Masking significantly reduces, but does not eliminate COVID-19 infection in a spatial agent-based simulation of a University dormitory floor

Abstract

COVID-ADAPT is a stochastic, discrete-space, agent-based simulation model of airborne infection and public health interventions, capable of modeling SARS-CoV-2 transmission. Using a map of a real university dormitory floor, the model simulates agents moving about and potentially infecting each other. Health interventions tested are vaccination and masking, including whether the solitary initial infectious agent was masked. Universal masking with N95 masks and 100% vaccination of susceptible people resulted in significantly lower prevalence after 3 weeks compared to all other scenarios, but still led to a substantial number of infections. Increased vaccination levels from 52% to 100% by itself did not result in a significant difference in prevalence due to symptomatic and asymptomatic breakthrough infections. These results suggest that vaccination alone is insufficient to stem outbreaks, and the best way to reduce COVID-19 infections is to ensure that all infectious people are masked. However, because asymptomatic infections are common, the only way to ensure this is universal masking, which also reduces prevalence by protecting susceptible individuals from infection. Universal masking is the best way forward, especially under threat of the Delta variant, to keep facilities open and safe for occupants, while minimizing the number of infections.