Agent-based modelling for epidemiological applications

Abstract

Malaria is an infectious disease induced by protozoan parasite called Plasmodium and transmitted by Anopheles mosquitoes which still imposes a health threat to virtually half of the human populace and causes severe mortality, especially in sub-Saharan Africa.

The characteristics of malaria transmission depend on many relevant factors, such as socio-economic conditions, availability of control measures and local composition of mosquito species. Mathematical modelling of disease propagation is conducted to evaluate the impact of these factors and to ameliorate the lack of experimental data in this field.

In this thesis an agent-based model of short-range mosquito host-seeking behaviour in the presence of Long-Lasting Insecticidal Nets (LLINs) is introduced and further calibrated by the means of experimental data. The model is extended to quantify the overall impact of intervention strategies in community-level situations. As demonstrated in the present study, other optional factors can be included into the simulations, such as parasite ecology and socio-economic conditions. Model simulations are conducted for a short-term period of one day. Additionally, when coupled with regression models based on the simulation results under various conditions, the proposed approach allows transition from the in situ mechanisms of mosquito behaviour to commonly estimated characteristics of malaria transmission.