Psychological Factors in Agent-Based Epidemic Models: how Behavior Shapes Disease Outcomes

Abstract

In addition to the use of real-world behavioral data, pandemic research profits from behavioral modeling in simulations. In this research, an agent-based model (ABM) was applied to examine the spread of infections under two behavioral assumptions. The modeling approach was implemented using the German Epidemic Modeling System (GEMS) in Julia. The MPICE model, which accounts for individual perceptions and social influences on adherence to self-isolation measures, was applied to a 120-day simulation, in which we examined the spread of a communicable infectious disease within a population (N = 10,000). A Two-Sample t-test revealed a significantly higher disease spread for modeled behavior than for random behavior. This showed that unlike the random behavior model, where preventive actions are implemented probabilistically and uniformly, the psychological model likely introduced delays and inconsistencies in adherence due to variability in individual perceptions and social influences. Sensitivity analysis further showed the relevance of the model‘s components.