Abstract
This study used a simulation to examine nonpharmaceutical interventions (NPIs) that could have been implemented early in a COVID-19 surge to avoid a large wave of infections, deaths, and an overwhelmed hospital system. The authors integrated a dynamic agent-based travel model with an infection dynamic model. Both models were developed with and calibrated to local data from Los Angeles County (LAC), resulting in a synthetic population of 10 million agents with detailed socio-economic and activity-based characteristics representative of the County’s population. The study focused on the time of the second wave of COVID-19 in LAC (November 1, 2020, to February 10, 2021), before vaccines were introduced. The model accounted for mandated and self-imposed interventions at the time, by incorporating mobile device data providing observed reductions in activity patterns from pre-pandemic norm, and it represented multiple employment categories with literature-informed contact distributions. The combination of NPIs—such as masks, antigen testing, and reduced contact intensity—was the most effective, among the least restrictive, means to reduce infections. The findings may be relevant to public health policy interventions in the community and at the workplace. The study demonstrates that investments in activity-based travel models, including detailed individual-level socio-demographic characteristics and activity behaviors, can facilitate the evaluation of NPIs to reduce infectious disease epidemics, including COVID-19. The framework developed is generalizable across SARS-COV-2 variants, or even other viral infections, with minimal modifications to the modeling infrastructure.
