Fighting Malaria Together Newsletter - December 2010

Using mathematical modeling to inform malaria control

An interview with Nakul Chitnis (Swiss Tropical Institute) the lead author of a new report that introduces modeling to the malaria community and expands the dialogue on priority decision-making and when and how modeling can help.

What is the significance of the modeling report?

Mathematical modeling has a useful role to play in malaria control and elimination at all levels.  It is particularly important in informing planning and resourcing activities. Modeling can help guide global prioritization of interventions and target product profiles of new interventions. It can determine optimal combinations of interventions in various settings and provide a theoretical underpinning to expected reductions in malaria measures from target coverage levels. The report gives three examples of how mathematical modeling can be applied in a country setting: building models to understand the potential role of surveillance, to introduce malaria vaccines, and to understand the potential implications of combining bednet distribution with indoor residual spraying.

What are some of the challenges facing the use of mathematical modeling in malaria control?

The malaria modeling and malaria control communities often exist in separate spheres with little communication between them. There needs to be more communication and interaction between them for modeling to play the kind of role I just described. The modeling community needs to engage more with the control community, use more accessible language in expressing their results, and define the limitations of their research and results.

The malaria control community can best obtain modeling data by interacting with the modeling community and understanding the implications and limitations of modeling results.

Is modeling useful for malaria advocacy?

Yes. Modeling can help to determine priorities for funding of deployment of current interventions and research and development of new interventions. It can also help to show and document the potential benefits of applying a given control strategy.

Where do you see malaria modeling 10 years from now?

I would hope that modeling would move in the direction where weather forecast models are now. Anyone with an internet connection has access to a simple interface that shows the predicted weather (from an ensemble of models) and the uncertainly underlying the prediction, without needing to understand the details of the weather models.

While models for malaria do not need to go that far, a goal would be have an ensemble of different models with an easy-to-use interface that malaria control planners can use to obtain model results and the uncertainty underlying them.

Anything else you'd like to mention?

I think it's important for the users of model results to remember that all models are built upon a set of assumptions and parameter values determined from data that inherently contains errors. When used in situations where the assumptions hold, models can provide reliable and useful information. However, when the results are used where the assumptions fail, they are unreliable and sometimes dangerous. On the other side, it behooves the modeling community to be clear about the limitations of model results and their underlying uncertainty.