Investigating Leptospira dynamics in a multi-host community using an agent-based modelling approach.

Leptospirosis, a neglected bacterial zoonosis, is a global public health issue disproportionately affecting impoverished communities such as urban slums in the developing world. A variety of animal species, including peridomestic rodents and dogs, can be infected with different strains of leptospirosis. Humans contract leptospirosis via exposure to water or soil contaminated with the urine of infected animals. Due to the unavailability of safe and effective vaccines, preventive strategies mainly focus on minimizing human exposure to contaminated environment. In marginalized communities, this approach is ineffective due to infrastructure deficiencies and the difficulties in implementing sanitation and hygiene practices. Moreover, continuing the expansion of urban slums worldwide will likely contribute to the increase in outbreaks of leptospirosis. Effective prevention of leptospirosis outbreaks will therefore require a thorough understanding of Leptospira transmission dynamics in impoverished, high-density settings. We developed the agent-based model MHMSLeptoDy to investigate Leptospira dynamics in a realistic, in silico high-density community of rodents, dogs and human hosts, and two host-adapted Leptospira strains. Virtual explorations using MHMSLeptoDy were undertaken to evaluate alternate interventions and to assess the zoonotic transmission risk of leptospirosis. A key finding from model explorations is that rodents are the main contributors of rodent-adapted as well as dog-adapted strains in the environment, whereas dogs play an important role in distributing the rodent-adapted strain. Alternate leptospirosis control strategies can be evaluated using the open-source, customizable agent-based model, MHMSLeptoDy. This modelling approach provides a sophisticated mechanism to quantitatively evaluate nuanced intervention strategies and inform the design of rational, locally relevant leptospirosis control programmes.