The impact of heterogeneous transmission on the establishment and spread of antimalarial drug resistance

• We embedded within-host competition in an agent-based model of malaria.
• We examined the role of heterogeneous biting on the establishment and spread of drug resistance.
• We find that heterogeneous transmission slows the establishment of resistance in a population.
• We find that heterogeneous transmission speeds the spread of resistance once established.

Despite the important insights gained by extending the classical models of malaria, other factors, such as immunity, heterogeneous biting, and differential patterns of drug use have not been fully explored due to the complexity of modeling multiple simultaneous malaria infections competing within a host. Understanding these factors is important for understanding how to control the spread of drug resistance to artemisinin which is just emerging in Southeast Asia. The emergence of resistance plays out at the population level, but is the result of competition within individuals for transmission events. Most studies of drug resistance evolution have focused on transmission between hosts and ignored the role of within-host competition due to the inherent complexity of modeling at multiple scales. To embed within-host competition in the model, we used an agent-based framework that was developed to understand how deviations from the classical assumptions of the Ross-MacDonald type models, which have been well-described and analyzed, impact the dynamics of disease. While structured to be a stochastic analog to classical Ross-Macdonald type models, the model is nonetheless based on individuals, and thus aspects of within-host competition can be explored. We use this framework to explore the role of heterogeneous biting and transmission on the establishment and spread of resistance in a population. We find that heterogeneous transmission slows the establishment of resistance in a population, but once resistance is established, it speeds the spread of resistance through the population. These results are due to the skewed distribution of biting which makes onward transmission a low probability and suggests that targeting the “core” group of individuals that provide the vast majority of bites could significantly slow the spread of resistance.