A Statistical Examination of Distinct Characteristics Influencing the Performance of Vector-Borne Epidemiological Agent-Based Simulation Models
Academic Article
Overview
Research
Identity
Additional Document Info
Other
View All
Overview
abstract
The spread of infectious diseases is a complex system in which pathogens, humans, the environment, and sometimes vectors interact. Mathematical and simulation modelling is a suitable approach to investigate the dynamics of such complex systems. The 2019 novel coronavirus (COVID-19) pandemic reinforced the importance of agent-based simulation models to quickly and accurately provide information about the disease spread that would be otherwise hard or risky to obtain, and how this information can be used to support infectious disease control decisions. Due to the trade-offs between complexity, time, and accuracy, many assumptions are frequently made in epidemiological models. With respect to vector-borne diseases, these assumptions lead to epidemiological models that are usually bounded to single-strain and single-vector scenarios, where human behavior is modeled in a simplistic manner or ignored, and where data quality is usually not evaluated. In order to leverage these models from theoretical tools to decision-making support tools, it is important to understand how information quality, human behavior, multi-vector, and multi-strain affect the results. For this, an agent-based simulation model with different parameter values and different scenarios was considered. Its results were compared with the results of a traditional compartmental model with respect to three outputs: total number of infected individuals, duration of the epidemic, and number of epidemic waves. Paired t-test showed that, in most cases, data quality, human behavior, multi-vector, and multi-strain were characteristics that lead to statistically different results, while the computational costs to consider them were not high. Therefore, these characteristics should be investigated in more detail and be accounted for in epidemiological models in order to obtain more reliable results that can assist the decision-making process during epidemics.
