The spread of SARS-CoV-2 virus in Hungary – agent-based modelling scenarios for the first wave

Abstract

The COVID-19 pandemic, caused by the SARS-CoV-2 virus is one of the biggest challenges of the 21st century humanity had to face so far. In the hopes of contributing to the scientific discussion about the timeline and the spatial patterns of the spread, the paper presents a modelling study of the first wave of the epidemic in Hungary. The self-designed program implements agent-based methodology, with the agents representing the inhabitants of Hungary. Six types of social situations with a risk of infection were distinguished: household contacts, education, workplace, everyday activities, leisure activities, transportation. Based on their socioeconomic characteristics, the agents were distributed into the following eight activity groups: children under three years; kindergarteners and primary school pupils; high schoolers; university students; commuters; local employees; people between the age of 15-64 and unemployed, economically inactive or in home office; elderly. Each activity group shows a different frequency of social contacts in the defined social situations. The created programme also takes the linkages between the municipalities into consideration, and in order to simulate the intermunicipal flows, the gravity-model was adapted. To explore the different possibilities, 12 scenarios were created. The scenarios differ from each other in their base reproduction number, the initial location of the infection, the existence of restrictions, and the time-delay of the restrictions.

The unrestricted scenarios draw out the expected exponential growth pattern with six-day duplication. The results indicate that under unrestricted circumstances, workplace contacts have the highest significance in the spread, followed by contacts during education and within households. Under unrestricted scenarios, university students and commuters are the most susceptible to the infection. The spread between the municipalities shows a hierarchical diffusion pattern, combined with dispersion within the commuting zones. At an early stage, the virus shows a tendency for concentration, due to the significance surplus of the largest cities. The results indicate that the restrictions successfully limited the hierarchical diffusion of the virus. After the full implementation of the restrictions, the effective reproduction number sunk below 1 in every scenario, for every activity group. However, the relative significance of infections during household contacts increased, which may contribute to an extended epidemic. The probability of the scenarios was discussed in the light of the official numbers and the H-UNCOVER representative survey. A correlation analysis between the two most likely scenarios and the registered infections was also completed. The 4A scenario was deemed to have the highest probability, which indicates a considerably earlier epidemic onset than the date of first official detection.