Complexity

Malaria is complex. To build realistic models, we will need to develop a deeper understanding of malaria epidemiology, malaria transmission dynamics, mosquito ecology, health systems, and vector control. Each one of these can be understood as a separate system that is affected by other parts of the system, but to make any headway, they must all fit together, both conceptually and computationally.

Simple models are a good way to introduce the basic concepts, but they often get it wrong. Perhaps of greater concern, simple models are simply not complex enough to describe some important features of malaria –
they lack the skill sets required to address policy questions. Does the model have a way of representing disease or severe disease by age, spatial dynamics, effects of waxing and waning immunity, or the metrics that are being used for evaluation?

We can also go too far with the complexity. Malaria is complex and messy, but our models don’t need to get every feature of malaria right. There’s a trap in chasing complexity and realism too far. All models are approximations, and we just need to get the gist of it, We probably don’t need to know everything to make good decisions, and we will rarely be able to offer a high degree of certainty. We don’t need a perfect model, but it should be pretty damned good. To develop robust analytics, it’s important to remember that all models are wrong, but some models are useful. (This statement, often used in one form or another, was paraphrased from George P. Box. This approximates the actual quote, but it’s still useful.)

We need some way of doing analysis for policy that can help us make malaria policy without knowing everything. We need some way of building models that are not too complex for the questions we need to address, but that can be repurposed later. In many cases, we will need to elaborate, and we want to guarantee that we don’t get stuck. This is one of the reasons why we developed ramp.xds and a suite of software packages to support nimble model building for simulation-based analytics, or SimBA.

If, as malaria analysts, we want to have a strong understanding of malaria theory to support malaria policies, then we must start simple and add complexity. To get it right, we need to understand malaria spatial dynamics, exogenous forcing, malaria and other human diseases, malaria immunity and care seeking, various kinds of heterogeneity and complexity. Malaria is a complex adaptive system with the major interacting domains: malaria epidemiology (in the narrow sense), malaria transmission dynamics, mosquito ecology, health systems, and vector control.

If we want to understand malaria, it helps to take a birds-eye view of real-world problems, and to try to place malaria within some broader matrix of problems like it. When we do so, it becomes clear that malaria represents an example of a complex adaptive system.

Mathematical models, typically formulated as dynamical systems, can help extend our capacity to understand complex things, and they have have been indispensible for the study of complex adaptive systems. In the case of malaria, we want to use the models as tools to help us understand malaria in populations, to help us develop a rational basis for malaria control.

Also, see the vignette on complex adaptive systems.