop and talk to you. It was a reminder that medicine and research are, at heart, human endeavours, serving and preserving people and their dignity.
I am grateful to the Laidlaw Foundation and to Imperial's Laidlaw programme for supporting this placement, to Dr Romina Mariano for her supervision, and to the whole team at ACRN for taking a chance on a medical student from the other side of the world.
If you are interested in federated data, the more technical side of REMAP, or ACRN as an organisation, please do not hesitate to contact me (E: zkh22@ic.ac.uk).