Researchers at the University of Edinburgh are using maths to help people in remote communities access aid more quickly.
In collaboration with Médecins Sans Frontières (MSF) / Doctors Without Borders, they have developed an algorithm that helps aid agencies optimise where to site humanitarian relief, healthcare, or medical facilities. This could save thousands of hours of avoidable travel and delay in affected communities.
The University’s Dr Kit Searle – who led the algorithm’s development – modelled its potential impact in a test based on sites in South Madagascar. The algorithm optimised the placement of just five MSF facilities, reducing the time taken for local people to access aid by more than 40,000 hours.
This figure represents the total time saved, cumulatively, by everyone in the entire population of the impacted area.
Access aid
“In high-income countries it can be quite difficult to think about the importance of making aid accessible,” Dr Searle said.
“In very remote areas with little to no infrastructure, where the only transport option is to pick up your feet and walk, people can be forced to travel for hours or even days to receive care or the aid that they may urgently need.
“MSF has developed an awesome tool that measures accessibility, and our algorithm then uses this information to identify where to site resources in a region that will maximise the impact and accessibility of those new facilities.”
Open source
The algorithm enhances MSF’s existing open-source geographical model. This combines map data, including topography, vegetation cover, watercourses, and roads and other infrastructure, and population statistics.
The University’s algorithm extrapolates from that data. This determines how long people from each community could take to access aid in a particular location and how many are affected. The algorithm then recommends where services would be best placed to minimise travel time for the most people. Such services include, for example, health clinics, vaccination points, shelters, water and sanitation points, and aid distribution centres.
Interface
MSF’s planners and decision makers access the technology using a simple web-based interface. They can draw an initial area on a map to quickly receive suggested optimum locations for assets in that area.
Based on the resources available, they can adjust the number of facilities they are able to put into an area. This shows how having more, or fewer, facilities affects the recommendations. They can also modify the target area. By making it larger, smaller, or redrawing it completely, they immediately see how the algorithm’s suggestions change.
Another factor considered is the transport infrastructure available and its quality. This means the algorithm can update recommendations to mitigate for developing circumstances that affect people’s ability to access aid. This could include events such as a road or bridge becoming impassable.
Collaboration
Development of the algorithm has been a truly collaborative venture between the University’s staff and students, and MSF.
Dr Searle, from the School of Mathematics and Maxwell Institute for Mathematical Sciences, led the collaboration for the University. A lecturer and MSc programme director in Operational Research, his research interests include combinational optimisation, such as exact and heuristic solution approaches to solving practical problems in facility location, vehicle routing and scheduling.
He reached out to MSF’s geospatial consultant Dr Andries Heynes to see if they could offer a student placement. Dr Searle and Dr Heynes previously shared a PhD supervisor while studying at the University of Stellenbosch, South Africa.
Placement
Dr Searle takes up the story: ‘I knew Andries was at MSF doing geospatial analysis. I thought maybe there’s a fun project we can do for a student there.”
In the summer of 2024, Dr Searle supervised Qixuan Zhang, an MSc Operational Research student, on her dissertation. She wrote an initial algorithm determining the optimal location of a fixed number of medical facilities.
“Qixuan worked on the project for 12 weeks,” Dr Searle said. “MSF were great with her. They arranged travel for her to visit their office in London at the end of the placement, to present her findings to their director and other key stakeholders and were quite happy with everything that was done. Qixuan was amazing – she did really good work. It was a really good starting point that made the project easier to take forward for everyone involved.”
This resulted in the follow up project, funded by the International Centre for Mathematical Sciences. This has seen Dr Searle extend the algorithm. This included adding objectives such as making fairness a consideration when recommending facility locations, minimising the construction of new facilities, and integrating the algorithm with MSF’s online tool.
Next steps
Dr Searle’s work with MSF continues. He believes there is room to develop the algorithm further for even more nuanced analysis.
“Currently we’re at the strategic level, planning where to locate these kinds of facilities. Next, we want to consider all these reactive, emergency scenarios where, maybe there’s an earthquake, roads have been damaged so they’re no longer usable. Can we have confidence in the accessibility of facilities we’re building then?
“Maybe there’s flooding and the extent of the river changes so bridges and fords are affected – what does that do to accessibility? If there’s an outbreak of disease, maybe we need to put mobile clinics into a region and move them around.
“Accommodating these scenarios are the immediate next steps that we’re going to take and they’re much more difficult problems to solve from a mathematical point of view.”
Impact
MSF is an international medical humanitarian organisation. It provides lifesaving emergency relief and longer-term medical care to vulnerable and excluded communities in more than 70 countries around the world.
Jose Luis Álvarez Morán is Epidemiology and Public Health Coordinator for MSF. He says the collaboration with Dr Searle is creating a positive impact, now and for the future.
“The University’s work is helping us add new functionalities to our model, to help our projects better understand the best location for our clinics and activities.
“It directly contributes to our aim of bringing medical humanitarian assistance to populations in distress, victims of natural or man-made disasters, and victims of armed conflict, regardless of ethnic origin, religion, gender, or political affiliation.
“We hope the model becomes a regular tool for our staff, and not only for MSF but all agencies working in humanitarian settings.”
Image credits: Main image – Ashley Cooper/Getty. All other images – Andrew Perry.
