Modelling: from runways to bednets

M. Henley/ WHO

Can you explain what mathematical modelling is, and how it can be used in the public health world?

Simply put, we use mathematical modelling to improve our understanding of how the malaria parasite affects humans and mosquitoes, and then we use that information to inform how we fight the disease. Modelling is important, because things don’t always work in a linear way – for example, you can give mosquito nets to 20% of people in a certain area, but that doesn't necessarily mean that you will get a 20% reduction in malaria. The people sleeping under the nets will be protected, but the nets are also killing the mosquitoes that land on them – which indirectly benefits even more people.

A tool’s effectiveness will also be determined by the level of exposure a population has to infective mosquito bites. In areas of high transmission, an individual can be exposed to 10 infective mosquito bites each night. Even if you cut this exposure in half, that individual will still be exposed to 5 infective bites and will likely still get malaria. In that setting, you may need a much larger effort in order to see an impact. Modelling helps us understand how these different factors predict the potential impact of various tools in different settings.

You currently head the malaria modelling research group at Imperial College in London, but your early career focused on mathematics. How did you first get involved in malaria research?

I earned my Masters degree in operational research, which was essentially applied modelling. We looked at things like how to time aircraft landing on runways. I can’t say I was particularly good at it, all my airplanes crashed. I moved into epidemiology when I started my PhD work at Imperial College, and subsequently became interested in malaria during my time at the London School of Hygiene & Tropical Medicine. The thing I find most interesting about malaria is the complexity of the disease and the need to understand so many different facets in order to be effective in controlling it – epidemiology, biology, social settings, and so on.

What are some of the key skills needed to be a good modeller?

You need to be able to think in the abstract and ask the right questions. There is no one single model that can answer all of your questions. Instead, you must be able to know what question you want to ask and what abstraction will help you answer that question. In many ways, it’s less about maths and more about the thought processes involved.

Could you describe some of the most important ways mathematical modelling has contributed to the malaria fight in recent years?

Models are helpful along the entire spectrum, from research and development to implementation. In the product development phase, we conduct randomized trials of drugs, vaccines, and other interventions at the community level. Some of these trials are complex, and models are quite helpful for generating simulations that inform trial designs, as well as the interpretation and generalization of results.

In terms of implementation, various interventions have been scaled up in different ways in different places. Modelling has been applied to help us determine the impact of these efforts. It can also be used to help us determine the impact of different combinations of tools and the best rational use of existing resources.

Modelling also helps inform country applications for funding. Donors look for results-based frameworks – they want you to write down plans and budgets that translate into coverage estimates and what the expected change in malaria burden is going to be. That last bit is the hardest for anyone to do, because it requires projection. And that’s where modelling comes in. Do the impact projections generated by modelling match your end goal? If not, what can be changed?

How would you respond to those who might suggest that there is a lot of uncertainty in the numbers that these models generate?

In terms of accuracy, I think it’s is important to remember that we don’t even know exactly how many people there are in the world, so no one expects to have perfect estimates of the number of people that have a particular disease. So yes, there is a certain amount of uncertainty around those estimates, but that’s because the raw data simply doesn't exist in many countries – particularly those with underdeveloped health systems that have the highest malaria burden. As countries improve their data systems, this uncertainty declines. Thus, strengthening surveillance in malaria endemic countries is critical.